Emanuele Regni - Ingegnere Civile - A.A. 2014-2015 by Tema Grafico

Università degli Studi di Ferrara

UNIVERSITA’ DEGLI STUDI DI FERRARA

CORSO DI LAUREA MAGISTRALE IN INGEGNERIA CIVILE

MECHANICAL CHARACTERIZATION OF HISTORICAL MORTARS AND BRICKS USING MINOR DESTRUCTIVE TECHNIQUES

Laureando Emanuele Regni Relatore: Prof. ing. Alessandra Aprile Secondo Relatore: Prof. ing. Luca Pelà

Anno Accademico 2014/2015

Ringraziamenti La presente tesi è stata sviluppata all’ UPC, Universitat Politècnica de Catalunya. E’ un lavoro di ricerca svolto sotto la supervisione del professor Luca Pelà. Vorrei ringraziarlo moltissimo per il suo aiuto nei momenti difficili, per quanto riguarda le questioni puramente tecniche, ma anche per il sostgno morale e la motivazione che ha saputo darmi. Ed un ringraziamento non meno importante anche al professor Pere Roca per la sua disponibilità, pazienza e profonda generosità. Sono molto grato alla professoressa Alessandra Aprile, non solo per l’opportunità fantastica che mi ha regalato ma anche per i messaggi positivi che ha sempre voluto e saputo trasmettermi ogni volta che mi sono confrontato con lei. Ha a cuore il futuro degli studenti e lo dimostra sempre. Lo staff del laboratorio mi ha aiutato tantissimo nella campagna sperimentale, mostrando una pazieza straordinaria, creando anche un ambiente di lavoro accogliente e sereno. Come non ricordare quindi Tomás Garcia, Camilo Bernard, Robert Michael, e Jordi Cabrerizo. Avrò sempre dei bellissimi ricordi. Ora un ringraziamento a tutti i miei colleghi e soprattutto amici che mi hanno accompagnato ed aiutato in questi anni di studi. Devo assolutamente ricordare Alessandro Boccato, il primo grande amico all’università, con il quale ho preparato molti esami, i carissimi Federico Pizzarulli, Simone Totolo, Michele Valarin, Federico Ferro, voglio menzionarli uno ad uno per il loro aiuto e per ciò che abbiamo condiviso assieme; ci sarebbe molto da scrivere. Porterò sempre nel cuore Simone Terenzi e Mattia Ferri come i due coinquilini più straordinari del mondo. Un ringraziamento speciale ai miei due fantastici coinquilini del soggiorno a Barcellona: Diego e Stefano, per il loro altruismo. I ringraziamenti più importanti alla mia famiglia: oggi si sta laureando con me mia madre. Ha sofferto e gioito con me ad ogni passo, mi ha supportato in tutto, come diciamo spesso io e mia sorella: dovrebbero darle una laurea honoris causa. Mia sorella, un esempio per me, per la vita, oltre che la mia migliore amica. Ma non sarei arrivato fin qui senza mio padre, un uomo di

grande cuore, che mi ha sostenuto e mi ha sempre ricordato qunto valgo, portando pazienza, capendo i miei nervosismi, non facendomi pesare talvolta i miei malumori. Voglio ringraziare loro e tutti gli amati parenti. Un ringraziamento ai miei amici di una vita. Voglio menzionare Stefano (un fratello maggiore per me), Andrea, Corinna, persone che mi hanno davvero voluto bene, ed altri, Christian, Stefano, Fabio. Vorrei nominare tutte le persone che ho a cuore, in questo fantastico giorno. Unâ&#x20AC;&#x2122;importantissima tappa si Ă¨ conclusa ed una nuova storia nella mia vita sta per cominciare. In questa avventura voglio ricambiare almeno parte del bene ricevuto e lo farĂ˛ trovando la mia realizzazione, la mia felicitĂ e restituendo nei fatti qualcosa di buono a tutte le persone che mi hanno voluto bene e che mi hanno sostenuto!

Abstract The thesis is focused on the study and characterization of historical mortars and bricks using two instrument wich allow to performe minor-destructive tests. The instruments are: ď&#x192;ź Helifix screw-pull-out system, wich measures the shear strength of a small cylinder or annulus of the material which is enganged by a helical self-tapped screw. ď&#x192;ź Windsor Pin System, which shoots a steel pin into the surface of concrete or mortar and the penetration depth can be measured. From the datas obtained with these two instruments the aim is to estimate the mechanical properties such as tensile and especially compressive strength of the materials. For the two instruments the existing correlation curves are often inconsistent, and absolutely not suitable for historical mortars. As can be seen from previous researches, the compressive strengths are often overestimated. Our goal is to improve the curves, or to create new curves by testing of a lot of mortars and bricks of different expected resistances. Another purpose is to develop analytical models for determining the resistance from the data provided by the two apparatus. The materials used to build the specimens were handmade bricks and hydraulic lime mortar, without cement content, to reproduce an historical existing masonry. To investigate a certain range of resistances, some wallets were built with different mortars. On these wallets the tests were performed at different ages of maturation. Minor destructive tests and direct destructive tests on the same materials were made to permite appropriate correlations. A test program was carried out on a historical building to verify if the techniques developed and the analytical interpretations shall be valid for a real existing structure. A general purpose of this thesis is to promote future works on these innovative minor destructive test techniques and to encourage the improvement of the existing standards.

Sommario

Titolo: “CARATTERIZZAZIONE MECCANICA DI MALTE E MATTONI STORICI ATTRAVERSO L’USO DI TECNICHE NON INVASIVE” La tesi si focalizza sullo studio a piccola scala di malte e mattoni storici, tramite l’uso di due strumentazioni: “The Helifix Load Test Unit” e “The Windsor Pin System”.  La “Helifix Load Test Unit”, che fondamentalmente misura la resistenza al taglio di una piccola corona cilindrica che si fissa e viene estratta attraverso un’ elica metallica. La resistenza al taglio è quindi relativa alla superficie cilindrica di rottura del materiale.  Il Windsor Pin System, che spara un chiodo metallico sulla superficie del materiale, e successivamente è possibile misurare accuratamente la profondità di penetrazione. Dai dati ottenuti, rispettivamente la forza di estrazione della corona cilindrica e la profondità di penetrazione, l’obbiettivo è misurare la resistenza a compressione del materiale. Esistono infatti curve di correlazione, ottenute a seguito di numerosi test di compressione diretta e prove non distruttive su malte di cemento e calcestruzzi, che forniscono la resistenza a compressione di questi materiali a fronte dei dati rilevati dai due strumenti. Queste curve forniscono risultati ancora troppo incerti e non affidabili, e si è osservato che non sono assolutamente adatte a descrivere il comportamento meccanico delle malte storiche: si hanno sempre delle sovrastime. Per gli edifici storici sarebbe invece importantissimo poter ricavare le proprietà meccaniche dei materiali tramite test non invasivi, in modo da salvaguardare l’architettura e l’estetica, evitando il prelievo di campioni, che può addirittura ammalorare ulteriormente strutture già datate e rovinate. Il nostro obbiettivo è quindi quello creare nuove curve che descrivano bene il comportamento delle malte più deboli. Studiando il fenomeno fisico alla base dei due test, si cercherà inoltre di fornire un modello analitico capace di stimare la resistenza a compressione a fronte delle letture strumentali. I materiali utilizzati per costruire i provini sono mattoni di terracotta fatti a mano e malte idrauliche ed aeree, prive di cemento, per riprodurre le malte storiche. Per investigare un certo

intervallo di resistenze, saranno costruiti alcuni muretti con diversi tipi di malta. Test non invasivi e test diretti di compressione e flessione sugli stessi materiali, a differenti etĂ di maturazione, saranno effettuati in modo da consentire le appropriate correlazioni. Un programma di test sarĂ inoltre effettuato su un edificio esistente, in modo da verificare che le interpretazioni ed i modelli forniti funzionino su una struttura esistente, dove le proprietĂ dei materiali, che hanno subito le influenze climatiche e del tempo, non sono direttamente controllabili come in laboratorio.

Table of Contents 1. Introduction ..................................................................................................... 1 1.1 Motivation of the research ......................................................................................... 1 1.2 Aims and Methodology .............................................................................................. 2 1.3 Outline of the thesis .................................................................................................... 3

2. State of the art.................................................................................................. 5 2.1 Overview on Lime Mortars ....................................................................................... 5 2.1.1 History of lime mortar ........................................................................................... 5 2.1.2 Types of lime: hydraulic lime and aerial lime....................................................... 5 2.1.4 Mechanical characterization and laboratory testing of mortars ............................ 9 2.2 Bricks ......................................................................................................................... 11 2.2.1 History and origin................................................................................................ 11 2.2.2 The brick making process ................................................................................... 12 2.2.3 Bricks: general properties.................................................................................... 17 2.2.4 Mechanical characterization and laboratory testing of bricks ............................. 19 2.3 In-situ minor destructive testing of mortar and bricks ........................................ 20 2.3.1 Windsor pin Penetromer system ......................................................................... 20 2.3.2 Helifix screw pull-out system ............................................................................. 24 2.3.3 Other minor-destructive tests (MDT) .................................................................. 30

3. Laboratory experimental program ............................................................. 35 3.1 First campaign on lime mortar ............................................................................... 35 3.1.1 Materials .............................................................................................................. 36 3.1.2 Construction of wallets and specimens ............................................................... 43 3.1.3 Flexural tests on prismatic specimens of mortar ................................................. 49 3.1.4 Compressive tests on mortar prismatic specimens .............................................. 57 3.1.5 Brazilian tests on prismatic specimens of mortar................................................ 64 3.1.6 Double Punch Tests on mortar joints .................................................................. 69 3.1.7 Helifix screw pull-out tests on wallet joints ........................................................ 84 3.2 Second campaign on lime mortars .......................................................................... 90 3.2.1 Materials .............................................................................................................. 90 I

Table of contents 3.2.2 Construction of wallets and specimens ............................................................... 91 3.2.3 Flexural tests on prismatic specimens of mortar ................................................. 93 3.2.4 Compressive tests on prismatic specimens of mortar ......................................... 94 3.2.5 Double Punching Tests on mortar joints ............................................................. 96 3.2.6 Helifix screw pull-out tests on walletâ&#x20AC;&#x2122;s joints ................................................... 102 3.2.7 Windsor Pin tests on wallet joints ..................................................................... 109 3.3 Campaign on bricks ............................................................................................... 113 3.3.1 Materials ............................................................................................................ 114 3.3.2 Compression tests of units................................................................................. 115 3.3.3 Compressive tests of cubic and prismatic specimens........................................ 120 3.3.4 Compressive tests of cylindrical specimens ...................................................... 124 3.3.5 Flexural tests on prismatic specimens of brick ................................................. 126 3.3.6 Brazilian tests on prismatic specimens of bricks .............................................. 127 3.3.7 Tensile tests on prismatic specimens of bricks ................................................. 130 3.3.8 Helifix screw pull-out tests on bricks ................................................................ 133 3.3.9 Windsor Pin tests on bricks ............................................................................... 135

4. In-situ experimental program .................................................................... 137 4.1 Mechanical characterization of the materials of Puig i Cadafalch house, Argentona, Spain .......................................................................................................... 137 4.1.1 Description of the case study ............................................................................ 137 4.1.2 Analysis of the available material and preparation of the specimens ............... 138 4.1.3 Double Punch Tests on mortar joints ................................................................ 141 4.1.5 Laboratory tests on bricks ................................................................................. 143 4.1.4 Helifix screw pull-out tests on mortar ............................................................... 145 4.1.5 Windsor Pin test on mortar................................................................................ 148 4.1.6 Minor destructive tests on bricks ...................................................................... 149

5. Discussion of results .................................................................................... 150 5.1 Results of tests on lime mortar .............................................................................. 150 5.1.1 Compressive tests VS Double Punch Tests....................................................... 150 5.1.2 Compressive tests VS Flexural Tests ................................................................ 155 5.1.3 Brazilian tests VS Flexural Tests ...................................................................... 157 5.1.4 Double Punch Test VS Helifix screw pull-out tests .......................................... 159 II

Table of contents 5.1.5 Compressive standard test VS Helifix screw pull-out tests .............................. 170 5.1.6 Double Punching Tests VS Windsor Pin tests .................................................. 172 5.1.7 Compressive tests VS Windsor Pin tests........................................................... 174 5.1.8 Flexural tests VS Helifix screw pull-out tests ................................................... 176 5.1.9 Flexural tests VS Windsor Pin tests .................................................................. 177 5.2 Results of in-situ experimental program .............................................................. 178 5.2.1 Helifix screw pull-out test on mortar joints ...................................................... 179 5.2.2 Windsor pin test on mortar joints ...................................................................... 182 5.2.3 Compressive strength of the mortar predicted from the pull-out force ............. 183

6. Conclusions .................................................................................................. 185 6.1 Summary ................................................................................................................. 185 6.2 General and specific conclusions .......................................................................... 186 6.3 Suggestions for future work .................................................................................. 187

III

List of figures Figure 2.1.1 – Building lime family chart according to EN-459-1 ............................................ 6 Figure 2.1.2 – View (a and results (b of the double punch test as a function of the thickness. 10 Figure 2.2.1 - a) Zigurat Al-Untash (Irán); b) Temples of Bagan (Birman). ........................... 12 Figure 2.2.2 - Diagrammatic Representation of Manufacturing Process ................................. 13 Figure 2.2.3 – Clay or shale being crushed and transported to storage area. ........................... 14 Figure 2.2.4 – Clay is thoroughly mixed with waterin pug mill before extrusion. .................. 14 Figure 2.2.5 – After mining, clay is extruded through a die and ............................................. 15 Figure 2.2.6 – Brick Enter Tunnel Kiln for Firing ................................................................... 16 Figure 2.2.7 – Bricks dimensions. ............................................................................................ 19 Figure 2.3.1 – Windsor pin system WP-2000 .......................................................................... 21 Figure 2.3.2 – Determination of pin reusability; a) Two possible pin configuration after the test; ........................................................................................................................................... 21 Figure 2.3.3 – Correct Style to performing penetrometer test. ................................................ 22 Figure 2.3.4 – Cleaning of the hole by the rubber bulb-type blower. ...................................... 22 Figure 2.3.5 – Measuring hole depth; a) Phase 1; b) Phase 2. ................................................. 23 Figure 2.3.6 – a) Circumstance of reading micrometer-Manual of Windsor pin system; b) Micrometer. .............................................................................................................................. 23 Figure 2.3.7 – Installation of helical tie for screw pull-out test. .............................................. 24 Figure 2.3.8 – Selection of the points and drilling holes.......................................................... 25 Figure 2.3.9 – Installation of the tie. ........................................................................................ 25 Figure 2.3.10 –Gripping device attaching. ............................................................................... 26 Figure 2.3.11 – Load testing the tie. ......................................................................................... 26 Figure 2.3.12 – Vekey 1991 “In –situ tests for maronry”; a) Calibration of screw-pull-out for mortars;

b) Calibration of screw-pull-out for UK AAC blocks. .......................................... 27

Figure 2.3.13 – a) Scattergram of lower strength masonry unit tests; b) Scattergram of mortar test results. ................................................................................................................................ 29 Figure 2.3.14 – Inspections: a) Coring (Binda et al., 2001); b) Endoscopy (Binda et al., 2001). .................................................................................................................................................. 31 Figure 2.3.15 – Equipment and setup for flat jack test. ........................................................... 31

List of figures Figure 2.3.16 – Xdrill: photo and geometry ............................................................................. 32 Figure 2.3.17 – The X-drill is normally knocked 15-20 mm into the predrilled hole.............. 33 Figure 2.3.18 – Measurement of the maximum moment of torsion in the moment of fracture. .................................................................................................................................................. 34 Figure 3.1.1 – a) Kerakoll mortar (Biocalce Muro); b) NHL3,5_1:3 “Cementos Tigre” c) CL90_powder aerial lime “Ciaries”..........................................................................................37 Figure 3.1.2 – Slaked lime CL90 “Segarra e Hernandez”; a) Sack of slaked lime (20 Kg); b) Hydration in the mixing machine with an excess water; c) Storage of lime in two plastic buckets and periodical mixing with helical accessory connected to a drill............................39 Figure 3.1.3 – Particle size in the technical data sheet of the product...................................40 Figure 3.1.4 – a) River sand 0-2 mm “Arids Catalunya Sorigué; b) Vibrating screen; c) Remaining sand in each sieve; d) Remaining sand weighing...................................................42 Figure 3.1.5 – Results of particle size made in the laboratory.................................................42 Figure 3.1.6 – a) Terra Cuita bricks; b) Nominal dimensions 345x145x45 mm....................43 Figure 3.1.7 – Disposition of the walls in the laboratory..........................................................44 Figure 3.1.8 – a) Cutting bricks by circular saw; b) Immersing and keeping under water for 15 – 20 min...................................................................................................................................46 Figure 3.1.9 – a) Vertical and horizontal joints leveled; b) Wallet dimensions (7 rows, 2 bricks each row)..................................................................................................................................46 Figure 3.1.10 – 16 wallets built on wooden beams...................................................................47 Figure 3.1.11 – a) Handmade wooden molds; b) 6 specimens in steel molds and 3 specimens in wood mold...........................................................................................................................47 Figure 3.1.12 – Molds making procedure: a) Sprinkling oil inside molds surface; b) 25 strokes with a piston; c) 25 strokes with the compacting machine; d) Excess of mortar removing; e) Putting the molds in plastic bags, to get humidity between 90-95%; f) 4 days in the climatic chamber; g) Specimens Maturation near the wallets, at the same climatic conditions..................................................................................................................................49 Figure 3.1.13 - Flexure test 1015:11; a) Setup of the test; b) Correct position of the sample; c) Failure mode of the sample......................................................................................................50 Figure 3.1.14 – Flexural strength VS Time (EN 1015:11)_ mortar NHL 3,5_1:3; a) f flex VS Time, specimens made in STEEL MOLDS; b) fflex VS Time, specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS.....................................52

List of figures Figure 3.1.15 – Flexural strength VS Time (EN 1015:11)_Kerakoll mortar “Biocalce muro M5; a) fflex VS Time, specimens made in STEEL MOLDS; b) fflex VS Time, specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS......................54 Figure 3.1.16 – Flexural strength VS Time (EN 1015:11)__ mortar CL90_1:3 “Ciaries” (with powder aerial lime); a) fflex VS Time, specimens made in STEEL MOLDS; b) fflex VS Time, specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS.....................................................................................................................................55 Figure 3.1.17 – Flexural strength VS Time (EN 1015:11)_ mortar CL90_1:3 “Segarra” (with slaked lime) a) fflex VS Time, specimens made in STEEL MOLDS;

b) fflex VS Time,

specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS.....................................................................................................................................56 Figure 3.1.18 - Compression test EN 1052:11:2007................................................................57 Figure 3.1.19 - NHL 3,5_1:3: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparison of results: STEEL MOLDS VS WOOD MOLDS................................................................................................................58 Figure 3.1.20 – Kerakoll mortar M5: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparison of results: STEEL MOLDS VS WOOD MOLDS.................................................................................................60 Figure 3.1.21 – CL90_1:3 “Ciaries”: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparation of results: STEEL MOLDS VS WOOD MOLDS.................................................................................................62 Figure 3.1.22 – CL90_1:3 “Segarra”: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparison of results: STEEL MOLDS VS WOOD MOLDS.................................................................................................63 Figure 3.1.23 - Brazilian test; a) Test setup; b) Failure in the whole prisms

Figure 3.1.24 – Brazilian test results, Tensile strength VS Time; a) Kerakoll mortar M5; b) NHL 3,5_13 “Cementos Tigre”; c) CL90_1:3 “Ciaries”; d) CL90_1:3 “Segarra”................68 Figure 3.1.25 – DPT, specimens' preparation: a) Removin joints by hammer and chisel; b) Drawing squares 5x5 cm on the removed joints; c) Cutting removed joints by a circular saw; d) Samples (at least 4 each joint); e) Gypsum paste on half group of joints; f) Measuring specimens' thickness................................................................................................................69 Figure 3.1.26 – Test setup (DIN 8555-9:1999).........................................................................71 Figure 3.1.27 – Double Punch Test setup: a) Ibertest and configuration with fixed punches; b) Centered specimen; c) Typical hourglass failure mode of central zone.................................71 VI

List of figures Figure 3.1.28 – a) Typical failure mode of a specimen with gypsum layer...........................72 Figure 3.1.29 – DPT_Kerakoll mortar M5_Campaign 1_AGE2 (28 Days)...........................74 Figure 3.1.30 – a) and b) show difficulties in joint’s extraction..............................................75 Figure 3.1.31 – NHL 3,5_1:3 DPT: fcDPT VS Time..................................................................76 Figure 3.1.32 – Kerakoll mortar M5_DPT: fcDPT VS Time.....................................................79 Figure 3.1.33 – CL90_1:3 “Ciaries”_DPT: fcDPT VS Time; a) Specimens with gypsum; b) Specimens without gypsum; c) fcDPT VS Time__with gypsum VS without gypsum..............81 Figure 3.1.34 – CL90_1:3 “Segarra”_DPT: fcDPT VS Time; a) Specimens with gypsum.......82 Figure 3.1.35 - Helifix screw pull-out: operational phases - a) Cut in half the helices with a pincer; b) Make the pilot hole with a screw drill (diameter 4mm or 3 mm); c) Insert the helix in the mortar with a regolable support; d) Gently applie the support necessary to the extrection; e) Pull out the helix within the appropriate system making slowly 1/3 lap at time............................................................................................................................................85 Figure 3.1.36 - a) Helices on horizontal joints; b) Helices on vertical joints.........................85 Figure 3.1.37 - Helifix screw pull-out Mortar NHL3.5_1:3; Extraction force VS time...........87 Figure 3.1.38 – Helifix screw pull-out Kerakoll mortar M5; Extraction force VS time..........90 Figure 3.2.1 – Material used in experimental campaign 2; a) Kerakoll mortar “Biocalce Muro”; b) Terra Cuita bricks .................................................................................................... 90 Figure 3.2.2 – Sheets of polyethylene around the wallets, according to UNE-EN 1052-1...... 92 Figure 3.2.3 – KERAKOLL mortar M5; Campaign 2; Flexural strength VS Time ................ 94 Figure 3.2.4 – Kerakoll mortar "Biocalce Muro" M5; Campaign 2_Cp. Strength VS time .... 96 Figure 3.2.5 – DPT, gypsum powder on specimens surfaces. a) Gypsum powder on lower punch; b) Gypsum powder on upper specimen’s surface; c) Typical failure mode of a specimen treated with gypsum; d) Frequent failure mode of an untreated specimen. ............. 97 Figure 3.2.6 - Kerakoll mortar M5, DPT: Cp. Strength VS Time; a) Specimens with gypsum; b) Specimens without gypsum; c) Specimens with gypsum powder VS Specimens without gypsum powder. ..................................................................................................................... 100 Figure 3.2.7 – Kerakoll mortar M5, Exp. Campaign 2_Helifix pull-out tests: FEXT VS time a) Holes made with screw drill ϕ=3mm; b) Holes made with screw drill ϕ=4mm; c) Comparison between FEXT VS time, ϕ=3mm VS ϕ=4mm.......................................................................... 106 Figure 3.2.8 - Kerakoll mortar M5, Exp. Campaign 2. Helifix screw pull-out (screw drill ϕ3mm); ................................................................................................................................... 108

VII

List of figures Figure 3.2.9 - Kerakoll mortar M5 "Biocalce Muro", Exp. Campaign 2, Windsor Pin test procedure; a) Pin shooting into the joint; b) Pin partly penetrated into the joint; c) Cleaning hole with the pipette; .............................................................................................................. 110 Figure 3.2.10 – Micrometer of Windsor pin system; ............................................................. 110 Figure 3.2.11 – Kerakoll mortar M5, Experimental Campaign 2; Windsor Pin test; P.depth [mm] VS Time [days] ............................................................................................................ 112 Figure 3.3.1 – Tested Bricks; a) “Terra Cuita” bricks; b) Bon Dia bricks (unbaked bricks); c) “Piera_claro” bricks; d) Bricks from Puig’s house. ............................................................... 114 Figure 3.3.2 – a) Ibertest machine with load cell of 3000 KN; b) Test set-up; c) Terra Cuita failure mode; d) Bon Dia failure mode. ................................................................................. 116 Figure 3.3.3 – a) fc VS diplacement “Terra Cuita” bricks; .................................................... 117 Figure 3.3.4 – Secant modulus in a diagram Stress VS Strain. .............................................. 118 Figure 3.3.5 – fc*/δ VS δ: a) Piera Claro bricks; b) Terra Cuita bricks; c) Bon Dia bricks. . 119 Figure 3.3.6 – Compressive directions; a) Direction Y, perpendicular to the bricks header face; b) Direction X, perpendicular to the brick stretcher; c) Direction Z, perpendicolar to the brick bed. ................................................................................................................................ 120 Figure 3.3.7 – Prior operations on the bricks; a) Faces regularization by grinding machine;..................................................................................................................................121 Figure 3.3.8 – Types of surface treatments; a) Formwork around the specimen; b) Specimens surfaces treated with glue X60; c) Compressive test in the ibertest (load cell 10 KN): gypsum application on the surfaces. .................................................................................................... 122 Figure 3.3.9 - Coring machine “Hilti”; extraction of cylindrical samples ϕ 35mm. .............. 124 Figure 3.3.10 – Cylinders tested; a) Along the direction X, specimens 2:1; b) Along the direction Y, specimens 2:1; c) Specimens 1:1, direction Z, tested in the IBERTEST with load cell of 10 KN. ......................................................................................................................... 124 Figure 3.3.11 – Flexural tests; a) Traction direction Y; b) Traction direction X; c) Support for the test in the IBERTEST (load cell 10KN); d) Failure mode (rupture of the middle section). ................................................................................................................................................ 126 Figure 3.3.12 – Setup details: Brazilian tests of cylindrical samples 2:1 ............................... 128 Figure 3.3.13 – Vertical lines on the header and stretcher faces to ensure the correct tracting direction; a) Specimens 2:1, tensile direction y; b) Specimen failure (traction y); c) Specimens 2:1, tensile direction x; d) Specimen failure (traction x); ...................................................... 129

VIII

List of figures Figure 3.3.14 – a) Setup details: brazilian tests of cylindrical samples 1:1; b) All the tested specimens. .............................................................................................................................. 129 Figure 3.3.15 - Setup details: a) Configuration details of the INSTRON; b) Specimen fixed to the plates by the epoxy resin X60. ......................................................................................... 131 Figure 3.3.16 – Uniaxial tension of brick samples; a) Application of the glue; .................... 132 Figure 3.3.17 – Terra Cuita samples; η VS displacement. ..................................................... 132 Figure 3.3.18 – Helifix screw pull-out tests on the bricks; a) “Terra Cuita” bricks; ............. 134 Figure 3.3.19 – Windsor pin tests on the bricks; a) “Terra Cuita” bricks; b) “Piera Claro” bricks; c) “P y C” bricks. ........................................................................................................ 135 Figure 4.1.1 – Puig i Cadafalch house; a) Before the weather event; b) After the meteorological event (strong wind)........................................................................................ 138 Figure 4.1.2 – Battlements fallen from the terrace and from the roof. ................................... 139 Figure 4.1. 3 – Cleaning of the bricks; a) Removing of the mortar using hammer and chisel; b) Cleaning of bricks faces using a steel brush; c) Pieces of brick before cleaning; d) Pieces after cleaning. ................................................................................................................................. 139 Figure 4.1.4 – Mortar from Puig i Cadafalch house: classification; a) Clear mortar; b) Dark mortar (it probably contains a little percentage of cement). .................................................. 140 Figure 4.1.5 – Specimen preparation; a) Marking of the specimens; b) Cutting of the specimens ............................................................................................................................... 141 Figure 4.1.6 – Double punch test on mortar specimens (Puig i Cadafalch house); a) Specimen between the punches, before the test; b) Central zone who breaks forming an hourglass; c) Putting of gypsum powder on the upper specimen face; d) Failure mode. ............................ 142 Figure 4.1.7 – Compressive tests on cubic specimens; a) Direction Z; b) Cube between the load platens; c) Direction Y; d) Direction X. ......................................................................... 143 Figure 4.1.8 – Helifix screw pull-out tests on the mortar of the wall; a) Ground floor; ........ 147 Figure 4.1.9 – Windsor pin test on the mortars of the walls. ................................................. 148 Figure 5.1.1 – Comparison between compressive strengths averages: fcST,steel VS fcDPT,gy; a) Hydraulic mortar NHL 3,5_1:3; b) Kerakoll mortar M5 “Biocalce muro”; c) Aerial mortar CL90_1:3 “Ciaries”; d) Aerial mortar CL90_1:3 “Segarra” ................................................ 153 Figure 5.1.2 – Comparison between compressive strengths averages: All mortars. .............. 154 Figure 5.1.3 – fc VS fflex_all types of mortar. Prisms made in Steel molds. .......................... 155

List of figures Figure 5.1.4 – Brazilian test; a) Test setup; b) Splitting mechanism characterized by two wedge regions under the punches and a vertical connection crack connecting them ............ 157 Figure 5.1.5 – ζ t VS fflex all types of mortar. Prisms made in Steel molds. ........................... 159 Figure 5.1.6 – fcDPT,gy VS FEXThel; a) Averages made for each joint; b) Averages made for whole wallets. ......................................................................................................................... 161 Figure 5.1.7 – Theory of thick pipe; a) Circular crown with internal and external forces

and

. b) Infinitesimal element subject to the forces’ system.................................................... 163 Figure 5.1.8 – Analytical model for Helifix pull-out; a) Development of radial stresses . ................................................................................................................................. 165 Figure 5.1.9 – fcDPT VS FEXT,hel – Comparison between trendline and analytical curves obtained by theory of thick pipe. a) Analytical curves (assuming ϕ = 30°, 35° and 40°) VS Trendline (averages made for each joint); b) Analytical curves (ϕ = 30°, 35° and 40°) VS Trendline (averages made for whole wallets); c) Analytical curve with ϕ = 35° VS Trendline (averages of each joint, less worse data); d) Analytical curve with ϕ = 35° VS Trendline (averages of whole wallets, less worse data); ........................................................................ 169 Figure 5.1.10 – fcDPT VS FEXT,hel – Analytical curve of thick pipe model: experimental points enclosed in the region [An(ϕ35°)*0,70÷ An(ϕ35°)*1,30] – a) Averages for single joint; b) averages for waal wallets ....................................................................................................... 170 Figure 5.1.11 – fc VS FEXT,hel – Comparison between trendline (averages made for whole wallets) and analytical curves obtained by theory of thick pipe; a) Analytical curves (assuming ϕ = 30°, 35° and 40°) VS Trendline Y; b) Analytical curve (ϕ = 30°) VS Trendline Y (averages made ignoring the worse data); .......................................................................... 171 Figure 5.1.12 – fcDPT VS P.depth– a) Averages made for each joint; b) Averages made for whole wallets; c) Correlation curves (fc VS P.depth) provided by the product data sheets.......................................................................................................................................173 Figure 5.1.13 – fcST VS P.depth – a) Trendline Y; b) Trendline Y compared with the lines made with the tabular data of the instrument. ........................................................................ 175 Figure 5.1.14 - fflex VS FEXT – FEXT averages for whole wallets ............................................ 177 Figure 5.1.15 – fflex VS P.depth – Averages for whole wallets............................................... 178 Figure 5.2.1 – FEXT VS fC,DPT; analytical curve, ϕ=35°..........................................................180 Figure 5.2.2 – FEXT VS P.depth; averages for the mortar of each analyzed hole. ............... 183 Figure 5.2.3 – Mortar exposed: a) Holes at the first floor; b) Test set-up in a hole. .............. 184

List of Tables Table 2.1.1 – Types of natural hydraulic limes ......................................................................... 7 Table 2.1.2 – Types of natural hydraulic limes (EN459-2:2010) ............................................. 7 Table 2.1.3 Types of calcium lime ............................................................................................ 9 Table 2.1.4 – Types of calcium lime (EN 459-2)...................................................................... 9 Table 3.1.1 – Chemical requirements for the lime, expressed as characteristic values_EN 4591 ................................................................................................................................................ 38 Table 3.1.2 – Example of flexural test session: NHL 3.5 prisms. ............................................ 51 Table 3.1.3 – Results of flexural tests on NHL 3,5_1:3 mortars.............................................. 51 Table 3.1.4 – Example of flexural test session on Kerakoll mortar ......................................... 53 Table 3.1.5 – Results obtained from the flexural tests on Kerakoll mortar M5 ....................... 53 Table 3.1.6 – Results obtained from flexure tests on mortar CL90_1:3 “Ciaries” (with powder aerial lime)................................................................................................................................ 55 Table 3.1.7 - Flexural tests on mortar CL90_1:3 “Segarra” (with slaked lime) ...................... 56 Table 3.1.8 – Results obtained from the compressive tests on NHL 3,5_1:3 mortar .............. 58 Table 3.1.9 – Example of compressive test session on KERAKOLL mortar M5 ................... 59 Table 3.1.10 – Results obtained in the compressive tests on Kerakoll mortar “Biocalce muro” M5. ........................................................................................................................................... 60 Table 3.1.11 - Results obtained in the compression tests on mortar CL90_1:3 “Ciaries” ....... 61 Table 3.1.12 - Results obtained in the compressive tests on mortar CL90_1:3 “Segarra”.......63 Table 3.1.13 – Example of brazilian test session on NHL 3,5_1:3 mortar .............................. 65 Table 3.1.14 - NHL 3,5_1:3 mortar; Results obtained from brazilian tests ............................. 66 Table 3.1.15 – Kerakoll “Biocalce Muro” M5; Results obtained from Brazilian tests ............ 66 Table 3.1.16 – CL90_1:3 “Ciaries”; Results obtained from Brazilian tests ............................ 67 Table 3.1.17 – CL90_1:3 “Segarra”; Results obtained from Brazilian tests............................ 67 Table 3.1.18 – Example of DPT test session on Kerakoll mortar M5 (AGE 2) ...................... 73 Table 3.1.19 – NHL 3.5_1:3; Results obtained from DPT on square specimens of mortar. ... 74 Table 3.1.20 – DPT_NHL 3.5_1:3 Ratio of Cp. Strengths averages between specimens with and without gypsum. ................................................................................................................ 77 Table 3.1.21 – Kerakoll mortar M5: DPT on square specimens, all results. ........................... 77 Table 3.1.22 – DPT__Kerakoll mortar M5 Ratio of Cp. Strengths averages between specimens with and without gypsum ....................................................................................... 79 XI

List of tables Table 3.1.23 – CL90_1:3 “Ciaries”: DPT on square specimens, all results. ........................... 80 Table 3.1.24 – CL90_1:3 “Ciaries”: DPT on square specimens, all results. ........................... 82 Table 3.1.25 – DPT_ Ratio of Cp. Strengths averages between specimens with and without gypsum; a) CL90_1:3 “Ciaries”; b) CL90_1:3 “Segarra” ....................................................... 83 Table 3.1.26 – Example of test session on NHL 3.5_1:3 mortar ............................................. 86 Table 3.1.27 – All data, helifix screw pull-out on mortar joints (NHL 3,5_1:3) ..................... 87 Table 3.1.28 – Kerakoll mortar M5: helifix screw pull-out on mortar joints, FEXT (all data).............................................................................................................................89 Table 3.2.1 – Example of flexural test session (Campaign 2).................................................. 93 Table 3.2.2 – All data from flexural tests on Kerakoll mortar M5 .......................................... 93 Table 3.2.3 – Example of compression test data_Kerakoll mortar “Biocalce muro” .............. 95 Table 3.2.4 – Kerakoll mortar, Campaign 2_Compressive strength, all data .......................... 95 Table 3.2.5: Kerakoll mortar “Biocalce Muro”_Example of DPT test session (AGE 4) ........ 98 Table 3.2.6 – Kerakoll mortar M5, DPT, Campaign 2, all data. .............................................. 99 Table 3.2.7 – Kerakoll mortar, Campaign 2; Ratio between compressive strength (with/without gypsum) ........................................................................................................... 101 Table 3.2.8 – Kerakoll mortar M5, Helifix screw-pull-out, Campaign 2, Age 4 ................... 103 Table 3.2.9 – Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out, horizontal joints, all results...................................................................................................................... 104 Table 3.2.10 – Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out ................... 106 Table 3.2.11 – Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out, vertical joints: all results. ............................................................................................................................... 107 Table 3.2.12 – Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out; Comparison between extraction forces of vertical and horizontal joints.................................................... 109 Table 3.2.13 – Kerakoll "Biocalce Muro", Campaign 2, Windsor pin tests; Typical test session on a wallet: at least 15 tests each wallet ................................................................................. 111 Table 3.2.14 – Kerakoll "Biocalce Muro", Campaign 2, Windsor pin test: all data .............. 112 Table 3.2.15 – Comparison between the correlations Pin-in VS Compressive strength provided .................................................................................................................................. 113 Table 3.3.1 – Compressive tests on units: data and averages ................................................ 118 Table 3.3.2 – Compressive tests on units, calculated from the ratio fc*/δ ............................ 119 Table 3.3.3 - Compressive tests of cubic specimens (direction Z): data and averages .......... 123 XII

List of tables Table 3.3.4 – Compressive tests of cubic specimens (directions X and Y): data and averages ................................................................................................................................................ 123 Table 3.3.5 – Compressive tests, cylindrical specimens 2:1, ϕ=35mm (“Terra Cuita” bricks) ................................................................................................................................................ 125 Table 3.3.6 – Compressive tests, cylindrical specimens 1:1, (h=35mm; ϕ=35mm); a) Direction Z; b) Direction Y (perpendicular to header); c) Direction X (perpendicular to stretcher);....125 Table 3.3.7 – Flexural strengths of all the kinds of brick. Direction of traction Y ................ 127 Table 3.3.8 – Flexural strengths of all the kinds of brick. Direction of traction: X ............... 127 Table 3.3.9 – Tensile strengths ζ tx and ζ ty of the ϕ=35mm samples extracted from the bricks; a) Samples 2:1; b) Samples 1:1. ............................................................................................. 130 Table 3.3.10 – Uniaxial tension tests: results ......................................................................... 133 Table 3.3.11 – Helifix screw pull out on bricks: results ........................................................ 135 Table 3.3.12 – Windsor Pin tests on bricks: results. .............................................................. 136 Table 4.1.1 – DPT results (compressive strengths) ................................................................ 143 Table 4.1.2 – Compressive strengths of the cubes (35x35x35mm) ....................................... 144 Table 4.1.3 – Bricks of Puig’s house, compressive tests on prismatic specimens, direction Y ................................................................................................................................................ 144 Table 4.1.4 – Bricks of Puig’s house, compressive tests on prismatic specimens, direction X................................................................................................................................145 Table 4.1.5 – Puig’s house; screw pull-out on the mortar of the battlements ........................ 146 Table 4.1.6 - Helifix screw pull-out on the mortars of the walls ........................................... 147 Table 4.1.7 – Windsor pin test on the mortar of the walls ..................................................... 148 Table 5.1.1 – Comparison between compressive strentghs averages fcST,steel VS fcDPT,gy; a) Hydraulic mortar NHL 3,5_1:3; b) Kerakoll mortar M5 “Biocalce muro”; c) Aerial mortar CL90_1:3 “Ciaries”; d) Aerial mortar CL90_1:3 “Segarra” ................................................. 151 Table 5.1.2 – fcDPTgy VS fcSt kerakoll mortar; a) Campaign 1; b) Campaign 2..................... 154 Table 5.1.3 – fc VS fflex

a) NHL 3.5_1:3; b) Kerakoll M5 Campaign 1; c) CL90_1:3

“Ciaries”; b) CL90_1:3 “Segarra”.......................................................................................... 156 Table 5.1.4 – Ratio fflex /ζ t for all types of mortar.................................................................. 158 Table 5.1.5 – Compressive strengths of the joints VS Helifix screw pull-out: fcDPTgy VS FEXT,hel Experimental Campaign 1__a) Nhl 3,5_1:3 “Cementos Tigre”; b) Kerakoll “Biocalc Muro” M5. .............................................................................................................................. 160 XIII

List of tables Table 5.1.6 – Compressive strengths of the joints VS Helifix screw pull-out: fcDPTgy VS FEXT,hel..................................................................................................................................... 160 Table 5.1.7 – fcDPTgy VS FEXT,hel; averages for whole wallets ................................................ 161 Table 5.1.8 – fcDPT VS FEXT – Relative error εr , Trendline VS Analytical line ϕ35° a) Averages made for each joint; b) Averages made for whole wallet ...................................... 168 Table 5.1.9 – fcDPTgy VS FEXT; averages for whole wallets.................................................... 170 Table 5.1.10 – fc VS FEXT – Relative error εr , Trendline VS Analytical line a) Double Punch Test VS Helifix screw pull-out, analytical line with ϕ=35°; b) Standard compression test VS Helifix screw pull-out, analytical line with ϕ=30°; ................................................................ 172 Table 5.1.11 – fcDPT VS P.depth – Averages for each mortar joint ........................................ 172 Table 5.1.12 – fcDPT VS P.depth – Averages for whole wallets ............................................. 172 Table 5.1.13 – fcDPT VS P.depth – Averages for whole wallets............................................. 175 Table 5.1.14 – fflex VS FEXT – FEXT averages for whole wallets ............................................. 176 Table 5.1.15 – fflex VS P.depth – P.depth averages for whole wallets ................................... 177 Table 5.2.1 – Parameters for analtycal curve (equation 5.1.22, see figure 5.2.1) .................. 180 Table 5.2.2 – Parameters for analytical curve (equation 5.1.22), corrected to describe the mortar joints of Puig i Cadafalch’s house .............................................................................. 181 Table 5.2.3 – Comparison between cp. strengths estimated and detected. ............................ 181 Table 5.2.4 – FEXT VS P.depth; averages for the mortar of each analyzed hole. ................... 182

XIV

1. Introduction 1.1 Motivation of the research Compressive strength is one of the most important mechanical properties in masonry buildings. Testing and determining compressive strength in masonry building are so complicated because of several reasons and sometimes it is not possible. In most of the time removing and carrying some parts of masonry building to lab is not practical. Also historical buildings usually has been built making use of several materials and completed in long time. So each building part could have different mechanical characterizations. Nowadays minor and non-destructive tests can produce quick data about mechanical characterization especially compressive strength. This characterization can not be considered 100% accurate, but it gives an order of magnitude of the mechanical properties. It would be important by careful and well-planned researches to improve the correlations votes to derive the compressive strength of mortar and bricks from data provided by noninvasive testing. Also understanding the physical phenomenas at the base of each noninvasive test it should be possible to obtain by simple, but effective analytical formulas, the foundamental properties such as the compressive strength. Windsor pin penetrometer is an instrument designed to determine compressive strength of materials like mortars, concrete, bricks, and blocks in situ. For the tests described in this thesis Windsor pin system WP-2000 from NDT JAMES INSTRUMENT Company has been used. This instrument is mostly made for testing on cement mortars and concrete. The spring of this instrument produce 108NM forces to test compressive strength of concrete to a maximum of 5300 PSI (36.9MPa). Obviously compressive strength of historical mortars is pretty lower than new mortars (cement base). So the compressive strength table of this instrument, provided from the company, is not suitable for historical mortars. An aim is to give an interpretating method for weak mortars. A lot of tests will be made on different types of mortar measuring the pin-penetration depth and on the same mortars direct destructive tests will be made to find the correlations (P.depth VS Compressive strength and P.depth VS Flexural strength). The helifix screw-pull-out is a very simple test. It is foundamentally a shear strength measuring of a small cylinder or annulus of the material which is enganged by a helical self1

1 – Introduction tapped screw (Vekey, 1997). The methallic helix is previously screwed into a small hole made by a drill. The screw enganges a little cylindrical crown of the tested material, than it is ectracted. During the extraction the screw only translates, and the maximum pull-out force at the external surface of the cylindrical crown in measured. Until now it has been done few studies (for example Ferguson, 1994 and Vekey, 1997) and the correlating curves for mortars and concrete don’t give satisfactory and fully acceptable results, especially for weak historical mortars. A lot of screw-pull-out tests will be made on mortars and bricks in this research. The same material will be tested by direct destructive tests and the correlation curves (Pull-out force VS Compressive strength and Pull-out force VS Flexural strength) will be gived.

1.2 Aims and Methodology To find correlation between compressive strength and instruments’ reading (Windsor and Heifix) probably the only way is to perform non-invasive tests and destructive tests on specimens with same material and properties. Micrometer data and pull-out force measured by the Helifix Load Test Unit cannot directly give us compressive strength of material unlike compressive strength test. Compressive strength is equal to force divided by area. In case of mortars finding correlation between compressive strength and micrometer reading data is more complicated. As we know mortars harden over time. Hardening process of some mortars, like aerial lime mortars, continues over many years even more than one century. So different kinds of mortars depend on the age and material will have different compressive strength. In the case of ancient mortars we cannot find two mortars with same chemical and mechanical properties. It is also very interesting to perform penetrometric and compressive tests in order to follow the hardening process. Lime mortars are the most frequently used in the historical buildings, so this thesis focused on them. During the first laboratory program, 16 wallets were built, 4 wallets for each type of mortar, with “Terra Cuita” handmade bricks. The types of mortar were:  A hydraulic lime mortar with ratio lime:sand = 1:3;  A prefabricated mixture based on hydraulic lime mortar with additives which improve the performance. The mixture was provided by the Kerakoll Company, and for this type of mortar higher resistance is expected than the other mortars;  An aerial lime mortar with powder lime CL90, with ratio lime:sand = 1:3;  An aerial lime mortar with slaked lime CL90, with ratio lime:sand = 1:3.

1 – Introduction Furthermore according to European standard EN-196-1, 6 prismatic specimens were made for each wall with the same mortars of the joints. 3 prisms were made in steel molds and 3 prisms in handmade wooden molds, to investigate how the wood absorption affects on the mortars’ hardening process. The minor-destructive tests were made on the mortar joints and on the bricks. From the same joints, square mortar specimens were cutted and compressed. The data from flexural tests and compressive tests provide other informations about the hardening process and the reached resistance. In fact each wall was tested at a different age of maturation. AIMS  Determining compressive strength of historical mortars by Windsor pin penetrometer test and Helifix Load Test Unit;  Calibration of compressive strength table of the two instruments, for lime mortars;  Improve all the test methods;  Develop analytical models capable to providing the compressive strength from the data of non-invasive tests;  Verify if the obtained results are also valid for an existing building. In addition a second experimental program in laboratory was conduced, to complete the studies. This time 5 wallets were built, but only with kerakoll mortar and three prisms molded in steel were made for each wallet. Finally a test program was made in the house of Puig i Cadafalch (Dolores Monserdà Street, 3-5, Argentona, Barcelona).

1.3 Outline of the thesis The present thesis is divided into six chapters. The first chapter deals with the objectives and focus of the thesis. The second chapter regards the state of the art. Some basic information about the story and the types and properties of mortar and bricks are given. The third chapter deals with the laboratory experimental program, along with the research results on bricks, mortar joints and mortar prisms. In chapter four the results of in-situ experimental program on brick and mortar joints are resumed.

1 â&#x20AC;&#x201C; Introduction In chapter five all the results exposed in the chapters 3 and 4 are analysed. The experimental curves are compared with the analytical curves, and the different testing methods are compared. Chapter six concludes the thesis with the summary and main outcomes of the present research and suggestions for the future works.

2. State of the art

2.1 Overview on Lime Mortars Mortar is a workable paste used to bind masonry blocks together and fill the gaps between them (M. Como, 2013). It becomes hard when it sets and it gains stiffness and resistance over the time, resulting in a rigid aggregate structure. The functions carried out by mortar in the masonry, are mainly three (Martinez et al., 2001): fill the joints, avoiding the passage of water; regularize the disposition of bricks and uniformly distribute the load; cooperate to lead horizontal stresses until foundations. The common materials of masonry construction are brick, stone, marble, granite, travertine, limestone, cast stone, concrete block, earth, glass block, stucco, and tile. Mortars typically are composed of binder, aggregates, water and mixture. Mortar is used for different applications. The properties and characteristics of the mortars mainly depend on the binder. We can mention bitumen, gypsum, clay, lime, cement and etc. as a binder. Admixtures materials (natural or artificial) have been added to mortar for avoiding of shrinkage, crack and for increasing total strength. Different materials like blood, egg, fig juice, pig grease, manure and straw have been used as admixture in different country and periods.

2.1.1 History of lime mortar Egyptian used lime for building constructions in 2600â&#x20AC;&#x201C;2500 BC. Hydraulic lime mortar seems to have been appeared in Mediterranean countries under Roman influence, Deposits of volcanic ash and Pozzuoli materials used with lime from early time for Roman concrete in constructions. Hydraulic lime is produced in higher temperature than aerial lime and maybe ancient people before romans couldnâ&#x20AC;&#x2122;t make enough temperature to producing hydraulic lime.

2.1.2 Types of lime: hydraulic lime and aerial lime Building lime divided to two big groups, hydraulic lime and non-hydraulic lime (aerial lime). Sub-families and forms of hydraulic and aerial limes are given in figure 10 respectively according to EN-459, the European standard for Building Lime (Figure 2.1.1).

2 â&#x20AC;&#x201C; State of the art

Figure 2.1.1 â&#x20AC;&#x201C; Building lime family chart according to EN-459-1

HIDRAULIC LIME Hydraulic lime was the principal binder for mortar up to the mid 1800â&#x20AC;&#x2122;s when Portland cement was developed as a product. Although relatively weak and slow in setting and developing strength, when compared to cement based mortars, mortars produced with hydraulic lime were suitable for the relatively thick walls and lower stresses that generally characterized the more massive masonry construction of former times. All hydraulic limes have the property of setting and hardening under water but atmospheric carbon dioxide contributes to the hardening process in the longer term.

2 â&#x20AC;&#x201C; State of the art Three grades of hydraulic lime are identified in EN459 the European standard for Building Lime: -

Natural hydraulic lime (NHL): this is produced by burning more or less argillaceous or siliceous limestone and then reducing it to a powder by slaking with or without grinding; It has the property of setting and hardening when mixed with water and by reaction with carbon dioxide from the air (carbonation). The hydraulic properties exclusively result from the special chemical composition of the natural raw material. Grinding agents up to 0,1 % are allowed. Natural hydraulic lime does not contain any other additions (EN 459-1). Natural hydraulic lime shall be classified according to the notation given in Table 2.1.1 and 2.1.2. Table 2.1.1 â&#x20AC;&#x201C; Types of natural hydraulic limes

Table 2.1.2 â&#x20AC;&#x201C; Types of natural hydraulic limes (EN459-2:2010)

Formulated lime (FL): it consists of air lime and/or natural hydraulic lime with added hydraulic or pozzolanic material. It has the property of setting and hardening when

2 – State of the art mixed with water and by reaction with carbon dioxide from the air (carbonation) (EN 459-1.2). -

Hydraulic lime (HL): this is a binder consisting of lime and other materials such as cement, blast furnace slag, limestone filler and other suitable materials. It has the property of setting and hardening under water. Atmospheric carbon dioxide contributes to the hardening process (EN 459-1.2).

AERIAL LIME Non-hydraulic lime or aerial lime is the principal binder of most traditional mortars and plasters. Air lime is used for the preparation or the production of materials used in building construction as well as in civil engineering. Air lime when appropriately batched and mixed with water, forms a paste that improves the workability (values of flow and penetration) and water retention of mortars. The carbonation of hydrates in contact with atmospheric carbon dioxide forms calcium carbonate which develops strength and contributes to the durability of mortars containing building lime (hence the name of air lime). Aerial limes forms divide to quicklime and hydrated lime. -

Quicklime (Q) is an air lime mainly in the oxide form which reacts exothermically on contact with water. Quicklime is available in a range of sizes from lump to powder.

Hydrated lime (S, S PL or S ML) is an air lime mainly in the hydroxide form produced by the controlled slaking of quicklime. Hydrated lime is available as:  powder (S);  putty (S and PL);  slurry or milk of lime (S ML);

Aerial limes chemical compositions divide to calcium lime (CL) and dolomitic lime (DL). I.

Calcium lime (CL) Calcium lime shall be classified according to the notation given in Table 2.1.3 and its total (CaO + MgO) content in accordance with Table 2.1.4.

2 – State of the art Table 2.1.3 Types of calcium lime

The properties of the type of calcium lime shown in Table 2.1.3 determined in accordance with EN 459-2 shall conform to the requirements in that table. Calcium lime is an air lime consisting mainly of calcium oxide and/or calcium hydroxide without any hydraulic or pozzolanic addition. The chimical compositions of the types of calcium lime are shown in Table 2.1.4. Table 2.1.4 – Types of calcium lime (EN 459-2)

II.

Dolomitic lime (DL)

Dolomitic lime is an air lime consisting mainly of calcium magnesium oxide and/or calcium magnesium hydroxide without any hydraulic or pozzolanic addition. (EN 4591,2)

2.1.4 Mechanical characterization and laboratory testing of mortars The main tests on mortars (M. Bošnjak-Klečina, S. Lozančić, 2010) are the following:  Mortar flexural-test: The flexural strength is evaluated according to the EN 1015-11:2007. The flexural strength

[

] is:

2 – State of the art

Where

[ ], ,

 Mortar compressive-test: The compressive strength is evaluated directly by the compression test of the two resulting parts of mortar prisms, after the flexural test according to the EN 101511:2007. The compressive strength,

[

] is calculated like:

Where: F = maximum load [N]; 402 = load platens’area [mm2]. This test will be better explained in chapter 3.  Double Punch Test: Some methods have been proposed in the literature for estimating mortar mechanical propertyes, by making use of small, non-standard sample (Drdacky et al., 2008, Drdacky, 2011). In particular, one of the most used methods is the double punch test. This test was proposed by Henzel and Karl, 1987. In the same investigation the authors found the optimal diameter for the two punches, equal to 20mm. Investigation on the influence of mortar quality, mortar porosity, mortar curing and confining effect of mortar surrounding the loaded are were reported in the literature (Henzel and Karl, 1987; Pelà et al., 2012; Sassoni and Mazzotti, 2013). The test and the typical curve obtained are shown in Figure 2.1.2.

Figure 2.1.2 – View (a and results (b of the double punch test as a function of the thickness.

2 – State of the art Exponentially decreasing strength is obtained for increasing ratios. McNary and Abrams (1985) had highlighted that the compressive strength increment with the confining effect. Therefore the behaviour shown in Figure 2.1.2.b has a possible explanation in the confinement effect.

2.2 Bricks In this section the bricks mechanical and chemical properties are exposed. The thesis aims to analyze the properties of the single brick. The main aim is to give a method to better interpret the data of minor destructive instruments, Windsor Pin System and Helifix Screw Pull-out. The tolls locally affect the materials, so it is intersting to point out the micro-properties of mortars and bricks.

2.2.1 History and origin The appearance of the bricks takes place during the transition between nomadic and sedentary life, because of the humans’ need to settle and form a community. This period is known as Neolithic, and it is dated between 10000 and 8000 BC. In this period the construction of the first city is detected. The oldest brick was found in Jericho (1952), during an excavation. Jericho is considered the first place in the world were the sedentary life started. Near the Jordan River the ruins of “Tell el Sultan” are situated and it is the first human settlement in the world, with a system of urban fortification. For these reasons, Jerico has always been considered the cornerstone of civilization and human development. However, the ceramic appears later (4000 BC), in the first villages. The bricks were made in raw form, dried under the sun. These kinds of brick consist of sand, clay, water and sometimes fiber or organic materials such as straw, twigs or excrement. The Sumerian, inhabitants of the ancient region of Mesopotamia, introduced the baked bricks. They understood that rainwater attack the unfired bricks. Today a great buildig made with the unfired bricks “adobe” is steel present in Iràn. It is the famous Zigurat (Figure 2.2.1.a).

2 â&#x20AC;&#x201C; State of the art

Figure 2.21 - a) Zigurat Al-Untash (IrĂĄn); b) Temples of Bagan (Birman).

2.2.2 The brick making process The fundamentals of brick manufacturing have not changed over time. However, technological advancements have made contemporary brick plants substantially more efficient and have improved the overall quality of the products. A more complete knowledge of raw materials and their properties, better control of firing, improved kiln designs and more advanced mechanization have all contributed to advancing the brick industry. TYPES OF CLAY Clays occur in three principal forms, all of which have similar chemical compositions but different physical characteristics. -

Surface Clays: surface clays may be the upthrusts of older deposits or of more recent sedimentary formations. As the name implies, they are found near the earth surface.

Shales: shales are clays that have been subjected to high pressures until they have nearly hardened into slate.

Fire Clays: fire clays are usually mined at deeper levels than other clays and have refractory qualities. Surface and fire clays have a different physical structure from shales but are similar in chemical composition.

All three types of clay are composed of silica and alumina with varying amounts of metallic oxides. Metallic oxides act as fluxes promoting fusion of the particles at lower temperatures. Metallic oxides (particularly those of iron, magnesium and calcium) influence the color of the fired brick. The manufacturer minimizes variations in chemical composition and physical properties by mixing clays from different sources and different locations in the pit. Chemical composition 12

2 â&#x20AC;&#x201C; State of the art varies within the pit, and the differences are compensated for by varying manufacturing processes. As a result, brick from the same manufacturer will have slightly different properties in subsequent production runs. Further, brick from different manufacturers that have the same appearance may differ in other properties. PHASES OF MANUFACTURING The manufacturing process has six general phases (Brick Industry Association, 2006) (Figure 2.2.2): 1) Mining and storage of raw materials; 2) Preparing raw materials; 3) Forming the brick; 4) Drying; 5) Hacking; 6) Firing and cooling; 7) De-hacking and storing finished products.

Figure 2.2.2 - Diagrammatic Representation of Manufacturing Process

1) Mining and storage of raw materials Surface clays, shales and some fire clays are mined in open pits with power equipment. Then the clay or shale mixtures are transported to plant storage areas (Figure 2.2.3).

2 â&#x20AC;&#x201C; State of the art

Figure 2.2.3 â&#x20AC;&#x201C; Clay or shale being crushed and transported to storage area.

Continuous brick production regardless of weather conditions is ensured by storing sufficient quantities of raw materials required for many days of plant operation. Normally, several storage areas (one for each source) are used to facilitate blending of the clays. Blending produces more uniform raw materials, helps control color and allows raw material control for manufacturing a certain brick body. 2) Preparing raw materials To break up large clay lumps and stones, the material is processed through size-reduction machines before mixing the raw material. Usually the material is processed through inclined vibrating screens to control particle size. 3) Forming the brick Tempering, the first step in the forming process, produces a homogeneous, plastic clay mass. Usually, this is achieved by adding water to the clay in a pug mill (Figure 2.2.4).

Figure 2.2.4 â&#x20AC;&#x201C; Clay is thoroughly mixed with waterin pug mill before extrusion.

2 – State of the art -

Stiff-Mud Process: In the stiff-mud or extrusion process (Figure 2.2.5) water in the range of 10 to 15 percent is mixed into the clay to produce plasticity. After pugging, the tempered clay goes through a de-airing chamber that maintains a vacuum of 15 to 29 in (375 to 725 mm of mercury). De-airing removes air holes and bubbles, giving the clay increased workability and plasticity, resulting in greater strength. Next, the clay is extruded through a die to produce a column of clay. As the clay column leaves the die, textures or surface coatings may be applied. An automatic cutter then slices through the clay column to create the individual brick. Cutter spacings and die sizes must be carefully calculated to compensate for normal shrinkage that occurs during drying and firing.

Figure 2.2.5 – After mining, clay is extruded through a die and trimmed to specified dimension before firing

Soft-Mud Process: The soft-mud or molded process is particularly suitable for clays containing too much water to be extruded by the stiff-mud process. Clays are mixed to contain 20 to 30 percent water and then formed into brick in molds. To prevent clay from sticking, the molds are lubricated with either sand or water to produce “sandstruck” or “water-struck” brick. Brick may be produced in this manner by machine or by hand.

Dry-Press Process: This process is particularly suited to clays of very low plasticity. Clay is mixed with a minimal amount of water (up to 10 percent), then pressed into steel molds under pressures from 3.4 to 10.3 MPa by hydraulic or compressed air rams.

2 – State of the art 4) Drying: Wet brick from molding or cutting machines contain 7 to 30 percent moisture, depending upon the forming method. Before the firing process begins, most of this water is evaporated in dryer chambers at temperatures ranging from about 38 ºC to 204 ºC. The extent of drying time, which varies with different clays, usually is between 24 to 48 hours. Although heat may be generated specifically for dryer chambers, it usually is supplied from the exhaust heat of kilns to maximize thermal efficiency. In all cases, heat and humidity must be carefully regulated to avoid cracking in the brick. 5) Hacking: Hacking is the process of loading a kiln car or kiln with brick. The number of brick on the kiln car is determined by kiln size. The brick are typically placed by robots or mechanical means. The setting pattern has some influence on appearance. Brick placed faceto-face will have a more uniform color than brick that are cross-set or placed face-to-back. 6) Firing and cooling -

Firing: Brick are fired between 10 and 40 hours, depending upon kiln type and other variables. There are several types of kilns used by manufacturers. The most common type is a tunnel kiln, followed by periodic kilns. Fuel may be natural gas, coal, sawdust, and methane gas from landfills or a combination of these fuels. In a tunnel kiln (Figure 2.2.6) brick are loaded onto kiln cars, which pass through various temperature zones as they travel through the tunnel. The heat conditions in each zone are carefully controlled, and the kiln is continuously operated. A periodic kiln is one that is loaded, fired, allowed to cool and unloaded, after which the same steps are repeated. Dried brick are set in periodic kilns according to a prescribed pattern that permits circulation of hot kiln gases.

Figure 2.2.6 – Brick Enter Tunnel Kiln for Firing

2 – State of the art Firing may be divided into five general stages: 1) final drying (evaporating free water); 2) dehydration; 3) oxidation; 4) vitrification; and 5) flashing or reduction firing. All except flashing are associated with rising temperatures in the kiln. Although the actual temperatures will differ with clay or shale, final drying takes place at temperatures up to about 205 ºC, dehydration from about 150 ºC to 980 ºC, oxidation from 540 ºC to 980 ºC, and vitrification from 870 ºC to 1320 ºC. -

Cooling: after the temperature has peaked and is maintained for a prescribed time, the cooling process begins. Cooling time rarely exceeds 10 hours for tunnel kilns and from 5 to 24 hours in periodic kilns. Cooling is an important stage in brick manufacturing because the rate of cooling has a direct effect on color.

De-hacking and storing finished products: De-hacking is the process of unloading a kiln or kiln car after the brick have cooled, a job often performed by robots. Brick are sorted, graded and packaged. Then they are placed in a storage yard or loaded onto rail cars or trucks for delivery. The majority of brick today are packaged in self-contained, strapped cubes, which can be broken down into individual strapped packages for ease of handling on the jobsite. The packages and cubes are configured to provide openings for handling by forklifts.

2.2.3 Bricks: general properties Bricks are blocks of a ceramic material that are produced in common or standard sizes. There are a large variety of bricks, in relation with the material (bricks might be made from clay, lime and sand, concrete, stone, etc.), the type (solid, hollow, perforated, etc.), dimensions, as well as the mechanical properties. As regards the dimensions of bricks, they vary widely depending on the age and the production areas, in function of the available raw materials and technology of the time. Furthermore the sizes of the bricks are completely changed in consequence of the evolution of industrial technology, therefore of the cooking process. Although the size of bricks may vary, the proportions between the three dimensions are usually constants: the longest side is about twice of the shorter side, and the shorter side is not less than twice the height. An optimum brick should be free from any impurities, it should present fine and uniform grained, plane faces, free of cracks but with a certain roughness; it should not contain more 17

2 – State of the art than 0.05% of sulfuric anhydride and it must resist to sea water and atmospheric agents without delamination (Gambarotta, 2001). The characterization of old clay bricks is a difficult task, due to the high variability in production and additional variability caused by deterioration from the weather or chemical agents, such as soluble salts, freeze–thawing cycles or load–unload cycles. Moreover, clay bricks in a given structural element or building can belong to different construction periods or productions. Finally, the experimental test set-up conditions (dimensions and moisture content of the sample, boundary conditions, temperature, etc.) can also influence the results (Fernandes et al., 2009). Nevertheless, the analysis of the bricks properties is fundamental both to understand the mechanisms of damage and evaluate the safety, both to make decisions on reusing and replacing materials. A great quantity of studies have demonstrated that ancient materials generally show low characteristics in comparison with the modern ones, such as high porosity and absorption, low compressive strength, and elastic modulus (Lourenço et al., 2010). Is thus a need the knowledge of the physical, chemical and mechanical properties of bricks in order to avoid the use of unsuitable new material to replacing the old one. The mechanical properties of bricks, in particular the compressive strength, the modulus of elasticity and the Poisson’s ratio, are very relevant for the structural behaviour of historical constructions, as these are the main influence factors on the compressive strength of masonry. The compressive strength of bricks (fc) is strongly influenced by the characteristics of the raw material and by the production process. It is known that the raw clay of old bricks were often of low quality and the manufacturing process was relatively primitive and inefficient. This parameter is characterized by large variability and the range of values found in the literature is quite wide (about 1.5-32 MPa), meaning that in situ testing or destructive testing of samples must be carried out when the compressive strength of the brick is required. Other characteristics of existing old bricks can provide an indication about compressive strength, such as mineral composition, texture, crack pattern and porosity level, by revealing the conditions of drying and firing. The modulus of elasticity (E) is frequently found in the literature and is also characterized by large variability. It is not always clear how authors measured the values presented, even if most standards refer the use of the linear part of the stress–strain curve in a range of 30–50% 18

2 – State of the art of the maximum stress value. The values found range from 1 to 18 GPa, which represents a range between 125 and 1,400 fc, where fc is the compressive strength. Most common values are in the range of 200 fc, with an average value of 350 fc (Fernandes et al., 2009). The Poisson’s ratio (ν) influences the relative deformability and appears in certain analytical models of the behaviour of masonry. In the literature, values from 0.15 to 0.20 for the stone and from 0.10 to 0.15 for brick can be found (Martínez et al. 2001). This study analyzes solid clay brick, since this is the typology most used in historic buildings, because it was easy to produce, lighter than stone and formed a wall that was fire resistant and durable.

2.2.4 Mechanical characterization and laboratory testing of bricks The main tests on brick units (M. Bošnjak-Klečina, S. Lozančić, 2010) are the following:  Determination of dimensions: The bricks’ dimensions (figure 2.2.7) (length, width and height) must be determinate according to EN772-16.

Figure 2.2.7 – Bricks dimensions.

 Initial capillary water absorption coefficient: The initial capillary water absorption coefficient of the brick expressed in kg/m2·min is made according to EN772-11.  Water absorption: The water absorption is determined by the procedure prescribed by the norm EN772-13.  The percentage of pores: The percentage of pores in brick elements is determined by the procedure prescribed by the norm EN772-3. The percentage of pores portion of the brick element is the quotient of the difference between gross and net volume and gross volume. 19

2 – State of the art

 Compressive test: The compressive strength of brick elements is tested following the procedure described in EN772-1. Some examples of this kind of test are indicated in chapter 3.  Flexural test: a standard which give the rules for this kind of test on bricks doesn’t exists. But it can be make following the standard EN 772-6 for mortar prisms.

2.3 In-situ minor destructive testing of mortar and bricks 2.3.1 Windsor pin Penetromer system Penetrometer test is a Minor destructive testing (MDT) method to quickly estimate compressive strength of concrete or mortar. The Windsor pin system WP-2000 (figure 2.3.1) shoots a steel pin into the surface of concrete or mortar and the depth of penetration can be measured. The spring is loaded by tightening the reaction nut at the top of the instrument until the trigger mechanism latch closes to hold the spring in place. The stored potential energy is 91 Lbs (108Nm). With the spring loaded it is compressed to a distance of 0.8 inches. Once the trigger is pulled there is enough force to test compressive strength of concrete to a maximum of 5300 PSI (36.9MPa). The device drives a steel pin into the surface of the material and measures the compressive strength by measuring overall penetration. The tables of the data sheets of the instrument provide an estimation of the compressive strength of the material, depending on the pin penetration depth. There is a Concrete Strength Table and a Mortar Strength Table. Main applications of Windsor pin system according to factory catalogue: 1. Test new concrete products and structures for early strength; 2. Evaluate the in situ strength in existing structures, e.g., after suspected fire damage; 3. Test strength of block, brick and mortar joints within an existing structure, e.g., load bearing walls; 4. Test polymer concrete and patching compound; 5. Quality control of precast elements such as block, brick, slabs and pipe;

2 â&#x20AC;&#x201C; State of the art

Figure 2.3.1 â&#x20AC;&#x201C; Windsor pin system WP-2000

Determination of pin reusability The pin is made of high strength steel specifically for building material penetration. After each shot be sure to check the length of the steel pin to verify if it is reusable (Figure 2.3.2). To do this, use the provided go/no-go gauge. If the pin can easily pass through the slot on the gauge we should discard the pin and get new one.

Figure 2.3.2 â&#x20AC;&#x201C; Determination of pin reusability; a) Two possible pin configuration after the test; b) Determination of pin reusability, passing the pin through the slot of the gauge.

Performation of a Pin penetration test: all the phases 1. Insert a pin into the removable ring or V-shaped barrel. 2. Using the wrench tighten the loading nut until the trigger mechanism latch closes to hold the spring in place. 3. Back off the loading nut completely off the top of the load screw before pulling trigger.

2 â&#x20AC;&#x201C; State of the art 4. Place the instrument on a smooth flat surface of the material to be tested. If necessary, use a grindstone to prepare the surface. 5. Place the instrument perpendicular to the test surface and pull the trigger (figure 2.3.3). The instrument should be held firmly against the surface. Particularly when testing vertical walls and ceiling.

Figure 2.3.3 â&#x20AC;&#x201C; Correct Style to performing penetrometer test.

6. Remove the instrument and with the rubber bulb-type blower (figure 2.3.4), clean out the small hole made in the material surface.

Figure 2.3.4 â&#x20AC;&#x201C; Cleaning of the hole by the rubber bulb-type blower.

7. Place the micrometer over the hole, making sure that the reference surface of the micrometer is flat on the material. For measuring mortar joints, the micrometer utilizes a V-shaped barrel similar to the pin drive. 8. Insert micrometer probe to the bottom of the hole using the knurled thimble on the head of the micrometer.

2 â&#x20AC;&#x201C; State of the art 9. Read and note the micrometer reading. To find the penetration, subtract the measuring reading from 1 inch. Micrometer reading 1. Hold the micrometer over the hole by grasping the grip. The grip is the first cylinder of the micrometer which the ring attaches to and has no numbers etched into it. 2. Turn the thimble at the very top of the micrometer above the grip and the sleeve with numbers marked horizontally on its surface. Depending on which way the thimble is turned, the spindle will extend or retract (Figure 2.3.5).

Figure 2.3.5 â&#x20AC;&#x201C; Measuring hole depth; a) Phase 1; b) Phase 2.

3. Turn the thimble until the spindle is extended to the bottom (figure 2.3.5.b). 4. Write down the number (Figure 2.3.6.a9 on the sleeve that the bottom of the thimble (Figure 2.3.6.b) is lined up.

Figure 2.3.6 â&#x20AC;&#x201C; a) Circumstance of reading micrometer-Manual of Windsor pin system; b) Micrometer.

2 â&#x20AC;&#x201C; State of the art 5. Subtract that measurement from 1 inch to get the actual reading.

2.3.2 Helifix screw pull-out system Another important instrument to perform in-situ minor-destructive tests on mortar, concrete or bricks is the helifix screw pull-out system. This is a very simple test. It is fundamentally a measurement of the shear strength of a small cylinder or annulus of the material which is hooked by a helical self-tapped screw (Vekey, 1997). Figure 2.3.7 shows a typical specimen and helix. As with all similar proxy tests the pull-out resistance has to be calibrated against some standard accepted strength measure of the material.

Figure 2.3.7 â&#x20AC;&#x201C; Installation of helical tie for screw pull-out test.

Scope According to (RILEM TC 127-MS 1997), the technique can give useful data on the following charatcteristics: 1) Batch-to-batch variability of strength or general quality; 2) Variation of quality in relation to a reference sample; 3) Changes of properties with time. Strength increases due to hardening and the effects of weather conditions and additives; Thus, on building sites the method has application both for quality control purposes and as a mean for troubleshooting. In the laboratory, it is useful for control of hardness development during curing and hardening, or as a monitoring/control parameter for other tests such as acid rain and freeze â&#x20AC;&#x201C; thaw tests. The method is limited by the yield strength of the helical steel screw used to a maximum strength of approximately 10 N/mm2. Above this strength, the failure is by yield of the steel and the test value is not longer proportional to the strength of the mortar. Thus, for stronger mortars the technique is only suitable for proof testing. The method is not suitable for mortar beds less than 8 mm thick.

2 â&#x20AC;&#x201C; State of the art Procedure 1. Randomly select position within an area not exceeding 2 m2, and at each position drill a 4 mm in diameter hole in the middle of the thickness of the mortar bed (figure 2.3.8). In low density materials, the helix can be driven straight in with no preparation. In mortars and medium density materials a pilot hole is drilled so that only the outer periphery of the helix is enganged (Vekey 1991).

Figure 2.3.8 â&#x20AC;&#x201C; Selection of the points and drilling holes.

2. Mount a helical tie into the driving tool and then, holding horizontal the sleeve of the tool, push the exposed end of the tie gently into the pilot hole. 3. Hammer the helical tie firmly, but not violently, into the hole so that the specified length of its thread (L) is embedded in the mortar using the sleeved driving tool (figure 2.3.9).

Figure 2.3.9 â&#x20AC;&#x201C; Installation of the tie.

2 â&#x20AC;&#x201C; State of the art 4. After installation, a gripper is then screwed onto the end of the tie. This holds the tie fixed during the test, restraining it from rotating and ensures a shear type failure in the test material (figure 2.3.10).

Figure 2.3.10 â&#x20AC;&#x201C;Gripping device attaching.

5. The proof loading device is then attached to the gripper and the assembly is rotated to screw down the tie and take up any slack (figure 2.3.11). The tie is loaded until failure. The load applied to the tie should be increased steadily. The peak load reached during a test is held by the read needle on the dial, and is recorded as the pull-out load.

Figure 2.3.11 â&#x20AC;&#x201C; Load testing the tie.

Test Results Where the data is used to indicate monitor quality variation, to follow changes due to hardening or cyclic actions in durability testing, it is sufficient to calculate the mean of the ten (or more) measurements. If absolute indications of equivalent strength properties such as cube strength, flexural strength or tensile strength are required, the data must be transformed using a suitable 26

2 – State of the art calibration curve. Previous calibration trials (figure 2.3.12) indicate that the relationship between pull-out force and strength properties is not linear (Vekey, 1991), thus each individual measurement must be transformed before mean is calculated.

Figure 2.3.12 – Vekey 1991 “In –situ tests for maronry”; a) Calibration of screw-pull-out for mortars; b) Calibration of screw-pull-out for UK AAC blocks.

Figure 2.3.12.a shows the relationship between mortar cube strength and pull-out force for a 6mm in diameter helix for a range of mortars. This indicates that the variability is high but the relationship is quite consistent and is not particularly sensitive for the weaker mortars where the accuracy is more critical. The slight flattening out of the curve for higher strength mortars is probably largely due to the plastic failure of the helix but for such high strength mortars it is rarely critical to precisely measure strength. Figure 2.3.12.b indicates AAC over a limited strength range. This technique has been used on a number of field investigations of understrength and degraded mortar and the derived data has generally well correlated with data from other tests. Ferguson, 1994: program and procedure of his research Another important research was performed by W.A.Ferguson and J.Skandamoorthy (Ferguson, 1994). The work was made as following. A wide range of test materials have been tested:  Autoclaved aerated concrete (AAC) blocks;  Lightweight, medium density and high density concrete blocks;  A range of brick types and strengths, and a wide range of mortar mixes; 27

2 â&#x20AC;&#x201C; State of the art The results showed that the method appeared to be suitable for testing materials of strengths ranging up to 10 MPa; with stronger materials, the method did not provide satisfactory results. Essentially there were two problems with stronger materials: 1. Firstly, the failure mechanism differed from that found in the weaker units. Rather than failing in shear, the material failed in compression under the thread of the tie; 2. Secondly, at higher pull-out loads, it was found that the gripper used to hold the tie and stop it turning. On site, the pull-out tests are carried out using a standard proof loading device; for the calibration tests, though, all testing was carried out in a universal testing machine. Whatever the means of testing, however, a gripper device is attached to the tie and the load applied through it; the gripper secures the helix and prevents it from turning, and unscrewing from the test material, during test. With the exception of AAC units, where the tie was driven directly into the test material, the same test procedure was followed for all materials tested: holes were drilled into the test material 5 mm deeper than the intended insertion depth, using a rotary-hammer drill with a 4,5 mm masonry bit, and a helical tie was then carefully hammered into it to a depth of 30 to 50 mm. For this operation, the ties were fitted into a sleeved device that allowed the tie to rotate during insertion, allowing the tie to cut a thread in the material. The gripper was then screwed onto the end of the tie and the load increased on the tie until the test material failed. The pull-out force was recorded. The pull-out tests on masonry units were carried out in their stretcher faces. In practice, the number of tests carried out in each individual unit varies depending on the size of the unit. In addition, the depth to which the tie was inserted into the units also varied: in most of the units types 30mm was used, although 50mm was used with some of the weakest AAC units, to ensure that a reasonable pull-out load was reached. During testing, each unit was used to provide 3 tests results for stretcher face, although some of the blocks were only tested in one face. In the weaker AAC blocks, the depth of insertion of the tie was increased to 50mm. For each unit the pull-out test was carried out first, followed by tests to determine the density, the water-absorption and then the compressive strength. Mortars: the test programme used to calibrate the screw pull-out test in mortars used 6 mortar mixes and four unit types, covering a range of water absorptions. The tests were carried out in nominally 10mm mortar beds, cast between two bricks in a couplet. Six pull-out tests were 28

2 â&#x20AC;&#x201C; State of the art carried out on each specimen, three for stretcher face. The couplets were lightly precompressed both before and during a test as it was felt that the insertion of the tie into unloaded couplets could cause some of them to break. Six mortar mixes were used, with different proportions of cement, lime and sand, by volume. Five couplets and three 100mm mortar cubes were made for each test mortar and unit type. All the specimens, including the cubes, were cured under polythene sheeting, and testing of both the couplets and the cubes was carried out 28 days after the specimens were manufactured. The pull-out testing took place on an Instron universal testing machine. Fergusonâ&#x20AC;&#x2122;s research: results on masonry units and on the mortars A scatter graph showing the results for all the pull-out tests carried out on medium strength masonry units is in figure 2.3.13.a. It is clear from the diagram that the results for materials with a compressive strength near 10 MPa show a poor relationship between the compressive strength of the material and the pul-out load. They clearly show that for these units either the test method needs to be refined or some other method needs to be devised to apply the load to the tie, thereby ensuring that the same mode of failure occurs whenever the material strength.

Figure 2.3.13 â&#x20AC;&#x201C; a) Scattergram of lower strength masonry unit tests; b) Scattergram of mortar test results.

For the mortars, as every combination of the 4 brick types and 6 mortars was tested, there were 24 sets of results. Each set of data was given an identification reflecting the combination of its constituents. Unfortunately, some of the couplets split before all six tests had been completed on them; in addition, different numbers of specimens of each type were made. As a result, the numbers of test results available for each combination of mortar and brick varied 29

2 â&#x20AC;&#x201C; State of the art consistently. A scattergram showing pull-out load plotted against compressive cube strength of the mortar for all the results of the pull-out tests is shown in figure 2.3.13.b. The results of the tests carried out on to access the screw pull-out test have been shown that the method provides a practical way of assessing the compressive strength of weaker masonry materials. However, the limitations placed on the system by both the gripper device and the change in mode of failure that occurs in stronger materials limits the use of the system to these weaker materials, and likely to continue to do so. For Ferguson it is possible that the use of smaller diameter ties may favorite shear failures in some stronger materials. However, the apparent change in the failure mode in stronger materials, from shear to compressive failure, is unlikely to be easy to overcome. Nowadays very little researches have been done on the helifix screw pull-out system, and to correlate the pull-out force to the compressive strengths only the graphs in figure 2.3.12 and 2.3.13 are available for mortars and masonry units. The helifix screw pull-out system can give information about batch-to-batch variability of strength or general quality, variation of quality in relation to a reference sample, and changes of properties with time. If we want to estimate the compressive strength, there are still too many uncertainties. In this thesis the study will be deepened. It will be proposed a method to estimate the mortar compressive strength from the pull-out force. Moreover we will try to interpret the physical phenomenon at the base of the pull-out test. In this way it will be possible to propose an analytical model able to univocall estimate the compressive strength of the mortar from the pull-out force.

2.3.3 Other minor-destructive tests (MDT) Due to the minor damage induced onto the structures, slightly destructive testing techniques are especially convenient when testing valuable historical buildings. The MDT can be split into four main categories: inspections, coring, endoscopies and flat jack. For a comprehensive introduction the reader is referred to Boving (1989) and Suprenant and Shuller (1994). CORING TECHNIQUE In the case of masonry composed of multi-layers, the coring technique is often used (see Figure 2.3.14.a). This method consists in the coring of small diameter boreholes and taking samples in the most representative sections, which can be mechanically tested. The boreholes

2 – State of the art can be used later for endoscopies (see Figure 2.3.14.b), which can provide valuable information about the existence of internal cavities and cracks.

Figure 2.3.14 – Inspections: a) Coring (Binda et al., 2001); b) Endoscopy (Binda et al., 2001).

FLAT- JACK TESTING Another common minor destructive technique is the flat-jack test. First used in the field of rock mechanics, flat-jack testing was later adapted by Rossi (1982) to be used on masonry structures. Nowadays, the flat-jack technique is used in the following tests:  Evaluation of the compressive stress state of masonry;  Evaluation of the compressive deformability properties of masonry;  Evaluation of the shear strength along the mortar joints. The compressive stress state is evaluated using a single flat-jack (Figure 2.3.15.a) placed inside a cut mortar bed joint (Figure 2.3.15.b). To evaluate the deformability characteristics of masonry, a cut parallel to the first one is made and a second flat-jack is inserted in this second cut. Therefore, the uniaxial compressive deformability properties of the masonry sample between the two parallel horizontal cuts can be assessed, including loading-unloading behaviour.

Figure 2.3.15 – Equipment and setup for flat jack test.

2 – State of the art

The flat-jack method also allows the measurement of the shear strength along a mortar joint, although this technique is seldom used. This test implies the removal of a brick from the centre of the masonry sample delimited by the two flat-jacks. A hydraulic jack is then put in the place of the removed brick and shear load is applied. This test allows obtain the peak and residual shear strength of the mortar joints. By performing this test on other places on the structure with different compressive stress states, it is possible to compute the friction angle and the cohesion of the mortar joints. All these evaluations can be done with minimum disruption to the masonry, since flat-jack testing requires only the removal of a portion of mortar joints and some individual bricks, which can be easily repaired to its original condition. X-DRILL SYSTEM The concept is to measure the maximum moment of torsion of the X-drill “knocked” into a pre drilled small hole in a mortar joint. The moment of torsion per unit length (knocked into the joint) is linear proportional to fm (compressive strength of the mortar joints) and the method gives a slightly more accurate prediction of fk (compressive strength of masonry) compared to the standardized method of measuring fm through EN 1015-11 (Christiansen,2011). The moment of torsion must be related to fm or fb (compressive strength of the mortar joints). According to EN 1996-1-1 (3.2), fk is:

The X-drill is shown in figure 2.3.16.

Figure 2.3.16 – Xdrill: photo and geometry

2 – State of the art

Up till approximately a M5 mortar an X-drill made of stainless steel is used. For mortar stronger than M5 high-yield steel is used. X-DRILL – the method (Cristiansen, 2011): 1. A Φ6 hole of 70-80 mm length is predrilled in the mortar joint. This hole, approximately equivalent to the inner diameter of the X-drill, will steer the X-drill and secure: a minimum of crushing when knocking the X-drill into the mortar joint and an X-drill only fixed through the flanges. 2. Afterwards a Φ 10 hole of 10 mm length is drilled. This minor hole will ensure a reduced influence on fm from any strong pointing mortar. 3. With a hammer the X-drill is knocked approximately 15-20 mm into the hole measured from the point of resistance (figure 2.3.17). This length is nominated Li. A specific value is not important since the relevant torque moment is per unit length. Li should be minimum 10 mm due to the accuracy though. Li should obviously be smaller than the shaft. For this prototype that is 50 mm.

Figure 2.3.17 – The X-drill is normally knocked 15-20 mm into the predrilled hole.

4. A torque-meter with a trailing pointer is used to determine the maximum moment of torsion under fracture (figure 2.3.18). This value is nominated Mv.

2 â&#x20AC;&#x201C; State of the art

Figure 2.3.18 â&#x20AC;&#x201C; Measurement of the maximum moment of torsion in the moment of fracture.

5. The value mv is determined as: mv= Mv/ Li. 6. The value mv,m is determined as the mean value of a series of mv; 7. The number of measurements should at least be 5 depending on the variation of the results. 8. When the X-drill is pulled out the adherence dust is briefly examined. If part of the dust is yellow or red as the units the results are ignored (because the drilling has erroneously been performed partly in the unit). 9. Extremely high values are ignored (while extremely low values are not). Extremely high values indicate that the X-drill was probably restrained in the units, while extremely low values indicate worn or washed joints.

3. Laboratory experimental program 3.1 First campaign on lime mortar In a first step, it was decided to construct a series of wallets, 4 for each type of lime which are described in section 3.3.1 of this thesis. In this way, for each type of lime it is possible to obtain experimental data about 4 ages of maturation and thus to have different resistance values, useful to adequately calibrate the Helifix screw pull-out by obtaining different points (

A summary of the objectives of the first experimental campaign on mortars:  Perform extraction tests with the Helifix screw pull-out system in the mortar joints of wallets, and, on the same joints, perform the double punch test. In this way it will be possible to have a precise correspondence between the compressive strength and the values obtained from the minor destructive tests, that allows the understanding of the correlation (

) for lime mortars.

 Understand exactly how the gypsum on the loading surfaces of the samples affects to the compressive strength. For this test the requirements of DIN 18555-9 were followed, especially in reference to chapter 5 ”Testing of mortar from masonry”, section “5.1 Method III”: «5.2.1 Principle: Specimens approximately 50mm square or 50mm in diameter are prepared from samples taken from masonry (e.g. by core drilling or masonry unit/joint assemblies). The specimens are then placed between a pair of loadingplatens measuring 20mm in diameter and tested for compressive strength. » This principle is followed in the destructive tests of the joints. DIN 18555-9 in paragraph 5.2.3 specifies that: «….Only specimens with plane and parallel upper and lower faces may be used for the test. Where necessary, surfaces may be levelled by applying a suitable material (e.g. gypsum plaster) to a thickness of not more than 1mm.» The test and procedure adopted will be described more accurately in section 3.1.6.  Study the relationship between the compressive strengths

and

(respectively the compressive strength from DPT, and the compressive strength according to the EN 1015-11 standard “Methods of test for mortar for masonry”). The masons’ procedures for the production of wallets, and therefore of the joints, are always the same, and much attention is paid to the regularity of these (more or less the

3 – Laboratory experimental program same thickness for all joints). The prismatic specimens making process is also always the same. Would it be possible that the differences between the DPT square specimens and the prismatic specimens are: compaction of the mortar, specimen shape factor and boundary conditions during the maturation. Also the ages for the tests will be the same, or as close as possible, so we think that it will be possible to deduce a correlation between the two types of test.  Understand the correlation between the compressive and the flexural strengths from a comparison of the experimental results in different ages of maturation.  Study the effect of a material capable to absorb water, on the maturation of lime specimens. By construction and use of wooden molds, it may be possible to obtain resistance values more comparable with the compressive strengths obtained from the joints, because wood and bricks absorb the water, perhaps favoring the maturation of lime especially during the first few days. The steel molds does not favour the loss of water.  Understand the differences between the behaviour and the mechanical properties of different types of lime.

3.1.1 Materials To better represent the historical mortars, it was decided to make two types of aerial lime mortars, and a type of hydraulic lime mortar, all with lime/sand ratio 1:3. Probably the same ratio will allows us to obtain comparable data on different types of mortar. Moreover there is a fourth premixed mortar, Kerakoll “Biocalce Muro”, which will give higher resistances than the first three mortars, and will reach full maturity more quickly.

3 – Laboratory experimental program

Figure 3.1.1 – a) Kerakoll mortar (Biocalce Muro); b) NHL3,5_1:3 “Cementos Tigre” c) CL90_powder aerial lime “Ciaries”

In addition for a premixed mortar, we expect lower experimental dispersion compared to those that would we have with mortar mixed by hand. So the materials used are: 1) First row of wallets: Kerakoll mortar “Biocalce Muro” Biocalce Muro (fig.3.1.1.a) is a product supplied us by Kerakoll S.p.a., an Italian company that operates in the construction sector. According to the data sheet of the product, it is a mortar made with hydraulic lime NHL 3,5, according to EN 459-1, very suitable for the works on historic restoration, and this is one of the reasons for our choice. The strength class is M5, according to EN 998-2, or rather the compressive strength according to EN 1015-11 (compressive test on prismatic specimens after 28 days from molding) is 5 MPa. Is important to say that this mortar, in addition to hydraulic lime NHL 3.5, contains other elements:  silica washed litter from fluvial quarry (grain size: 0,1÷0,5 mm);  silica washed litter from fluvial quarry (grain size: 0,1÷1 mm);  dolomitic selected limestone (grain size: 0÷2,5 mm);  powder of pure white marble of Carrara (0÷0,2 mm). In the data sheet a lot of informations are reported, such us density, temperature range for application between 5 and 35°C (range always respected in the laboratory), and the correct water mixture ratio:  4,4 l of water each 25 l of mixture are needed: 4,4 l (H2O) / 25 l

3 – Laboratory experimental program This rule will be followed very strictly by the mason. 2) Second row of wallets: hydraulic lime NHL 3,5 “Cementos Tigre” NHL 3,5 lime (fig.3.1.1.b) will be mixed with sand in ratio 1:3, to make a lime mortar with low mechanical performance. We want to make a comparison with kerakoll mortar, which contains hydraulic lime, and a comparison with the other two types of air lime mortar. This lime is made by a Catalan company, “Cementos Natural Tigre”. It is a NHL 3,5, according to EN 459-1, or rather the characteristic compressive strength calculated following the technical specifications of EN 196-1 must be

after 28 days

from molding. Obviously we talk about the lime properties without addition of sand, but when the ratio lime/sand=1/3, the mechanical strengths are lower. 3) Third row of wallets: “Ciaries”__aerial calcium lime CL90 (hydrated lime powder) It is a calcium lime powder (fig.3.1.1c), a calcium hydroxide. It was provide by a company of Barcelona called Ciaries. It derives from a hydration process of quicklime, which makes the lime “slaked”. According to the data sheets, certificates and statements of performance provided by the company, the product complies with the requirements in EN 459-1. It is called CL90 because it has a content of

(by mass). More precisely

it cumplies the requirements indicated in table2, section 4.4.2 of EN 459-1 (Table 3.1.1). Table 3.1.1 – Chemical requirements for the lime, expressed as characteristic values_EN 459-1

In addition, it complies with the physical requirements specified in paragraph 4.4.3 of the same standard.

3 – Laboratory experimental program 4) Fourth row of wallets: “Segarra & Hernandez”__aerial calcium lime CL90 (slaked lime)

Figure 3.1.2 – Slaked lime CL90 “Segarra e Hernandez”; a) Sack of slaked lime (20 Kg); b) Hydration in the mixing machine with an excess water; c) Storage of lime in two plastic buckets and periodical mixing with helical accessory connected to a drill.

The fourth row of the walls will be made using a mortar with slaked lime, always trying to keep the proportions lime:sand=1:3 (volume). The Catalan company Segarra provide us a lot of bags of slaked lime CL90 (Figure 3.1.2.a), left in a solution with an excess water during the production process and packaging. In a slaked lime, the lime in contact with water benefits from the time effects, in the process called “aging” or “maturing”, during which the hydroxide calcium crystals are subjected to important morphological and dimensional changes: plasticity, workability and water retention increase. Slaked lime carbonates faster than powder lime, with benefits for the building’s durability and resistance. In the thesis a comparison will be made between the strengths provided by specimens made with aerial lime powder and sand and specimens made with slaked lime and sand. An aim is to understand if and how the mechanical properties of the two different types of lime will change. Many bags were ruined and they lost a certain percentage of water. To obtain uniform lime hydration, the content of the bags was mixed within the mixing machine (Figure 3.1.2.b) putting an excess of water. Then all was put in 2 big plastic containers (Figure 3.1.2.c), mixing periodically to improve hydration of the product and avoid the formation of lumps. The requirements that the lime respects are the same of the lime powder (see the previous point). Physical and chemical requirements of the CL90: the technical data sheets ensure the quality of the product.

3 – Laboratory experimental program 5) Fine-grained river sand “Arids Catalunya” Relying on earlier work of thesis, where students used sands with particle sizes too large for lime mortars, with which were obtained not always satisfactory results, in this research it was decided to use sand with a little particle size, in the range 0-2 mm. It was used a river sand provided by the catalan company “Arids Catalunya Sorigué” (Figure 3.1.4.a) which has also provided the technical data sheet of the product (Figure 3.1.3).

Figure 3.1.3 – Particle size in the technical data sheet of the product.

Observing the particle size obtained following the standard EN 933-1, with sieves described in EN 933-2, we can understand that it is a very fine sand, with a passer of 97% by mass from the sieve with mesh of 1mm, and a passer of 79% from sieve with mesh of 0,5 mm. However, the grain size has been proven by an independent test, according to EN 933-1 and EN 933-2, but it is certainly very fine sand (results in Figure 3.1.5). The EN-933-1 standard was followed for the procedure, and it was chosen a group of sieves according to regulation 933-2, knowing to the maximum particle size. The procedure is summarized below: a) The samples were reduced following the 933-2, to obtain the required number of test portions; the size of each portion shall be at least 0,2 kg, according to table 1 of this standard (

. It was taken a sample with total weight

b) Then the test portion was dried, heated at 110 °C, until constant mass. It was weighed and

was registered.

c) The washing was made to remove micro particles, with size less than 0,063 mm, according to paragraph 7.1 of EN 933-1. During the washing the sand was placed in water and then sieved within 0,063mm sieve to eliminate suspensions. Then the sand was weighed and

was registered.

3 â&#x20AC;&#x201C; Laboratory experimental program First fine fraction is obtained from:

d) The sieves were placed in column, from largest opening sieve to the closest one, choosing the sieves according to EN 933-2. The last sieve is a protection sieve to collect the residual fine part. e) The dried sand was putted in the first sieve. The tower of sieves was mechanically shaked by vibrating screen (Figure 3.1.4.b) f) Lastly the remaining sand in each sieve was weighed (Figures 3.1.4.d and 3.1.4.c). The passing percentage to each sieve is calculated as: [ ]

â&#x2C6;&#x2018;

We have 7 sieves and the sieve with n=7 is the first one at the top, with mesh of 4mm. Particle size analysis made in the laboratory provided a particle size similar to that in the data sheets, with a percentage by mass of grain with size of 2 mm equal to 0.4%, while in sheets this percentage is 0%.

3 – Laboratory experimental program

Figure 3.1.4 – a) River sand 0-2 mm “Arids Catalunya Sorigué; b) Vibrating screen; c) Remaining sand in each sieve; d) Remaining sand weighing.

Figure 3.1.5 – Results of particle size made in the laboratory.

3 – Laboratory experimental program 6) Bricks: “Terra Cuita Piñol Pallarés S.L.” The units used in this research are handmade bricks, made by the company “Terra Cuita Piñol Pallarés S.L. Spain” (Figure 3.1.6.a). The nominal dimensions of the units are 345x145x45 mm (Figure 3.1.6.b). It is important to point out that there is a large variability of the brick dimensions due to its peculiar construction procedure. From the product data sheet supplied by the company, we can find the compressive normalized strength average, calculated according to UNE-EN-772-1. It is:

Figure 3.1.6 – a) Terra Cuita bricks; b) Nominal dimensions 345x145x45 mm.

3.1.2 Construction of wallets and specimens Firstly it was decided the place in the laboratory to be allocated for the walls construction.  16 walls were built, 4 for each type of mortar.  4 rows of wallets were made, on the wooden beams; 1), 2), 3) and 4) represent the ages of maturation, while rows a), b), c) and d) represent the groups of wallets for type of mortar (Figure 3.1.7).  So each type of mortar had four different ages of maturation, on which it was possible to perform the various programmed tests, according to their expected speeds of maturation. For example Kerakoll mortar was tested in shorter and closer ages of maturation than the other types of mortars, because the data sheets show a rapid growth of mechanical properties.  Each wall consists of 7 rows of two bricks, with regular in thickness joints (more or less 15mm), in order to facilitate the minor destructive tests (with Helifix and 43

3 – Laboratory experimental program Windsor), and to extract enough joints to make square specimens for DPT. In the first design we thought to test half wall with the two instruments and to destine the other half for joints extraction. After the first test on the first wallet, it was decided to continue making tests on whole wallets face, then extracting joints for get square specimens, because the two minor destructive tests didn’t absolutely disturb the specimens.

Figure 3.1.7 – Disposition of the walls in the laboratory.

Now all proportions lime/sand/water used for the mortar making process are explained. The ratio lime/sand was made by volume, using some metal pails with precise volumetric capacities, in order to reduce as much as possible the dosification errors. A) NHL 3,5__1:3__hydraulic lime mortar, with NHL 3,5 “Cementos Tigre”__dosification: -

15000 cm3 of sand;

5000 cm3 of NHL 3,5 “Cementos Tigre”

5200 cm3 of water.

B) Kerakoll mortar__hydraulic lime mortar M5 -

25 kg of mixture; 44

3 – Laboratory experimental program -

4400 cm3 of water.

This proportion was made in order to respect the fraction indicated on the sack. C) CL90__“Ciaries”__Aerial lime mortar with powder lime CL90__dosification: -

15000 cm3 of sand;

5000 cm3 of powder aerial lime CL90 “Ciaries”;

7000 cm3 of water.

D) CL90__“Segarra”__Aerial lime mortar with slaked lime CL90__dosification: -

11000 cm3 of slaked lime;

30000 cm3 of sand.

The fraction lime/water was 0,71 for “CL90_Ciaries” mortar, 0,96 for NHL 3,5 “Cementos Tigre” mortar and it isn’t exactly know for “CL90_Segarra” mortar, where 1,1 parts of slaked lime and 3 parts of sand were used. The doses were chosen in order to obtain mortars with lime and sand in proportion 1:3 and to obtain workability mortars, following previous experiments such us “M.Arandigoyen et al. 2005”. -

Construction of wallets:

As regards the construction method of the wallets, the mason followed standard rules, trying to make vertical and horizontal joints with the same thickness. In wallets’ faces designed for penetrometric tests, the joints must be at the level of the outer edge of the bricks, creating a surface as regular as possible and easily testable, with the objective of limiting reading errors especially in Windsor Pin tests. The steps are the following: -

Cutting some bricks to half by disc saw (Figure 3.1.8.a);

Immersing and keeping bricks under water for 15÷20 min (Figure 3.1.8.b): the bricks must be not too dry, otherwise the mortar loses quickly the moisture;

3 – Laboratory experimental program

Figure 3.1.8 – a) Cutting bricks by circular saw; b) Immersing and keeping under water for 15 – 20 min.

Preparing basement. In order to have enough space to performing penetrometer test we decided to build the walls 30 centimeters upper than ground. Walls were built on wooden beams (Figure 3.1.10);

During the walls building, vertical and horizontal direction were leveled (Figure 3.1.9. a);

Walls dimensions: 690x420x145 mm≈. 7 rows of two bricks (Figure 3.1.9.b);

Figure 3.1.9 – a) Vertical and horizontal joints leveled; b) Wallet dimensions (7 rows, 2 bricks each row).

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.1.10 â&#x20AC;&#x201C; 16 wallets built on wooden beams.

Specimens molding and maturing:

During the construction of the wallets, standard specimens 40x40x160 mm were made, according to the rules of UNE-EN-196-1. They were made 6 specimens in steel molds (3 for flexural and compressive tests, and 3 for Brazilian tests) and 3 specimens in wooden molds (for flexural and compressive tests in order to compare strengths of specimens made in steel and wood molds) (fig 3.1.11.b). Wood molds were handmade (fig 3.1.11.a).

Figure 3.1.11 â&#x20AC;&#x201C; a) Handmade wooden molds; b) 6 specimens in steel molds and 3 specimens in wood mold.

3 – Laboratory experimental program Each group of two steel molds plus one wooden mold represent the wallet at age-i, and the samples must be made with the same mortar of the walls and it must mature in the laboratory under the same climatic conditions of the wallets (Figure 3.1.12.g). All procedure is precisely described in “BS-EN 1015-11”  Firstly the internal molds surfaces were spread with a special oil, helpful to the subsequent specimen removing (Figure 3.1.12.a).  The mold was half-filled, giving 25 strokes with a special piston (Figure 3.1.12.b).  Then the mold was entirely filled and the mortar was compacted with 25 strokes with tool and 25 shots more on the compacting machine according to UNE-EN-196 (Figure 3.1.12.c).  The excess of mortar was removed by a metal ruler (Figure 3.1.12.d).  As regards the maturation, the guidelines of the “BS-EN 1015-, were followed but also the procedure described in the paper (Baronio et al. 1995). In the first two days the specimens were left at a relative humidity between 90-95%, putting the molds in plastic bags (Figure 3.1.12.e) and covering the upper surface by a plastic film, in a climatic chamber with a temperature around 20±2°C (Figure 3.1.12.f).  After two days the plastic bags were removed and the hydraulic lime mortar specimens continued the maturation near the wallets (Figure 3.1.12.g), at the same conditions, while the specimens with aerial lime were leaved 4 days more in the climatic chamber at a relative humidity of 75% and at 20°C. This was made to prevent the rapid loss of moisture and the consequential cracking. After four days in these conditions, also the aerial lime specimens continued their maturation near the wallets.

3 – Laboratory experimental program

Figure 3.1.12 – Molds making procedure: a) Sprinkling oil inside molds surface; b) 25 strokes with a piston; c) 25 strokes with the compacting machine; d) Excess of mortar removing; e) Putting the molds in plastic bags, to get humidity between 90-95%; f) 4 days in the climatic chamber; g) Specimens Maturation near the wallets, at the same climatic conditions.

Specimens naming:

In this thesis there are 4 types of mortar to test at different ages. For example N-S-3-1 means NNHL3,5 “Cementos Tigre”, SSteel mold, 3age3, 1specimen number 1. After flexural testS, they will be 1.acompression on the first half and 1.bcompression on the second half. It will be C “CL90 Ciaries”, S”CL90 Segarra”, K”Kerakoll mortar”. After a certain hardening period, the specimens are ready for the first tests.

3.1.3 Flexural tests on prismatic specimens of mortar The flexural strength was evaluated according to the EN 1015-11:2007. 49

3 â&#x20AC;&#x201C; Laboratory experimental program The flexural strength

] is:

[

Where

[ ], ,

Three prisms for each age were tested. The tests were performed in the Ibertest machine with a load cell of 10 kN, in load control (Figure 3.1.13). The load rate was kept constant during the test with a value which allows us have a breakage between 30 and 90 seconds. Precisely it was tried to keep the load to get failure in 60 second, calibrating according to previous experiments and expected strengths.

Figure 3.1.13 - Flexure test 1015:11; a) Setup of the test; b) Correct position of the sample; c) Failure mode of the sample.

For each mortar and age it was made a table like Table 3.1.2, with some photos and all data:

3 – Laboratory experimental program Table 3.1.2 – Example of flexural test session: NHL 3.5 prisms.

The results obtained over the time for each type of mortar are shown in the table 3.1.3 and in the graphics (Figure 3.1.14.a, .b, .c): Table 3.1.3 – Results of flexural tests on NHL 3,5_1:3 mortars.

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.1.14 â&#x20AC;&#x201C; Flexural strength VS Time (EN 1015:11)_ mortar NHL 3,5_1:3; a) fflex VS Time, specimens made in STEEL MOLDS; b) fflex VS Time, specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS

As it can be see in figure 3.1.14, the compressive strength growths fastly for specimens made in wooden molds, and the values remain constant after 28 days, while for specimens made in steel molds, after the peach there is a resistance drop. For Kerakoll mortar M5, data tables were made for each test session, for example (Table 3.1.4):

3 – Laboratory experimental program Table 3.1.4 – Example of flexural test session on Kerakoll mortar

For Kerakoll mortar the data of flexural strength in the first experimental campaign are (Table 3.1.5 and Figure 3.1.15.a, .b, .c): Table 3.1.5 – Results obtained from the flexural tests on Kerakoll mortar M5

3 – Laboratory experimental program

Figure 3.1.15 – Flexural strength VS Time (EN 1015:11)_Kerakoll mortar “Biocalce muro M5; a) fflex VS Time, specimens made in STEEL MOLDS; b) fflex VS Time, specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS

As reported in the graphs (Figure 3.1.15.a, .b, .c), the flexural strength behaviour presents an increase in the first 28 days, and a subsequent little degrowth. A similar trend of decrease after the peak was observed also in the same test with a different mortar mix, by Baronio, 1999 and Witt, 2014. The reason of such a decrease of strength is still not clear at all and might be related to ageing of mortar, deserving to be better investigated in further researches. For mortar CL90_1:3_”Ciaries” with aerial powder lime, flexural strength data are (Table 3.1.6 and Figure 3.1.16.a, .b, .c):

3 – Laboratory experimental program Table 3.1.6 – Results obtained from flexure tests on mortar CL90_1:3 “Ciaries” (with powder aerial lime)

Figure 3.1.16 – Flexural strength VS Time (EN 1015:11)__ mortar CL90_1:3 “Ciaries” (with powder aerial

lime); a) fflex VS Time, specimens made in STEEL MOLDS; b) fflex VS Time, specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS

As reported in the graphs (Figure 3.1.17.a, .b, .c), the flexural strength of specimens made in wood molds is significantly higher than the flexural strength of specimens made in wood molds. In facts the differents between the flexural strengths is more than 10%. The experimental dispersions are low and usually lower for specimens made in steel molds. For mortar CL90_1:3_”Segarra” with slaked lime, the data of flexural strength are (Table 3.1.7 and Figure 3.1.17.a, .b, .c): 55

3 – Laboratory experimental program Table 3.1.7 - Flexural tests on mortar CL90_1:3 “Segarra” (with slaked lime)

Figure 3.1.17 – Flexural strength VS Time (EN 1015:11)_ mortar CL90_1:3 “Segarra” (with slaked lime) a) fflex VS Time, specimens made in STEEL MOLDS; b) fflex VS Time, specimens made in WOOD MOLDS; c) fflex VS Time, STEEL MOLDS VS WOOD MOLDS

In the same way observed for the other mortars with aerial powder lime, for mortar with slaked lime CL90_1:3 “Segarra” the flexural strength averages of specimens molded in wood are higher than the flexural strength of specimens molded in steel, and the differents are more than 10%. So it can be assumed that for mortar with aerial lime, a material like wood that absorbs water in the first days of maturation greatly affects the flexural strength average. 56

3 – Laboratory experimental program

3.1.4 Compressive tests on mortar prismatic specimens The compressive strength was evaluated directly by the compression test of the two mortar prisms halves, according to EN 1015-11:2007. The compressive strength

[

]

calculated like:

Where: F = maximum load [N]; 402 = load platens’ area [mm2]. Three prisms for each age were tested. The tests were performed in the Ibertest machine with a load cell of 10 KN, in load control. The load rate was kept constant during the tests with a value which allows have failure between 30 and 90 seconds. Precisely it was tried to keep the load to get rupture in 60 seconds, calibrating according to previous experiments and expected strengths. In Figure 3.1.18 the test setup and the typical hourglass failure mode are shown.

Figure 3.1.18 - Compression test EN 1052:11:2007: a) Setup of the test; b) Failure; c) Typical hourglass failure mode

For mortar NHL 3,5_1:3_”Cementos tigre” the data of compressive strength are (Table 3.1.8) and the graphs in figure 3.1.19 shows the evolution of the compressive strength over time.

3 â&#x20AC;&#x201C; Laboratory experimental program Table 3.1.8 â&#x20AC;&#x201C; Results obtained from the compressive tests on NHL 3,5_1:3 mortar

Figure 3.1.19 - NHL 3,5_1:3: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparison of results: STEEL MOLDS VS WOOD MOLDS.

Considering Figure 3.1.19.a, 3.1.19.b, the compressive strength behaviour increases with time, but after 60 days it decreases, as for specimens made in steel molds as for specimens molded in wood. Like in the flexural tests, also in the compressive tests for specimens molded

3 – Laboratory experimental program in wood the scatter is larger and the maturation is faster than for specimens made in steel molds. The rapid maturation in the first days is due to the fact that the wood helps the elimination of the mixing water, approaching the bricks behaviour. The higher scatter for wood-molds specimens may be due to the fact that these molds are handmade and each piece of wood absorbs differently. Steel can not simulate the behaviour of the bricks during the specimens maturation but is a good material for the prisms making process, because its behaviour is constant and the molds are all the same. For comparisons with the data of the other experimental tests, the data of the tests made on samples molded in steel will be used. For mortar “Kerakoll M5”, and for the other mortar types, at each test session a data table like 3.1.9 was made. Table 3.1.9 – Example of compressive test session on KERAKOLL mortar M5

For mortar “Kerakoll mortar M5” all compressive strength results are in table 3.1.10 and the graphs in figure 3.1.20 show the compressive strength evolution over time.

3 – Laboratory experimental program Table 3.1.10 – Results obtained in the compressive tests on Kerakoll mortar “Biocalce muro” M5.

Figure 3.1.20 – Kerakoll mortar M5: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparison of results: STEEL MOLDS VS WOOD MOLDS.

3 – Laboratory experimental program Considering Figures 3.1.20.a, .b, for Kerakoll mortar the compressive Strength increases with time peaking at 28 days but after it decreases, both for specimens made in steel molds and for specimens molded in wood. Like in the flexural tests, also in the compressive tests for specimens molded in wood the scatter is bigger and the maturation is faster than for specimens molded in steel. The faster maturation in the first days, and the higher dispersion for specimens molded in steel, are do to the same reasons discussed for the NHL 3,5_1:3 mortar. It is also observable that after a certain aging period, for each type of mortar the compressive strength averages are the same for specimens molded in wood and specimens molded in steel. This suggests that the wood absorption effect is effective only in the early days of maturation. For mortar “CL90_1:3 Ciaries”, with powder aerial lime, all data of compressive strength are in table 3.1.11 and the graphic in figure 3.1.21 shows the compressive strength over time. Table 3.1.11 - Results obtained in the compression tests on mortar CL90_1:3 “Ciaries”

3 – Laboratory experimental program

Figure 3.1.21 – CL90_1:3 “Ciaries”: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparation of results: STEEL MOLDS VS WOOD MOLDS.

Considering figure 3.1.21.a, for specimens molded in steel the compressive strength does not vary over time. At the third age of maturation the compressive strength decreases, then increases again in the fourth age. Observing figure 3.1.21.b, for specimen made in wooden molds the compressive strength trend is more linar. After the peak at 125 days, the strength remains the same (same values in the third age). In the graph of figure 3.1.21.c it can be observed that the compressive strength averages are significantly higher for specimens molded in wood, and the difference is more than 10%. This happens in all maturation ages except for the first age, but in this session there have been difficulties in the phases of demolding, and some specimens were ruined. For mortar “CL90_1:3 Segarra” with slaked lime, all compressive strength data are in table 3.1.12 and the graphs in figure 3.1.22 show the compressive strength over time.

3 – Laboratory experimental program Table 3.1.12 - Results obtained in the compressive tests on mortar CL90_1:3 “Segarra”

Figure 3.1.22 – CL90_1:3 “Segarra”: Compressive strength VS Time; a) Specimens made in STEEL MOLDS; b) Specimens made in WOOD MOLDS; c) Comparison of results: STEEL MOLDS VS WOOD MOLDS.

Considering figures 3.1.22.a and 3.1.22.b, the compressive strength trend is similar for the two types of specimens, with a constant growth until the third maturation age. After the peack the compressive strength remains more or less constant. The scatter is very low, and this 63

3 – Laboratory experimental program means that the specimens making process and the test performing were successful. The resistance of the specimens molded in wood is constantly greater than the resistance of standard specimens molded in steel. The difference exceeds 10% and it is important. This was also the case of the mortar “CL90_1:3 “Ciaries”. It can be assumed that the production of the prismatic specimens with wooden molds is a good technique for mortars with aerial lime. The wood is a material that helps the loss of moisture, simulating the bricks behaviour, and the resistances obtained from the specimens may be more truthful.

3.1.5 Brazilian tests on prismatic specimens of mortar The assessment of the direct tensile strength is an important issue when the material is very brittle. Therefore, it was decided to perform the Brazilian test following the ASTM C496:1996, designed for concrete, and some available researches (Rocco et al. 1999). The Brazilian test was performed on the prismatic mortar samples with dimension of 40x40x160 mm3, for all types of mortar, at differents ages of maturation. In some test sessions, after testing the two halves obtained were tested again with the Brazilian test, since the dimensions exclude different results due to the size effect. This was not always done because the experimental dispersions for this test on this kind of mortars were very low, and we didn’t need to make many tests. The tests were made using a test frame expressly prepared for the Brazilian test. The machine used was the Ibertest with a load cell of 10 kN, and the load rate was chosen in order to have breakage in 60 seconds. An important characteristic of this test is the dimension of the strip of wood used to distribute the point load over the specimen. In this test the dimension of the strips was 4mm, which correspond to the 10% of the specimen base. The setup details are reported in Figure 3.1.23.

Figure 3.1.23 - Brazilian test; a) Test setup; b) Failure in the whole prisms

3 – Laboratory experimental program The failure mode showed a clear diametric fracture through the centre of the specimen (Figure 3.1.23.b and 3.1.23.b). The formula used to calculate the tensile strength, knowing the maximum force obtained, is:

Where Fmax = maximum force [N], b = section base [mm], h = section height [mm]. In each test session, the results were gathered in a table like this (Table 3.1.13): Table 3.1.13 – Example of brazilian test session on NHL 3,5_1:3 mortar

For example, NHL3,5_1:3_br_age3_st_1.1 means: -

NHL3,5  mortar with hydraulic lime NHL 3,5 “cementos tigre”;

1:3  lime:sand = 1:3, in volume;

br  brazilian test;

age3  age of maturation 3;

st  steel mold (brazilian tests were performed only on specimens molded in steel);

1.1  is an half of the specimen called 1, tested another time.

Following similar rules, in the other tests all the signs were given. For mortar NHL 3,5_1:3_”Cementos tigre” the data of tensile strength (obtained from brazilian test) are (Table 3.1.14 and Figure 3.1.24.a):

3 – Laboratory experimental program Table 3.1.14 - NHL 3,5_1:3 mortar; Results obtained from brazilian tests

For “Kerakoll mortar M5” the data of tensile strength are (Table 3.1.15 and Figure 3.1.24.b): Table 3.1.15 – Kerakoll “Biocalce Muro” M5; Results obtained from Brazilian tests

For mortar CL90_1:3 “Ciaries” the data of tensile strength are shown in Table 3.1.16 and Figure 3.1.24.c:

3 – Laboratory experimental program Table 3.1.16 – CL90_1:3 “Ciaries”; Results obtained from Brazilian tests

For mortar CL90_1:3 “Segarra” the data of tensile strength (obtained from Brazilian test) are shown in Table 3.1.17 and Figure 3.1.25.d: Table 3.1.17 – CL90_1:3 “Segarra”; Results obtained from Brazilian tests

As reported in the tables 3.1.14, 3.1.15, 3.1.16, 3.1.17 and in figure 3.1.24, the results obtained in the Brazilian test are hardly acceptable. The values are too low and if compared with the flexural strength, they are five times lower. For aerial mortars the trend is similar. The resistance poorly grows and maintains more or less constant values. The mortar with slaked lime shows higher values than mortar with aerial powder lime. For NHL 3,5_1:3 mortar, after 28 days the resistance values stabilize, while for kerakoll mortar the values lightly decrease, showing a trend similar to the flexural strength at the same ages, but the ratios

at any age are very different.

3 – Laboratory experimental program Such discrepancies might be due to the very low friction angle of the material, insufficient to develop the typical complete splitting mechanism characterized by two wedge regions under the punches and a vertical connection crack connecting them. The very low experimental values might be related to a local failure mechanism. Further investigations are necessary to better understand this phenomenon.

Figure 3.1.24 – Brazilian test results, Tensile strength VS Time; a) Kerakoll mortar M5; b) NHL 3,5_13 “Cementos Tigre”; c) CL90_1:3 “Ciaries”; d) CL90_1:3 “Segarra”

3 â&#x20AC;&#x201C; Laboratory experimental program

3.1.6 Double Punch Tests on mortar joints The specimen preparation is shown in the following figures:

Figure 3.1.25 â&#x20AC;&#x201C; DPT, specimens' preparation: a) Removin joints by hammer and chisel; b) Drawing squares 5x5 cm on the removed joints; c) Cutting removed joints by a circular saw; d) Samples (at least 4 each joint); e) Gypsum paste on half group of joints; f) Measuring specimens' thickness;

3 – Laboratory experimental program The punching tests were performed on the specimens obtained from each wallet, on which minor destructive testing were previously performed. Specifically, the rules indicated in chapter “5. Testing of mortar from masonry”, section “5.2 Method III” of DIN 18555-9 were followed. “Specimens approximately 50 mm square or 50mm in diameter are prepared from samples taken from masonry (e.g. by core drilling or masonry unit/joint assemblies). The specimens are then placed between a pair of loadingplatens measuring 20mm in diameter and tested for compressive strength.” The aim is obtain two series of well comparable data, because destructive and minor destructive tests were performed on the same materials (the joints). UIC, 2011; DIN 18555-9, 1999; and Binda et al, 2002 cover the specifications for this test. The steps were: 1) The joints were removed using hammer and chisel (Figure 3.1.25.a); 2) Squares 5x5 were marked on the joints (Figure 3.1.25.b); 3) The removed joints were cutted by a circular saw (Figure 3.1.25.c), trying to get 4 samples for each joint (Figures 3.1.25.d), 2 to be tested with gypsum paste, 2 to be tested without gypsum. In fact DIN 18555-9 standard suggests using specimens with parallel faces and whenever this is not possible it recommends putting a gypsum layer with a thickness lower than 1mm. We want to study the differences testing with and without gypsum, especially on medium resistance and the experimental dispersions. 4) On half group of joint a layer around 1 mm of gypsum paste was applied (Figures 3.1.25.e); 5) Before the experiments, the specimens’ thickness were always measured (Figures 3.1.25.f), to study how the thickness affects the compressive strength. The tests were performed with the Ibertest machine with a load cell of 10 kN. The load rate was chosen to get failure around 60 seconds. The specimens were centrally placed between the loading platens (D=20mm) of the testing machine, with a full contact between specimen and platens over the loading areas. The setup is shown in Figure 3.1.26. In Figure 3.1.27.a, 3.1.27.b and 3.1.27.c were shown respectively a test and the failure modes of the specimens, with and without gypsum.

3 – Laboratory experimental program

Figure 3.1.26 – Test setup (DIN 8555-9:1999)

Figure 3.1.27 – Double Punch Test setup: a) Ibertest and configuration with fixed punches; b) Centered specimen; c) Typical hourglass failure mode of central zone

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.1.28 â&#x20AC;&#x201C; a) Typical failure mode of a specimen with gypsum layer; b) Typical failure mode of many specimens without gypsum layer;

The Figure 3.1.27 shows the configuration of the machine and punches. We have chosen the fixed punches configuration, because in the hinged configuration the punches transmit to the specimens, uncontrollable components of moment, which depend to the surfaces parallelism. The figure 3.1.28.b shows a frequently type of breakage for the specimens without gypsum. Even if we choose specimens with parallel faces, the load is applied on the specimen with a non-full contact, and the edges of the punches cut the specimens in this way. The compressive strength is underestimated. When we test specimens with gypsum, the load is more uniformly distributed, and this is also known observing the breaking mechanism (Figure 3.1.28.a). There will be a central zone who breaks forming an hourglass (Figure 3.1.27.c), and in the surrounding part the cracks propagate. A table like 3.1.18 was made for each test session.

3 – Laboratory experimental program Table 3.1.18 – Example of DPT test session on Kerakoll mortar M5 (AGE 2)

As it can be seen from the tables, 4 specimens for each joint were tested, 2 treated with gypsum and 2 untreated. 24 samples for each wallet were tested. The specimens were cut in squares 50x50 mm. The thickness was not constant due to the nonperfect bricks flatness, with a range of 15÷25 mm. In the graphs of Figures 3.1.29.a and 3.1.29.b, two examples of compressive strength VS thickness are shown. In some experiments the compressive strength is higher for specimens less thick, and the results are in accordance with Benedetti, Pelà 2012. But in some other experiments, the trend is opposite.

3 – Laboratory experimental program

Figure 3.1.29 – DPT_Kerakoll mortar M5_Campaign 1_AGE2 (28 Days); a) fcDPT VS Thickness, joints treated with gypsum paste; b) fc DPT VS Thickness, joints untreated.

For mortar “NHL 3,5_1:3” all data of DPT compressive strength are (Table 3.1.19): Table 3.1.19 – NHL 3.5_1:3; Results obtained from DPT on square specimens of mortar.

The compressive strength fcDPT is calculated like:

In (3.4) fcDPT is the compressive strength [MPa]; Fmax is the maximum force that breaks the specimens [N] and π·202 = load area of the punches. For walls made with mortar NHL 3,5_1:3, it wasn’t always possible to obtain 4 specimens for each joint as planned, because the mortar bound strongly with the bricks, and often in the 74

3 – Laboratory experimental program extraction the joints splintered, and it was impossible to get all specimens with plane and parallel faces (Figure 3.1.30.a and 3.1.30.b):

Figure 3.1.30 – a) and b) show difficulties in joint’s extraction.

So for some sessions less than 24 tests ware made (12 with and 12 without gypsum) but sufficiently to respect the DIN 18555-9, which prescribes at least 10 tests. Observing the results (Figure 3.1.31.a, 3.1.31.b and Table 3.1.19), we can say that in general there are no great differences between the average resistances in the various age of matation: from first maturing age (14 days) to the fourth age (110 days), there is a growth of only 0,38 Mpa for the specimens without gypsum and 0,5 Mpa for the specimens with gypsum. To obtain greater differences for this type of mortar, probably it’s necessary to wait longer.

3 – Laboratory experimental program

Figure 3.1.31 – NHL 3,5_1:3 DPT: fcDPT VS Time; a) Specimens with gypsum; b) Specimens without gypsum; c) fcDPT VS Time__with gypsum VS without gypsum.

Comparing the results of tests with and without gypsum, (Figure 3.1.31.c and Table 3.1.20), we observe that the compressive strength average with gypsum is always higher than without, and there is an average ratio of:

In the second age the specimens with gypsum had average strength 49% higher than samples without gypsum. In fact in this session it was very difficult to get specimens with flat and parallel faces, and it was necessary to put a gypsum layer thicker than 1mm, to make testable the specimens according to DIN 18555-9. So the gypsum had a great influence on the compressive strength.We think that the influence of the plaster has several reasons:  Fill the microcavities on the specimens’ surfaces;

3 – Laboratory experimental program  It makes the surfaces plane and parallel, and the load from punches is more evenly distributed;  Creates a confinement on the specimens, which increase the compressive strength. Table 3.1.20 – DPT_NHL 3.5_1:3 Ratio of Cp. Strengths averages between specimens with and without gypsum.

For kerakoll mortar M5, all data of DPT compressive strength are (Table 3.1.21): Table 3.1.21 – Kerakoll mortar M5: DPT on square specimens, all results.

3 – Laboratory experimental program In figure 3.1.32.a and 3.1.32.b, an anomalous variation of strength over time for the two types of test is observable. More constant and linear growth was expected, but after 28 days the average strength decreases very much. Probably we didn’t make enough attention to the mortar hardening process: in a controlled environment like the laboratory, where the climatic conditions are sufficiently controlled we didn’t cover the wallets in the early days of maturation. The rules of UNE-EN 1052-1, section 7.2 were followed: “...To prevent the rapid loss of moisture in the first days after construction, cover the specimens (wallets) with sheets of polyethylene, or leave the wallets discovered in a lab environment...” The specimens were left discovered, but probably the heating air in the laboratory negatively affected the hardening process of the mortars in each wall. So it was not possible to observe a consistent resistance evolution of the joints over time. In the second experimental campaign (section 3.2 of this thesis) the specimens were covered.

3 – Laboratory experimental program

Figure 3.1.32 – Kerakoll mortar M5_DPT: fcDPT VS Time; a) Specimens with gypsum; b) Specimens without gypsum; c) fcDPT VS Time__with gypsum VS without gypsum.

In figure 3.1.32.c and table 3.1.22 a comparison between the strengths of the specimens with and without gypsum is shown. Table 3.1.22 – DPT__Kerakoll mortar M5 Ratio of Cp. Strengths averages between specimens with and without gypsum

3 – Laboratory experimental program The average ratio (Table 3.1.22) between the resistances is lower than for the other mortars, probably because the joints with kerakoll mixture are better removable and they have more regular surfaces, so the gypsum regularization effects are lower than the effects on weaker mortar, with more microcracks on the surfaces. In fact kerakoll mortar bound less with the bricks, and after removal, the joints are less damaged.

At a later time it will be interesting to compare the data of joints compressive strength with the data concerning the minor-destructive tests. For CL90_1:3 “Ciaries” with powder aerial lime, all data of DPT compressive strength are (Table 3.1.23 and figure 3.1.33):

Table 3.1.23 – CL90_1:3 “Ciaries”: DPT on square specimens, all results.

3 – Laboratory experimental program

Figure 3.1.33 – CL90_1:3 “Ciaries”_DPT: fcDPT VS Time; a) Specimens with gypsum; b) Specimens without gypsum; c) fcDPT VS Time__with gypsum VS without gypsum

Observing the charts (figure 3.1.33.a and 3.1.33.b), it can be noted that the average compressive strengths are not very high, but they linearly increase, and the values vary slightly in time. It is a weak material and probably needed a more large maturing time. The average resistances of specimens tested with gypsum are constantly higher than averages without gypsum (figure 3.1.33.c). For CL90_1:3 “Segarra” with slaked lime, all data of DPT compressive strength are (Table 3.1.24 and figure 3.1.34):

3 – Laboratory experimental program Table 3.1.24 – CL90_1:3 “Ciaries”: DPT on square specimens, all results.

Figure 3.1.34 – CL90_1:3 “Segarra”_DPT: fcDPT VS Time; a) Specimens with gypsum; b) Specimens without gypsum; c) fcDPT VS Time__with gypsum VS without gypsum

3 – Laboratory experimental program The behaviour is the same to that observed for mortar CL90_1:3 “Ciaries”. For mortar CL90_1:3 “Segarra” (figure 3.1.34.a and 3.1.34.b), the average compressive strengths are not very high, but they linearly increase and the values slightly differ in time. It’s a weak material and probably needed a more large aging time. The average resistances of specimens tested with gypsum are constantly higher than averages without gypsum (figure 3.1.34.c). Comparing the results of tests with and without gypsum, for the aerial mortars (Figure 3.1.33.c, 3.1.34.c and Tables 3.1.25.a and .b, we observe that the average compressive strength with gypsum is always higher than without, and there is an average ratio of:   Table 3.1.25 – DPT_ Ratio of Cp. Strengths averages between specimens with and without gypsum; a) CL90_1:3 “Ciaries”; b) CL90_1:3 “Segarra”

The behaviour is similar to that observed for the hydraulic lime mortar NHL3.5_1:3, which has more or less the same compressive strength (Table 3.1.20). This means that for weak joints, with a compressive strength around 1 MPa, the plaster layers (1mm thick) applied on the loading faces, increase 25% the average strengths. For kerakoll mortar M5 (Table 3.1.22) the average ratio between the two types of resistances is lower: the specimens treated with gypsum paste have a compressive strength 15-17% higher. For weak mortar joints, the extraction ruined the samples, and the influence of the plaster (which has a considerable compressive strength) is higher. In the second experimental campaign it was used another method, applying gypsum powder between the punches and the samples’ loading zones (section 3.2 of this thesis).

3 â&#x20AC;&#x201C; Laboratory experimental program

3.1.7 Helifix screw pull-out tests on wallet joints On the same joints tested in the DPT, the pull-out tests were performed using the Helifix screw pull-out system, with the aim to find a correlation between the extraction force of the Helix and the mortar compressive strength. Not having been performed a lot of research on this topic, another aim is to learn the best ways to perform the tests on different types of materials (bricks and mortars). For example in scientific papers like Ferguson 1994, investigators claimed that is better to use guide hole for the helix insertion with diameter < 4mm, when the mortar is weak. It is interesting to understand how the diameter affects the results. Another aim is to understand the differences between vertical and horizontal joints behaviours, at least in respect to extraction force, because it is difficult to adequately test the vertical joints with direct compressive tests, due to some difficulties (is difficult to obtain specimens 5x5cm, as it has been done for horizontal joints, and it is preferable to not introduce another error). The operational phases were the following (Figure 3.1.35): 1) The helices were halved hearts with a pincer (Figure 3.1.35.a); 2) The pilot holes for the helices were made, with a screw drill of 4mm in diameter (or 3mm in diameter) (Figure 3.1.35.b); 3) The helices were inserted in the mortar joints. For the insertion, a support able to adjust the insertion depth was used, and the helices were always inserted to 3 cm depth. This support is a tool different from the standard tool provided by the company, which is appropriate to the insertion of a helix part (length 3 cm). In the standard way, before the test it is necessary to cut the helix with a weighty pincer, which may disturb the area of mortar to be examined. So the data acquired may contain a very high error factor, which can disappear by cutting the helix before the test. The regolable support must be banged with an hammer and the helix pulls in the material translating and rotating (Figure 3.1.35.c); 4) The support necessary to the helix extraction was gently applied (Figure 3.1.35.d); 5) The helices were pulled out within the appropriate system, making slowly 1/3 lap at time. It was read the value indicated by the red arrow, which gives the instantaneous force, so is able to give the maximum force (Figure 3.1.35.e);

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.1.35 - Helifix screw pull-out: operational phases - a) Cut in half the helices with a pincer; b) Make the pilot hole with a screw drill (diameter 4mm or 3 mm); c) Insert the helix in the mortar with a regolable support; d) Gently applie the support necessary to the extrection; e) Pull out the helix within the appropriate system making slowly 1/3 lap at time.

Figure 3.1.36 - a) Helices on horizontal joints; b) Helices on vertical joints.

For each test session the data were collected in a table, where the thickness and the number of the tested joint were indicated. Numbering starts from the top downwards. Many tests were made, both on horizontal joints and on vertical joints (Figure 3.1.36). A large number of tests 85

3 – Laboratory experimental program is useful because of the high experimental dispersion. In the second experimental campaign the vertical joints were more carefully studied, because in the first campaign a lot of tests on vertical joints were not performed. An example of table is in Table 3.1.26: Table 3.1.26 – Example of test session on NHL 3.5_1:3 mortar

In the Table 3.1.26, in the column called joint (v) means vertical joint. A larger number of tests was made on horizontal joints because of the more available space for this type of test; in fact there are only 8 short vertical joints for each wallet. The tension of extraction in the bond area helix-mortar is:

Where η=tension; d=helix’s diameter (6mm); h= penetration depth of the helix (30mm). For the NHL 3,5_1:3 mortar, the average extraction forces for all ages are in table 3.1.27, and in figure 3.1.37 the chart Extraction force VS Time for the same mortar is shown.

3 – Laboratory experimental program Table 3.1.27 – All data, helifix screw pull-out on mortar joints (NHL 3,5_1:3)

Figure 3.1.37 - Helifix screw pull-out Mortar NHL3.5_1:3; Extraction force VS time

3 – Laboratory experimental program Unfortunately, the helifix screw pull-out results on the walls of mortar NHL3,5_1:3 are not satisfactory. There isn’t a linear growth over time of the pull-out force, as happened for the compressive strength on joints (DPT section 3.1.6). In the first 3 ages of maturation

remains more or less constant, but truly the test

procedures were always different, in order to get acceptable data. Often the outer part of the joints were too dry, the mortar was very friable and the holes made with the drill disturbed very much the material, and the helix didn’t bond with the mortar.  AGE1 (Helifix screw pull-out NHL3,5): it was possible to follow the standard procedures, performing a pilot hole with diameter ϕ=4mm, and entering the helix to a depth of 3 cm. The data of extraction force have never given result 0, but the experimental dispersion was quite high, CV=33%.  AGE2 (Helifix screw pull-out NHL3,5): it wasn’t possible to follow the standard procedures. Repeating the test in the same way like AGE1, it was got almost always result=0: mortar was very dry and friable; the helices didn’t bond with the mortar. Probably it was a problem of the external part of the mortar, and before the tests the outer more dry 2 cm of mortar was removed, because of the non-optimal maturation (wallets not covered with plastic film, and heating system sent hot air to the walls).  AGE3 (Helifix screw pull-out NHL3,5): The same procedures of age 2 were followed. Between ages 2 and 3, it has been a month, but the extraction resistance values did not vary. However the compressive strength data remained more or less constants in ages 2 and 3, so we cannot expect significant changes for the pull-out data.  AGE4 (Helifix screw pull-out NHL3,5): this time the outer parts of the joints weren’t removed, as done for the wallets of AGE 2 and 3. The tests always made drilling previously holes of in depth 3cm and 3 mm in diameter. The helix is diameter 6mm in diameter. It was unexpectedly found an average resistance

very much higher than

, although for the compressive strength of the joints there aren’t substantial differences between AGE3 and AGE4. Probably the fourth wall, being in a different position, suffered less the heating air then the wallets of AGE 2 and 3, so the outer part was less dry and damaged. The inner part of the joints, tested in Double Punch Test, matured in the same way of all wallets, and probably for this type of mortar 110 days are not enough to have a complete evolution of the compressive

3 â&#x20AC;&#x201C; Laboratory experimental program strength and comparing two ages, there arenâ&#x20AC;&#x2122;t big differences. The results will be better explained in chapter 4. At least in the implementation phase, the tests on kerakoll mortar gave fewer problems. It is a very durable mortar, and it is made with precise factory processes. However, the results on the extraction resistance are not very consistent and comparable with those obtained from the joints compressive tests. The results of Helifix screw pull-out on kerakoll mortar are (Table 3.1.28 and Figure 3.1.38). Table 3.1.28 â&#x20AC;&#x201C; Kerakoll mortar M5: helifix screw pull-out on mortar joints, FEXT (all data)

3 – Laboratory experimental program

Figure 3.1.38 – Helifix screw pull-out Kerakoll mortar M5; Extraction force VS time

As it can be seen in the chart (Figure 3.1.38), there is a growth of the average extraction force in time, and there is the same averages at ages 3 and 4: the resistance stabilizes. In chapter 4 we will see that there isn’t a clear correspondence with the compressive strengths of the same joints.

3.2 Second campaign on lime mortars In the second experimental campaign on the mortars, 5 wallets with “Terra Cuita bricks” were constructed, but only with “kerakoll mortar”, which was judge like the best-one for this type of experiments.

3.2.1 Materials

Figure 3.2.1 – Material used in experimental campaign 2; a) Kerakoll mortar “Biocalce Muro”; b) Terra Cuita bricks

3 – Laboratory experimental program The material used is: 1) Kerakoll mortar (Biocalce Muro) (Figure 3.2.1.a): Biocalce Muro is a product supplied by Kerakoll S.p.a., an Italian company that operates in the construction sector. According to the data sheet of the product, it is a mortar made with hydraulic lime NHL 3,5, and the strength class is M5, according to EN 998-2, or rather the compressive strength according to EN 1015-11 (compressive test on prismatic specimens after 28 days from molding) is 5 MPa. All the informations about the mortar are in the section 3.1.1 of this thesis. 2) Bricks: “Terra Cuita Piñol Pallarés S.L.” (Figure 3.2.1.b): The units used in this research are handmade Terra Cuita bricks, coming from the company “Terra Cuita Piñol Pallarés S.L.”, Spain. The nominal dimensions of the units are 345x145x45 mm. It is important to point out that there is a large variability of the bricks’ dimensions due to their peculiar construction procedure. From the product data sheet supplied by the company, we can find the compression normalized strength average, calculated according to UNE-EN-772-1. It is:

3.2.2 Construction of wallets and specimens As regards the construction of the wallets, the masons’ procedure has been the same to that adopted by the mason in the first experimental campaign (see in section 3.1.1). Summary: -

Halve the bricks using a circular saw;

Immersing and keeping the bricks under water for 15÷20 min: the bricks must be not too dry, otherwise the mortar loses quickly the moisture;

Preparing basement (in order to have enough space to perform penetrometer test we decided to build the walls 30 centimeters upper than ground. Walls were built on wooden beams);

During walls building vertical and horizontal direction were leveled;

Walls dimensions 690x420x145 mm ≈;

However, the masons made a serious mistake regarding the water/mortar ratio, not respecting the instructions provided by the company, which were: 4,4 l of water / 25 kg of mixture

3 – Laboratory experimental program Instead they put enough water to obtain workable mortar, but the same quantity for each amassed, so lower performances were expected from the mortar.

Figure 3.2.2 – Sheets of polyethylene around the wallets, according to UNE-EN 1052-1.

We followed the rules of UNE-EN 1052-1, section 7.2: “...To prevent the rapid loss of moisture, in the first days after construction, cover the specimens (wallets) with sheets of polyethylene, or leave the wallets discovered in a lab environment...” However, this time, to avoid the mistakes made in the previous experimental campaign, the wallets were covered with polyethylene sheets (Figure 3.2.2) in the first two days of maturation, so more homogeneous behaviour and less experimental dispersions were expected, because all joints mature in the same conditions (in the first days is very important to have the same conditions). Remember that in the previous experimental campaign, the wallets were left in the laboratory without covering by sheets of polyethylene, and the environmental conditions differently affected each specimen. As regards the prismatic specimens, only the prisms molded in steel were made according to BS-EN 1015-11. For each wallet 3 mortar prisms were made with a steel mold, in order to perform: 3 flexural tests; 6 compressive tests.

3 – Laboratory experimental program

3.2.3 Flexural tests on prismatic specimens of mortar The flexural strength was evaluated according to the EN 1015-11:2007, with (3.1) described in section 3.1.3. For each test session the data were collected in a table, where the load speed [Mpa/s] and the time taken to reach the breakage were wrote (Table 3.2.1): Table 3.2.1 – Example of flexural test session (Campaign 2)

For Kerakoll mortar used in the second experimental campaign, the flexural strength data are (Table 3.2.2): Table 3.2.2 – All data from flexural tests on Kerakoll mortar M5

The following chart (Figure 3.2.3) shows the the flexural strength evolution over time:

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.2.3 â&#x20AC;&#x201C; KERAKOLL mortar M5; Campaign 2; Flexural strength VS Time

As reported in (Figure 3.2.3), the behaviour of the flexural strength shows an increase in the first 23 days, and a subsequent little decrease. However the values of bending strength were higher in the first experimental campaign and this was expected, because the bricklayers putted more water than needed.

In both of experimental campaign there is a decrease of flexural strength, after the peak. A similar trend was observed also in the same test with a different mortar mix, by (Baronio, 1999; Witt, 2014), in the framework of the MICROPAR project. The reason of such a decrease of strength is still not clear at all and might be related to ageing of mortar, deserving to be better investigated in further research.

3.2.4 Compressive tests on prismatic specimens of mortar The compressive strength was evaluated directly by the compressive test of the two resulting parts of the mortar prisms, according to the EN 1015-11:2007 (see section 3.1.4 for more details). The compressive strength

[

] is calculated with equation (3.2) (see section

3.1.4). Like done in the first experimental campaign, for each test session they were collected the data in a table, where they were wrote the speed of load [Mpa/s], and the time taken to reach the break, trying to get it between 30 and 90 seconds (Table 3.2.3):

3 – Laboratory experimental program Table 3.2.3 – Example of compression test data_Kerakoll mortar “Biocalce muro”

For Kerakoll mortar, used in the second experimental campaign, the data of compressive strength are (Table 3.2.4): Table 3.2.4 – Kerakoll mortar, Campaign 2_Compressive strength, all data

The graph that shows the evolution of compressive strength is the following (Figure 3.2.4):

3 – Laboratory experimental program

Figure 3.2.4 – Kerakoll mortar "Biocalce Muro" M5; Campaign 2_Cp. Strength VS time

The evolution of compressive strength doesn’t show the same trend of the flexural strength. Already after 19 days there is a slight decrease. As expected, because of an error of the builders who putted too much water in the mortar, the compressive strength doesn’t reach the guaranteed value of 5 Mpa reported on the product’s data sheet (Biocalce Muro is a mortar of class M5 according to EN-998-2). Instead the maximum reached value is 4 MPa (19 Days).

3.2.5 Double Punching Tests on mortar joints The punching tests were performed on the specimens obtained from each wallet, on which the minor-destructive tests were performed. The aim is to obtain two series of well comparable data, because destructive and minor destructive tests were performed on the same materials (the joints). UIC 2011, DIN 18555-9 1999, and Binda et al 2002 cover the specifications for this test. Specifically, the rules indicated in chapter “5. Testing of mortar from masonry”, section “5.2 Method III” of DIN 18555-9, were followed: “Specimens approximately 50 mm square or 50mm in diameter are prepared from samples taken from masonry (e.g. by core drilling or masonry unit/joint assemblies). The specimens are then placed between a pair of loadingplatens measuring 20mm in diameter and tested for compressive strength.” For more details, see at the section 3.1.6. DIN 18555-9 standard suggests using specimens with parallel faces and whenever this is not possible it recommends putting a layer of gypsum with a thickness lower than 1mm. 96

3 – Laboratory experimental program Instead to prepare gypsum in the most common way, by mixing with water, it was tried another method, applying some gypsum powder between the specimen’s surfaces and the punches (Figure 3.2.5):

Figure 3.2.5 – DPT, gypsum powder on specimens surfaces. a) Gypsum powder on lower punch; b) Gypsum powder on upper specimen’s surface; c) Typical failure mode of a specimen treated with gypsum; d) Frequent failure mode of an untreated specimen.

This method was used because it was assumed that the gypsum powder rectifies the load filling the microcavities on the specimen surfaces. The particles of the gypsum can move during the load application, creating two optimal and parallel surfaces, so the load from the punches will be well distributed. Moreover due to its nature the powder will not bound with the specimens, and will not cause any confinement effects. With this method the measuring of the joints compressive strength was improved, and the experimental dispersions decreased very much. 12 tests with and 12 without gypsum powder were made, examining each joint. For each wallet data table like 3.2.5 was made: 97

3 – Laboratory experimental program Table 3.2.5: Kerakoll mortar “Biocalce Muro”_Example of DPT test session (AGE 4)

The nomenclature provides the informations about the tested specimen. For example: 

→ kerakoll mortar, with hydraulic lime, class M5 according to EN 998-2;

 DPT → Double Punch Test according to DIN 18555-9 (section 5.1, Method III);  GY → specimen with gypsum powder;  age4 → fourth age of maturation;  1.1 → joint n°1 (from the top), specimen 1. Figures 3.2.5 a) and b) show the procedure before the tests (how apply gypsum powder). Figure 3.2.5.d shows a frequent type of breakage for the specimens without gypsum, with a single longitudinal fracture that splits the specimen. Even if we choose samples with parallel faces, the load is applied on the specimen with non-full contact, and the edges of the punches

3 â&#x20AC;&#x201C; Laboratory experimental program cut the samples in this way. The compressive strength is underestimated. This type of failure never happened when gypsum powder was putted (Figure 3.2.5.c). Testing specimens with gypsum, the load is distributed more uniformly, and this is also known observing the rupture type. There will be a central zone who breaks forming an hourglass (Figure 3.1.27.c), and the surrounding part, where the cracks propagate. For Kerakoll mortar, used in the second experimental campaign, all data about DPT are in Table 3.2.6: Table 3.2.6 â&#x20AC;&#x201C; Kerakoll mortar M5, DPT, Campaign 2, all data.

Observing the experimental data, it is clear that at any age the experimental dispersion decreases almost constantly, comparing tests with and without gypsum. The coefficient of variation CV decreases by 7% for almost all tests, and for the last test decreases by 5%. Remember that:

The graphs are in Figures 3.2.6. There is continue and rapid growth of the joints compressive strength (Figure 3.2.6.a, 3.2.6.b). In the second experimental campaign it was chosen performing test at very close ages, because this type of mortar matures very quickly, and in this way maybe it was possible to get very distinct resistance values.

3 â&#x20AC;&#x201C; Laboratory experimental program For the specimens tested with gypsum powder (Figure 3.2.6.b) a little decrease at 23 days was observed, then the compressive strength backs to grow. The graph in Figure 3.2.6.c shows a comparison between the two types of compressive strength (with and without gypsum). Growth over time is similar for the two experimental curves, but the resistances averages on specimens treated with gypsum powder are higher.

Figure 3.2.6 - Kerakoll mortar M5, DPT: Cp. Strength VS Time; a) Specimens with gypsum; b) Specimens without gypsum; c) Specimens with gypsum powder VS Specimens without gypsum powder.

The Table 3.2.7 shows the ratios between the compressive strength averages (DPT with gypsum powder and DPT without gypsum powder).

100

3 – Laboratory experimental program Table 3.2.7 – Kerakoll mortar, Campaign 2; Ratio between compressive strength (with/without gypsum)

So using gypsum powder the compressive strength average increased by 11÷13%, while using gypsum paste increased by 17% (see section 3.1.6 of this thesis). Both methods are efficacious for the specimens’ regularization:  The load is uniformly distributed, and this can be seen observing the rupture mechanism. There is always a central hourglass and cracks that propagate in all directions of the sample (Figure 3.2.5.c);  Compressive strength average increases compared with specimens untreated by gypsum. Even if we choose specimens with parallel faces, on untreated samples the load is applied with a non-full contact, and often the edges of the punches cut the specimens in half (Figure 3.2.5.d). The compressive strength is underestimated. This type of failure never happens when gypsum powder is applied.  For the same reason the experimental dispersions decrease. The data on the compressive strength are more reliable, and they will be chosen for the comparison with the data of minor-destructive tests.  The technique using gypsum powder seems better than the technique using gypsum paste for several reasons: the powder is a material that regulates without creating confinement, and it doesn’t affect the material’s resistance. This is not true applying gypsum paste. The powder is a dry material, which doesn’t change the sample moisture content. Instead, with gypsum paste the moisture content changes and before testing is

101

3 – Laboratory experimental program necessary to leave the specimen in a climatic chamber until reaching the ideal moisture content, or rather constant mass: “...Constant mass is deemed to have been reached if the change in mass is not more than 0,1% within 24h.” (DIN 18555-9). Considering the time required for the preparation of the specimens with gypsum paste, and the hardening time in the climatic chamber, the technique, applying gypsum powder, is more rapid and convenient and provides satisfactory results.

3.2.6 Helifix screw pull-out tests on wallet’s joints The procedures of this type of experiment are explained at section 3.1.7 of this thesis. In the second experimental campaign, another aim is to understand the differences between results of extractions carried out making guide holes of 3mm in diameter and making guide holes of 4 mm in diameter. The question is: How the hole affects the bound helix-mortar, and the extraction force? The comparison was made only on the horizontal joints, while for vertical joints was made a comparison only with guide hole of 3 mm, to understand the differences between the bahaviour of vertical and horizontal joints. For each test session a table like the following was made (Table 3.2.8):

102

3 â&#x20AC;&#x201C; Laboratory experimental program Table 3.2.8 â&#x20AC;&#x201C; Kerakoll mortar M5, Helifix screw-pull-out, Campaign 2, Age 4

Where the different nomenclatures mean: -

readm.(hor.): is the mean (m.) of the readings made for all horizontal joints, without distinguishing the tests done practicing different guide holes 3 mm and 4 mm in diameter.

readm.(hor.4mm): is the mean (m.) of the readings made for all horizontal joints, practicing guide holes of 4 mm in diameter.

readm.(hor.3mm): is the mean (m.) of the readings made for all horizontal joints, practicing guide holes of 3 mm in diameter.

readm.(ve.3mm): is the mean (m.) of the readings made for all vertical joints, practicing guide holes of 3 mm (to test the vertical joints they were always made guide holes of 3mm in diameter).

Similarly, for each type of test, the corresponding CV (coefficient of variation) and SD (standard deviation) were defined. Observing the Table 3.2.8 and the other data tables for each age of maturation, it can be understand that the scatter without distinguishing tests done with different guide holes is very high. So it is clear that the guide hole is a factor that greatly affects the results of the tests. The experimental dispersions decrease distinguishing tests done with different guide holes, and it is minimal when the tests are done practicing guide holes of 103

3 – Laboratory experimental program 3 mm in diameter. It can be assumed that for weak mortars is better to make guide holes of 3 mm in diameter, disturbing less the testing zone. Experimental dispersion is higher for vertical joints than for horizontal joints, maybe because the compaction in the construction phase is different and the strength depends by several factors. Tests with guide holes of 3 and 4 mm in diameter have been made in different and random zones, trying to get a good comparison between the two ways for test, testing each joint. So they were chosen closed ages of maturation, trying to get distinct strength values, because in the first experimental campaign it was clear that Kerakoll “Biocalce Muro” mature very quickly, and if we waited too long for testing, we couldn’t get distinct values of resistance. In fact after 28 days the resistance more or less stabilizes for this type of mortar. All procedures are described at section 3.1.7. So a large number of test was made at ages 6, 13, 19, 23 and 35 days. All results for horizontal joints are shown in the following table (Table 3.2.9). Table 3.2.9 – Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out, horizontal joints, all results.

It can be observed that the average extraction resistance is always higher for tests performed with the guide hole of 3 mm in diameter, while the experimental dispersions are higher in the tests performed with the guide hole of 4mm in diameter. For this reason, we can think that for the weak mortars it is preferable to use a guide hole of 3 mm, as the on of 4 mm, that disturbs

104

3 – Laboratory experimental program too much the interaction zone mortar-helix (experimental dispersion generally higher, some values are too much lower than the average). We will see that instead for the bricks, is preferable to use the guide hole of 4mm, because with the hole 3 mm in diameter is difficult to insert the helix and the resistances are overestimated. The experiments with guide hole of ϕ=3mm in diameter provide higher pull-out resistances than with ϕ=4mm. In Both experiments the pull-out force early reaches the peak and then stabilizes. In Figure 3.2.7.a and 3.2.7.b they are shown the evolution of the pull-out forces over time, respectively for tests performed with guide hole of 4mm in diameter and hole of 3mm in diameter. In Figure 3.2.7.c there is a comparison “Pull-out force VS time” in the two ways to test: the trend is similar, but the resistances with guide holes of 3 mm in diameter are higher. In the first age (6 days) it wasn’t possible to test practicing holes 4mm in diameter, so the comparisons were made only in 4 ages of maturation. Although it was chosen to test at very close ages, it wasn’t possible to obtain very distinct pull-out force values, and there isn’t a big difference between the values obtained at any age, especially for tests performed with guide holes of 4 mm in diameter were the averages are more or less equal.

105

3 – Laboratory experimental program

Figure 3.2.7 – Kerakoll mortar M5, Exp. Campaign 2_Helifix pull-out tests: FEXT VS time a) Holes made with screw drill ϕ=3mm; b) Holes made with screw drill ϕ=4mm; c) Comparison between FEXT VS time, ϕ=3mm VS ϕ=4mm.

Table 3.2.10 – Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out Comparison between extraction forces with different guide holes (3 and 4 mm).

106

3 – Laboratory experimental program In the Table 3.2.10 there is a comparison between the averages obtained testing the two different ways: the ratio between two correspondents resistances show that the pull-out force is averagely 25% higher in the tests performed with guide holes of 3mm in diameter. -

FEXT.3 → Pull-out force of helices in tests performed practicing guide holes of 3mm in diameter;

FEXT.4 → Pull-out force of helices in tests performed practicing guide holes of 4mm in diameter.

The data regarding the extraction forces from vertical joints are collected in Table 3.2.11. It is clear that the averages decreased but the experimental scatter increased, comparing with the horizontal joints. Table 3.2.11 – Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out, vertical joints: all results.

In Figure 3.2.8.b the graph “Pull-out force VS Time” for vertical joints is shown, in Figure 3.2.8.a the graph “Pull-out force VS Time” for vertical joints, and in Figure 3.2.8.c the comparison between “Pull-out force VS Time” for vertical and horizontal joints (very similar trend, higher averages for horizontal joints). The comparison is made on tests performed practicing 3mm in diameter holes.

107

3 – Laboratory experimental program

Figure 3.2.8 - Kerakoll mortar M5, Exp. Campaign 2. Helifix screw pull-out (screw drill ϕ3mm); a) Fext on horizontal joints; b) Fext on vertical joints; c) Horizontal VS Vertical joints.

With the aim to understand how the pull-out force decreases for the vertical joints, it was calculated (Table 3.2.12) the ratio FEXT.ve/FEXT.hor for each age of maturation, and some statistical parameters of this ratio (average, standard deviation SD and coefficient of variation CV). -

FEXT.ve → Extraction force of helices from vertical joints (guide holes 3 mm in diameter);

FEXT.hor → Extracton force of helices from horizontal joints, considering only the tests performed practicing holes 3 mm in diameter.

108

3 â&#x20AC;&#x201C; Laboratory experimental program Table 3.2.12

â&#x20AC;&#x201C; Kerakoll "Biocalce Muro", Campaign 2, Helifix screw pull-out

Comparison between extraction forces of vertical and horizontal joints

There is a nearly constant ratio between the two extraction forces, and it can be assumed that FEXTve is more or less 75% of FEXT.hor (the very low scatter, with CV=4.6% makes this estimate reliable). For now, this can be said only for this type of mortar, but it is a data that deserves attention when similar tests on other types of mortar and on real buildings will be made.

3.2.7 Windsor Pin tests on wallet joints In the second experimental campaign, they were made also penetrometer tests with the Windsor Pin System, with the aim to obtain a correlation between the pin penetration depth into the joint, and the compressive strength of the same joint. The minor destructive test techniques and the destructive techniques were performed on the same material. This method is very effective because the correlations will be less affected by errors. The technique has also been used by other authors to calibrate tools that carry out minor-destructive tests in site. For example in F.Bovio, G.Bovio 2013 is described a method for the calibration of an instrument that perform the Lok test on existing structures of concrete. The Windsor pin tests were performed only on horizontal joints, because the vertical joints have been subjected only to pull-out tests and to a comparison with the values obtained in the pull-out tests on horizontal joints. In fact there were few and small vertical joints, and they did not offer much space for the execution of many tests, so it was made this choice. The Windsor-Pin System and its functioning are described in section 2.3.1 of this thesis, and in this section they were described the specifically operational procedures conducted in this research (Figure 3.2.9):

109

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.2.9 - Kerakoll mortar M5 "Biocalce Muro", Exp. Campaign 2, Windsor Pin test procedure; a) Pin shooting into the joint; b) Pin partly penetrated into the joint; c) Cleaning hole with the pipette; d) Measuring Pin penetration depth.

For each wallet a great number of tests on horizontal joints has been made, because the scatter is generally high. For each test session a table like in figuure 3.2.13 was made. The tests were performed on all joints (six in total), progressively numbered from top to bottom. The micrometer (Figure 3.2.10) measures the part of its tip that remains outside, and the measure is expressed in inches. So the pin penetration depth is calculated by the equation 3.8.

Figure 3.2.10 â&#x20AC;&#x201C; Micrometer of Windsor pin system; a) Reading of the micrometer tip that remains outside; b) Micrometer.

110

3 â&#x20AC;&#x201C; Laboratory experimental program [

]

(1 - read) because the micrometer measures the part of its tip that remains outside;

24.5 because 1 inch = 24,5mm. Table 3.2.13 â&#x20AC;&#x201C; Kerakoll "Biocalce Muro", Campaign 2, Windsor pin tests Typical test session on a wallet: at least 15 tests each wallet

111

3 – Laboratory experimental program All data of Windsor pin experiments are collected in Table 3.2.14: Table 3.2.14 – Kerakoll "Biocalce Muro", Campaign 2, Windsor pin test: all data

The scatters are not very high, with CV between 11% and 15%, and they are lower than the experimental scatters of screw pull-out tests on horizontal joints, with CV between 13% and 20% (Table 3.2.9). Thanks to the adopted experimental procedures (flat and well testable joints, at the same level of the bricks’ outer parts), it can be assumed that the data obtained from this tests are very reliable and affected by few experimental uncertainties.

Figure 3.2.11 – Kerakoll mortar M5, Experimental Campaign 2; Windsor Pin test; P.depth [mm] VS Time [days]

112

3 – Laboratory experimental program The mortar compressive strength should increase over time, so the pin should penetrate less deeply because the mortar is stronger and this in fact happens (Figure 3.2.11). In Table 3.2.15 it is shown a comparison between the compressive strength obtained from the tables provided by the technical data sheet of the instrument and the experimental evidences from the laboratory tests in this experimental campaign. It’s clear that the tables provided by the company are not suitable for analysis on weak mortars. For example it can be noted that for high values of penetration depth, the tables underestimate the compressive strengths. For low penetration depth values the table overestimate very much the compressive strengths. Table 3.2.15 – Comparison between the correlations Pin-in VS Compressive strength provided by the instrument’s tables and the experimental values

Another aim of this thesis is to give an effective empirical law for weak mortar, to estimate the compressive strength from the pin penetration depth.

3.3 Campaign on bricks The experimental campaign on bricks was made to improve the experimental techniques to estimate the mechanical properties of the bricks. To improve the kinds of techniques that allow us to find information about the brick is interesting to test small brick portions. The compressive strengths of brick pieces will be then related to the compressive strength of whole bricks. In a building analysis is often not possible to extract big portions from the structural elements for laboratory testing, and getting important data from tests of small portions is the only solution. For example the current standards don’t give correlations between the compressive strength of whole bricks and the compressive strength of cylindrical pieces that can be extracted from the brick. The corrlation would be very helpful, because it is easy to extract material with a coring machine without affecting the resistance and the esthetic of the structure.

113

3 – Laboratory experimental program

3.3.1 Materials Four types of bricks were used to verify whether the techniques developed and the correlations are effective on different types of material. The used bricks are shown in figure 3.3.1.

Figure 3.3.1 – Tested Bricks; a) “Terra Cuita” bricks; b) Bon Dia bricks (unbaked bricks); c) “Piera_claro” bricks; d) Bricks from Puig’s house.

So the bricks are: a) “Terra Cuita” bricks: are handmade bricks used in the wallets building. They were provided by the Catalan company “Terra Cuita Piñol Pallarés”. The medium dimensions are (290x137x45) mm; b) Bon Dia bricks: are not baked bricks made by the “Bòbila Bon Dia” Company. They are very weak bricks and they are used in the baking process of the Terra Cuita Company, which produces “Terra Cuita” bricks. The dimendions are (290x137x45) mm;

114

3 – Laboratory experimental program c) Piera Claro bricks: stone bricks provided by the “Ceràmica Piera S.L.” Company. These bricks are studied in other thesis made in the UPC, like (Cristina Usan Caño_2014) and (Alice Paverini 2014). The dimensions are (276x133x43) mm; d) Bricks from Puig’s house: they are bricks (280x133x35) mm, taken from some structural elements of “Puig i Catafalch” house, a building under renovation in Argentona (a village near Barcelona). The study done on this building will be described in section 4 of this thesis.

3.3.2 Compression tests of units The compressive test of the units was made with the ibertest machine with load cell of 3000 KN (Fig.3.3.2.a and Fig.3.3.2.b). The tests were made in displacement control, following the UNE-EN-772-1 for all phases. Before the compressions, the bricks faces were cleaned. The graphs (Figure 3.3.3) show a strange behaviour of the bricks. With very high and unacceptable deformations (Fig.3.3.2.c and 3.3.2.d), the load continues to grow. It was observed a linear part in the curve “fc VS displacement” (Fig.3.3.3), and it was decided that the brick compressive strength is one the extreme of this segment.

115

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.3.2 â&#x20AC;&#x201C; a) Ibertest machine with load cell of 3000 KN; b) Test set-up; c) Terra Cuita failure mode; d) Bon Dia failure mode.

The compressive strength is calculated with (3.9):

Where fc* is the compressive strength [MPa]; Fmax is the maximum force corresponding to the end of the linear part of the curve [N]; b and l are respectively the width and the length of the brick [mm]. According to the annex A.1 of the 772-1, the normalized compressive strength is (3.10):

Where fc* is the compressive strength and d=0,7 is the shape coefficient.

116

3 – Laboratory experimental program

Figure 3.3.3 – a) fc VS diplacement “Terra Cuita” bricks; b) fc VS diplacement “Piera Claro” bricks; c) fc VS diplacement “Bon Dia” bricks;

Three tests for each type of brick were made, but no test were performed on bricks of Puig’s house (section 5 of this thesis), because the bricks are too thin and their faces are too much irregular. The averages in table 3.3.1 show three different behaviours for the three kinds of analized bricks.

117

3 – Laboratory experimental program Table 3.3.1 – Compressive tests on units: data and averages

To estimate unambiguously the compressive strength fc*, we propose a method in which the secant modulus of the bricks is used. The secant modul is the ratio of stress to strain at any point on curve in a stress-strain diagram. It is the slope of a line from the origin to any point on a stress-strain curve (Figure 3.3.4).

Figure 3.3.4 – Secant modulus in a diagram Stress VS Strain.

So, considering the ratio fc/δ (Figure 3.3.5) where δ [mm] is the displacement, for each brick the maximum compressive strength fc* is calculated with the formula (3.11). (

)

Where δ [mm] is the brick deformation in correspondence of the maximum ratio fc*/ δ. The results are in table 3.3.2, with the normalized compressive strength fc (formula 3.10).

118

3 – Laboratory experimental program In Figure 3.3.5 the graphs fc*/δ VS δ are shown. For “Terra Cuita” and “Piera Claro” bricks, the maximum value fc*/δ is detectable. The fc* and the normalized compressive strength fc (equation 3.10) are listed in Table 3.3.2.

Figure 3.3.5 – fc*/δ VS δ: a) Piera Claro bricks; b) Terra Cuita bricks; c) Bon Dia bricks.

Table 3.3.2 – Compressive tests on units, calculated from the ratio fc*/δ

119

3 â&#x20AC;&#x201C; Laboratory experimental program The average compressive strength (Table 3.3.2) is very much higher compared to that estimated observing the graphs in Figure 3.3.3. The method requires to be enhanced in future researches.

3.3.3 Compressive tests of cubic and prismatic specimens These tests were carried out for several reasons. It was searched an appropriate method for studying the brick anisotropy, and specimens along the three main directions were tested, as it was also made by other researchers (Sassoni, Mazzotti 2013).

Figure 3.3.6 â&#x20AC;&#x201C; Compressive directions; a) Direction Y, perpendicular to the bricks header face; b) Direction X, perpendicular to the brick stretcher; c) Direction Z, perpendicolar to the brick bed.

The compressive strength along the directions x and y (Fig 3.3.6.a and 3.3.6.b) could be useful for the comparison with the minor destructive tests data, because in a building is possible to performe helifix and windsor tests only on header and stretcher faces. Tests on cubic specimens are also useful because they allow getting many data on the compressive strength from a small amount of material. In this way a brick could be accurately characterized. In fact the compressive test on units (section 3.3.2) is affected by many uncertainties. The bricks faces were cleaned using a grinding machine (Fig.3.3.7.a) then a large number of samples were obtained cutting the bricks with a circular saw (Fig.3.3.7.b).

120

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.3.7 â&#x20AC;&#x201C; Prior operations on the bricks; a) Faces regularization by grinding machine; b) Getting cubic samples from cutting by circular saw.

It was very difficult limiting experimental scatter in the first trial tests. Mostly for specimens tested along the directions x and y it was very difficult getting plain and regular faces, so the load could not be uniformly distributed on the whole loading faces. In this way the compressive strength was underestimated. So they were tried three different methods to regularize the loading zones: a) Manual cleaning using abrasive material; b) Application of gypsum powder on the loading areas, with two aims: regularize the surfaces and limit the confinement effect (Fig. 3.3.8.b). c) Leveling using glue X60. The paste was putted on the specimenâ&#x20AC;&#x2122;s faces, surrounded by a small formwork. This formwork is a simple metallic paper around the specimen which keeps the glue on the face and allows creating a flat surface during the hardening phase, thanks to the gravity effect (Fig. 3.3.8.a); Before the tests, the specimen cross sections were measured using a caliper, and the compressive strength of the sample was calculated as follows:

Where fc is the compressive strength [MPa]; Fmax is the maximum force [N]; b and h are respectively the width and the length of the cross section.

121

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.3.8 â&#x20AC;&#x201C; Types of surface treatments; a) Formwork around the specimen; b) Specimens surfaces treated with glue X60; c) Compressive test in the ibertest (load cell 10 KN): gypsum application on the surfaces.

Six cubes for all type of brick were tested to compare the methods and to choose the best one. The method c) (application of gypsum powder) provided the more high compressive strengths. With the two methods for treating the samples, low experimental scatters were obtained (table 3.3.3). The experiments made with gypsum powder gave the lowest results, but experimental scatter was very low compared with the results obtained from method a). It can be assumed that the best testing method is applying a small amount of gypsum powder on the loading zones. The gypsum cancels the confinement effect (lower compressive strengths) and more uniformly distributes the load. This is the reason why the scatter is lower (Table 3.3.3): the local effects of non-perfect regularity of the loading zones are restricted.

122

3 â&#x20AC;&#x201C; Laboratory experimental program Table 3.3.3 - Compressive tests of cubic specimens (direction Z): data and averages

So for the tests in the other two directions x and y, the method b) (with gypsum) was used. The results are in Tab. 3.3.4. Table 3.3.4 â&#x20AC;&#x201C; Compressive tests of cubic specimens (directions X and Y): data and averages

It can be noted that the compressive strength tends to be higher for specimens tested along direction x. This is the direction perpendicular to the bricks strercher face, the face on which the majority of the minor destructive tests is performed.

123

3 – Laboratory experimental program

3.3.4 Compressive tests of cylindrical specimens Another type of useful test could be the compression of small cylindrical specimens. This kind of specimens can be easily obtained from building using a coring machine (Fig.3.3.9). The experiment was made only for a type of brick (Terra Cuita) and another aim is to compare the results with the compressive strengths datas of cubic specimens.

Figure 3.3.9 - Coring machine “Hilti”; extraction of cylindrical samples ϕ 35mm.

The experiments are performed applying gypsum powder on the upper face and under the bottom face of the cylinder. They were made tests for cylinder with h/d=1/1 and h/d=2/1 where h is the high and d is the diameter.

Figure 3.3.10 – Cylinders tested; a) Along the direction X, specimens 2:1; b) Along the direction Y, specimens 2:1; c) Specimens 1:1, direction Z, tested in the IBERTEST with load cell of 10 KN.

124

3 – Laboratory experimental program The experiments were made for each brick direction (Fig. 3.3.10.a and 3.3.10.b), but along the direction Z only specimens 1/1 are testable (Fig. 3.3.10.c). The results can be observed in Table 3.3.5. Table 3.3.5

– Compressive tests, cylindrical specimens 2:1, ϕ=35mm (“Terra Cuita” bricks)

The compressive strengths are the same for the two analyzed directions. The results are very different from the compressive strengths of the cubes. In the table 3.3.6 the results for the cylinders 1:1 (h=35mm, d=35mm) are shown. Table 3.3.6

– Compressive tests, cylindrical specimens 1:1, (h=35mm; ϕ=35mm);

a) Direction Z; b) Direction Y (perpendicular to header); c) Direction X (perpendicular to stretcher);

125

3 â&#x20AC;&#x201C; Laboratory experimental program The compressive strengths in direction X and in direction Y are the same, and in direction Z is slightly lower. This also happened for cubic specimens. The analyzed bricks have a low anisotropy, but it can be assumed that the direction Z is the weakest.

3.3.5 Flexural tests on prismatic specimens of brick The flexural tests on prisms were performed following the standard UNE -EN 772-6, which is usually used for mortar specimens (section 3.1.3 of this thesis). Prismatic specimens were cut from the bricks using a circular saw (Fig. 3.3.7.b) then a small incision at the center of the specimen was made, with the aim to localize there the failure. Rupture occurs when the ultimate tensile strength is reached. The directions of traction are X and Y (Fig. 3.3.11.a and .b), because these are in reality the directions along which the brick normally works. The tests were carried out using the IBERTEST with load cell of 10 KN (Fig 3.3.11.c) and rupture always occurred in the center (Fig. 3.3.11.d).

Figure 3.3.11 â&#x20AC;&#x201C; Flexural tests; a) Traction direction Y; b) Traction direction X; c) Support for the test in the IBERTEST (load cell 10KN); d) Failure mode (rupture of the middle section).

The flexural strength is calculated like:

Where:

[ ], ,

. 126

3 – Laboratory experimental program The flexural test results are in Table 3.3.7 (direction of traction y) and 3.3.8 (direction of traction x). Table 3.3.7 – Flexural strengths of all the kinds of brick. Direction of traction Y

The first three types of bricks tested, have very similar strengths (Table 3.3.7). The unbaked brick has a very low resistance, ten times lower than the other bricks. When the traction is in direction X (along the width), the averages are slightly higher for all types of brick (Table 3.3.8). Table 3.3.8 – Flexural strengths of all the kinds of brick. Direction of traction: X

3.3.6 Brazilian tests on prismatic specimens of bricks It was decided to perform the Brazilian test following the ASTM C496:1996, designed for concrete, and some available researches (Rocco et al. 1999).

127

3 – Laboratory experimental program The Brazilian test was performed on the cylindrical brick samples with dimension of (ϕ=35mm; h=70mm) and (ϕ=35mm; h=35mm), for all types of brick. The cylinders were extracted only from Terra Cuita bricks (used for the construction of the wallets). The machine used was the Ibertest with a load cell of 10 KN, and the load rate was chosen in order to get breaking in 60 seconds. An important characteristic of this test is the dimension of the strip of wood used to distribute the point load over the specimen. In this test the dimension of the strips was 3.5mm, which correspond at the 10% of the diameter of the specimen. The setup details are reported in Figure 3.3.12.

Figure 3.3.12 – Setup details: Brazilian tests of cylindrical samples 2:1

The failure mode showed a clear diametric fracture through the length of the specimen (Figure 3.1.23.b and 3.1.23.b). To calculate the tensile strength, knowing the maximum force obtained, the formula used is:

Where Fmax = maximum force [N], b = base of the section [mm], h = height of the section [mm]. An important operation was made before the samples coring. The bricks faces were accurately marked with vertical lines perpendicular to the tracting direction (see Figure 3.3.13). In this

128

3 â&#x20AC;&#x201C; Laboratory experimental program way, paying attention during the sample placing, we were sure to analyze the tensile strengths in the designated directions (x or y). To perform the test, two plywood stripes were diametrically attached on the upper and lower side of the specimen, to ensure the uniform application of the load along the length of the cylinders (Fig. 3.3.13).

Figure 3.3.13 â&#x20AC;&#x201C; Vertical lines on the header and stretcher faces to ensure the correct tracting direction; a) Specimens 2:1, tensile direction y; b) Specimen failure (traction y); c) Specimens 2:1, tensile direction x; d) Specimen failure (traction x);

To analyze how dimensions affected the results, two specimen ratios (height/diameter) were tested (2:1 and 1:1). However, for the cylinder extracted from the bed, only the ratio 1:1 was examined, and the directions of traction were x and y (Fig. 3.3.14).

Figure 3.3.14 â&#x20AC;&#x201C; a) Setup details: brazilian tests of cylindrical samples 1:1; b) All the tested specimens.

The results obtained for samples with ratio height/diameter=2:1 are shown in Table 3.3.9.a, while those obtained for samples with ratio 1:1 are in Table 3.3.9.b. 129

3 – Laboratory experimental program Table 3.3.9 – Tensile strengths ζ tx and ζ ty of the ϕ=35mm samples extracted from the bricks; a) Samples 2:1; b) Samples 1:1.

It can be observed that the cylinders 1:1 are more resistant than the cylinders 2:1. This didn’t happen for Piera Claro samples, analized in the same way by another student (Alice Paverini 2014).

3.3.7 Tensile tests on prismatic specimens of bricks In this research program was tried a method for uniaxial tension of a brick sample. Some guidelines of a previous research (Lourenço et al. 2002) were followed. The INSTRON machine with load cell of 425 KN was used. This machine is normally used for tensile tests of steel samples. Samples (35x35x70) mm were used, with two incisions at the central section. In the machine configuration, two vises keep and pull the supports. On this support two metallic plates were enganged (Figure 3.3.15.a). The plates were cleaned and fixed to the ends of the specimens by a resistant epoxy resin (Figure 3.3.15.b). The glue must be strong enough, and the rupture must occur at the center of the samples.

130

3 â&#x20AC;&#x201C; Laboratory experimental program

Figure 3.3.15 - Setup details: a) Configuration details of the INSTRON; b) Specimen fixed to the plates by the epoxy resin X60.

In the first attempt, there were always problems. The experiment is very difficult. Care should be taken to the plates moving. Often specimens rupture occurred because of errors during the plates approaching to the specimens ends. The INSTRON moves very big loads, very much higher than the specimen compressive strength. In the first attempts no experiment has been successful and another problem was the rupture in the glue zone. The amount of resin used for the first tests was low, and the bonding area specimen-glue was insufficient. So to solve these problems, the hydraulic jacks of the INSTRON were carefully moved and the bonding zone glue-specimen were increased thanks to a small handmade formwork (Figure 3.3.16.a) To avoid the rupture of the specimen before the test, we hold the bottom plate and we approach slowly the upper plate. Despite these expedients, there have been other problems during the experiments. So it was decided to reduce more the middle section area, to get break in the weakest section (Fig.3.3.16.a). Finally the method was effective and it was possible to successfully realize two tests (Fig.3.3.16.b).

131

3 – Laboratory experimental program

Figure 3.3.16 – Uniaxial tension of brick samples; a) Application of the glue; b) Successful experiment: breaking at the middle section.

The graph in figure 3.3.17 shows the relation η (tension) – displacement. The tension is calculated like:

Where T is the tensile force [N]; b and h are the side of the central cross section, accurately measured [mm].

Figure 3.3.17 – Terra Cuita samples; η VS displacement.

132

3 – Laboratory experimental program The maximum tension occurs with an elongation of approximately 0.3 mm. The test is performed in displacement control and the “post peak” is observable. Initially the sample is compressed because the upper plate is supported by the upper face of the specimen. The upper jack doesn’t move, while the lower jack moves and compression decreases to zero. After this step the specimen is in traction (Fig.3.3.17). The rate of displacement is very low (0,05 mm/min), to avoid the premature rupture of the specimen. Table 3.3.10 – Uniaxial tension tests: results

Unfortunately it wasn’t possible to make more then two tests. Observing the Tab. 3.3.10, it can be assumed a tensile strength η ≅ 1.4 MPa, a lower value than founded in the brazilian tests (η = 1,5÷2 MPa). The data of uniaxial tension test are insufficient, but the adopted method is effective. In future researches it will be possible to make more tests. The bricks have a low tensile strength and to improve the quality and the accuracy of the data, a load cell of 10 KN is recommended. In these tests it were used a load cell of 450 KN, more suitable for very resistant materials, like steel.

3.3.8 Helifix screw pull-out tests on bricks To make a comparison with destructive tests on brick and to compare the relation fc VS Fext for bricks and for mortar, it were made a large number of helifix screw pull-out test on the 4 available types of brick. For “Terra Cuita” bricks the tests were performed on the units of a wallet made with kerakoll mortar after three moths from his construction. In this way the mortar is strong and it ensure the stability of the bricks during the tests (Fig.3.3.18.a). For Piera claro bricks the tests were performed on a wallet from a previous experimental research (Alice Paverini 2014). Also in this wallet the units were stable (Fig.3.3.18.b). For the unbaked bricks “Bon Dia” it wasn’t possible to build a wallet, because the bricks fragment in contact with water. So the tests were performed keeping the units with a metal vice (Fig.3.3.18.c).

133

3 – Laboratory experimental program The pull-out tests were also made for the units of “Puig i Cadafalch” house. These data belong to a group of “in situ tests”, and they will better explain in the chapter 4 of this thesis. The pull out test was made on some constructive elements that belong to the house terrace (Fig.3.3.18.d).

Figure 3.3.18 – Helifix screw pull-out tests on the bricks; a) “Terra Cuita” bricks; b) “Piera Claro” bricks; d) “P y C” bricks.

For the results, see the following table (Tab 3.3.11):

134

3 – Laboratory experimental program Table 3.3.11 – Helifix screw pull out on bricks: results

It can be observed that for the three types of normal bricks, the pull-out forces are very similar and that is what also happens for the compressive strengths. Instead the unbaked bricks show very much lower pull-out strength.

3.3.9 Windsor Pin tests on bricks As regards the tests on “Terra Cuita” and “Piera Claro” bricks, the windsor pin tests were made on the same wallets of screw pull-out tests (see Fig. 3.3.19.a and 3.3.19.b). It was not possible to get reliable data for unbaked bricks “Bon Dia”, because the metallic vice is not sufficient to stabilize the brick: when the pin hits the brick, the unit moves and a not quantifiable percentage of power is lost. In the “In situ experimental program”, the tests were made on the bricks of an internal wall of the house, were a portion of plaster was removed to expose the bricks for the visual analysis (Fig 3.3.19.c).

Figure 3.3.19 – Windsor pin tests on the bricks; a) “Terra Cuita” bricks; b) “Piera Claro” bricks; c) “P y C” bricks.

135

3 – Laboratory experimental program The results are in Tab. 3.3.12. “Piera Claro” and “Terra Cuita” bricks have the same trend, but for the “P y C” bricks the value is different. The pin penetrates averagely 1mm more. This could be due to the fact that during the tests, the wallet slightly moves and a little percentage of pin power is lost. Another possible cause is the different response that different bricks, with the same compressive strengths, would have against this kind of test. Perhaps not only the compressive strength affects the penetration depth but also other factors such as the porosity, the molecular structure and the shape of the units. Table 3.3.12 – Windsor Pin tests on bricks: results.

136

4. In-situ experimental program It was chosen to make an experimental campaign on materials of an historical building, to verify if the results are comparable to the ones obtained from the experimental campaign in laboratory. There are some differences between in-situ materials and the materials used in the laboratory, especially as regards the mortars. The time doesn’t affect very much the bricks mechanical properties, but it affects very much the mortar resistance. For example, aerial lime mortar reaches the maximum strength even 50 years after the maturation. The main aim is to establish if the experimental curves derivable from the laboratory data fits well the data obtained from in-situ experiments. The comparisons will be made in chapter 5 of this thesis, and it will be also verified if the analytical model, described in section 5.1.4, allows us to estimate the compressive strength for a mortar of a real building and of a mortar produced in laboratory.

4.1 Mechanical characterization of the materials of “Puig i Cadafalch house”, Argentona, Spain It was available an opportunity to perform our tests and techniques on the materials (mortar and bricks) of Puig i Cadafalch house. On this building it was planned a restoration work, and a detailed analysis on the materials was commissioned to the UPC laboratory, with the aim to estimate the mechanical properties.

4.1.1 Description of the case study The House Puig i Cadafalch was the home for the summer holidays of the famous architect Josep Puig i Cadafalch (Ministry of Culture, Historic Heritage, El Punt Avui). It was built between 1897 and 1905. It is located in Dolores Monserdà Street, 3-5, near Vender square, in Argentona (Barcelona). The house was built from the transformation of three buildings. The style has a medieval influence, and it presents a facade with modernist architectural elements (figure 4.1.1.a). Its state of conservation is very poor (figure 4.1.1.b). During December 2010 the walls surrounding the farm felled, because of a strong wind action. In January 2012 the house restoration program was announced, thanks to a loan of 246000 euros from the Government of Cataluña.

137

4 â&#x20AC;&#x201C; In-situ experimental program

Figure 4.1.1 â&#x20AC;&#x201C; Puig i Cadafalch house; a) Before the weather event; b) After the meteorological event (strong wind).

They are doing studies aimed to improve the resistance of the battlements surrounding the roof and the terrace. The battlements have fallen due to wind action, therefore to preserve the original architectural appearance they must be restored to their original position, with works that improve the mechanical performance against the wind. For the restoration, a big portion of the original material will be used. So it is necessary to study the mechanical properties of mortar and bricks. This was a good opportunity to try all the tests for a real case, and to compare the response of mortars made in laboratory and old mortars found in a building. These mortars are more than 100 years old, while the mortars produced in the laboratory matured for a few months.

4.1.2 Analysis of the available material and preparation of the specimens In a room many fallen battlements are stored (Figure 4.1.2.a). Bricks and mortars in the battlements and in the rest of the house are the same. Using hammer and chisel, from some battlements pieces to bring in laboratory were removed. All the pieces of mortar and bricks were accurately numbered, according to the battlement of origin (Figure 4.1.2.b).

138

4 – In-situ experimental program

Figure 4.1.2 – Battlements fallen from the terrace and from the roof.

In the laboratory, the bricks were cleaned: the mortar was accurately removed by hammer and chisel (Fig. 4.1.3.a), the bricks faces were cleaned using a steel brush (4.1.3.b), and observing the bricks before and after the cleaning, the result seemed good (Fig. 4.1.3.c and 4.1.3.d).

Figure 4.1. 3 – Cleaning of the bricks; a) Removing of the mortar using hammer and chisel; b) Cleaning of bricks faces using a steel brush; c) Pieces of brick before cleaning; d) Pieces after cleaning.

The available mortar had an additional classification according to the color (Figure 4.1.4). In fact, the darker mortar should be relative to a renovation work after the first construction. A lower resistance of the clearer mortar was expected.

139

4 – In-situ experimental program

Figure 4.1.4 – Mortar from Puig i Cadafalch house: classification; a) Clear mortar; b) Dark mortar (it probably contains a little percentage of cement).

Then the bricks and the mortar pieces were marked obtaining: -

At least three cubes 35x35x35 mm from the bricks (the bricks were very thin);

At least six prismatic specimens to make flexural tests. We wanted to make three tests with traction in direction y and three in direction x, where y and x are respectively the length and the width of the brick.

At least three mortar specimens to be tested at compression (Double punch test). This test is largely described at section 3.1.6 of this thesis. The adopted technique was the same as described in the section 3.2.5. Gypsum powder was putted on the specimen’s center before the load.

They were also made some prismatic specimens for compressive tests (at least three in direction x and in direction y). The aim is to compare the prisms compressive strength with that of cubes. The prisms were twice as high as the cubes, and this may affect very much the compressive strength, because the confinement effect of the load plates decreased.

By cutting with circular saw (figure 4.1.5.b), the mortar and brick specimens were obtained (figures 4.1.5.c and 4.1.5.d) and they were stored in the climatic chamber (T≥15°C, UR≤ 65%) until reaching constant mass, according to EN 772-1.

140

4 – In-situ experimental program

Figure 4.1. 5 – Specimen preparation; a) Marking of the specimens; b) Cutting of the specimens c) Storage of brick specimens; d) Storage of mortar specimens.

4.1.3 Double Punch Tests on mortar joints To perform these tests, they were followed the rules indicated in the section “5.2 Method III” of DIN 18555-9. Instead of preparing gypsum in the most common way, by mixing with water, it was tried another method, applying some gypsum powder between the specimen’s surfaces and the punches (Figures 4.1.6.a and 4.1.6.c). Thanks to the gypsum powder, the load is more uniformly distributed, and this is clear observing the failure mode. There is a central zone who breaks forming an hourglass (Figure 4.1.6.b), and cracks that propagate along the surrounding part (Figure 4.1.6.d). Observing the results in Table 4.1.1, it can be seen that the dark specimens have a compressive strength averagely higher than the clear specimens, as expected. The scatter is very high, with CV around 40% for each group of samples. The scattering is lowest for mortars prepared in laboratory (section 3.2.5 of this thesis). In a historic building there are many more uncertainties: we don’t know the doses (lime, sand and water) used in the mortar 141

4 â&#x20AC;&#x201C; In-situ experimental program making process, and the differences between different kinds of mortar are only detectable by visual analysis. A microscopic analysis to divide the mortars in different groups could be useful, but the resolution of the minor destructive instruments is not suitable for very precise analysis.

Figure 4.1.6 â&#x20AC;&#x201C; Double punch test on mortar specimens (Puig i Cadafalch house); a) Specimen between the punches, before the test; b) Central zone who breaks forming an hourglass; c) Putting of gypsum powder on the upper specimen face; d) Failure mode.

142

4 – In-situ experimental program Table 4.1.1 – DPT results (compressive strengths)

4.1.5 Laboratory tests on bricks To find the compressive strength of these bricks, they were made compression tests on cubic specimens. This kind of brick is too thin. It was not possible to make a compressive test of a whole brick.

Figure 4.1.7 – Compressive tests on cubic specimens; a) Direction Z; b) Cube between the load platens; c) Direction Y; d) Direction X.

On the PyC specimens they were made compression tests along the three main directions z, y and x (Figure 4.1.7.a, 4.1.7.c and 4.1.7.d respectively). All the tests were made putting 143

4 – In-situ experimental program gypsum powder on the loading faces. All the results are in the tables 3.3.2 and 3.3.3 (section 3.3.3 of this thesis), but they are in table 4.1.2 to compare the resistances along the three directions. Table 4.1.2 – Compressive strengths of the cubes (35x35x35mm)

The mean compressive strength is higher for the specimens tested along the direction Z (the direction along which the brick normally work). The scatter is lower in direction Z too, may be because the faces perpendicular to this direction are more parallel and better levellable than the other faces. On this kind of bricks, some tests on prismatic specimens were made, and the results are shown in table 4.1.3 and 4.1.4. Table 4.1.3 – Bricks of Puig’s house, compressive tests on prismatic specimens, direction Y

144

4 – In-situ experimental program Table 4.1.4 – Bricks of Puig’s house, compressive tests on prismatic specimens, direction X

The compressive strengths is too low and the scatter too high, comparing with cubic specimens results. This may be due to the not perfect vertically of the samples. The results are hardly acceptable and this kind of test was not repeated on other types of brick.

4.1.4 Helifix screw pull-out tests on mortar In a first inspection, we made pull-out tests on the battlements mortars (Table 4.1.5). We paid very much attention to the mortars colors, in order to compare the pull-out force to the compressive strength of same groups of mortar. In this way we could achieve acceptable (Fext; fc) points. As regards the battlements, we were sure to perform the DPT test ant the pull-out test on the same material.

145

4 â&#x20AC;&#x201C; In-situ experimental program Table 4.1.5 â&#x20AC;&#x201C; Puigâ&#x20AC;&#x2122;s house; screw pull-out on the mortar of the battlements

In addition to tests made on the battlements, some pull-outs are also carried out on the mortar of the internal walls. Each floor of the house was analyzed with the equipments, but it was not possible to get mortar from the walls for laboratory Double Punch Testing. In this case the minor destructive tests may determine some differences between the mortar compressive strengths of each floor, and if the analytical model described in section 5.1.2 of this thesis is valid, the compressive strength can be estimated without destructive tests. On each floor there were some walls areas where the external plaster was removed. In this way visual analyses were possible and some minor destructive tests were performed on mortar and bricks of the walls (Figure 4.1.8). In each hole a reasonable number of pull-out tests was performed, in order to get satisfactory informations, without compromise elements of great architectural importance. The data are in Table 4.1.6 and they will be used to estimate the compressive strength of mortar.

146

4 – In-situ experimental program

Figure 4.1.8 – Helifix screw pull-out tests on the mortar of the wall; a) Ground floor; b) Hole in a wall of the ground floor; c) Fixing screws; d) Screws fixed. Table 4.1.6 - Helifix screw pull-out on the mortars of the walls

147

4 â&#x20AC;&#x201C; In-situ experimental program

4.1.5 Windsor Pin test on mortar In the same holes some windsor pin tests were performed (Figure 4.1.9). It was already specified that the mortars in the wall were not available for direct compressive tests. So these tests are used to find a feedback with the pull-out tests. In the mortar where the extraction force is higher, the pin penetration depth should be lower. If this occurs, the results are more reliable. The data of penetration depth were also compared with the mortar fabricated in the laboratory. The aim is to estimate the compressive strength by these readings.

Figure 4.1. 9 â&#x20AC;&#x201C; Windsor pin test on the mortars of the walls.

Table 4.1.7 â&#x20AC;&#x201C; Windsor pin test on the mortar of the walls

From a first observation of the data (table 4.1.7), it is observable that the pin penetration depth is lower on mortars with higher pull-out force. These results strengthen the tests validity. The data will be better studied in chapter 5 of this thesis.

148

4 â&#x20AC;&#x201C; In-situ experimental program

4.1.6 Minor destructive tests on bricks On the bricks of Puig i Cadafalch house were also performed minor destructive tests with the tools Helifix screw pull-out system and Windsor pin system. The behaviour of the bricks is different from the behaviour of the mortars. The results are in chapter 3, with the data of the other tested bricks.

149

5. Discussion of results In this chapter all results anticipated in chapters 3 and 4 will be discussed and put in evidence. In fact, in the previous chapter they are shown all data obtained from all tests, the trends over time of the resistances for all types of mortar tested and also the results concerning the tests made on bricks. In this chapter the aim is to compare the results and try to understand if there are some correlations between the various strengths: the mortar in the joints is the same mortar used for the prisms making process according to UNE-EN 196-1, and the minor destructive tests have been made on the joints, which were then compressed (DPT). So there are good preconditions, even if it is clear that the many experimental uncertainties and the poor quality of some materials will make difficult some interpretations.

5.1 Results of tests on lime mortar Now some comparisons between the results of mortar tests will be described, trying to correlate the results obtained from the prisms and from the joints, and mainly the results of minor destructive tests performed with the Double Punch tests of the joints. In fact the penetration tests and DPT are made on the same materials, but the mortar forming the prisms matures in different conditions from the mortar forming the joints, since is not in contact with the bricks, and it is subjected to a different compaction during the preparation phases.

5.1.1 Compressive tests VS Double Punch Tests In this section, for all types of mortar it is made a comparison between the compressive strength obtained from the two different types of compressive test: the standard test of the two halves obtained after flexural test according to EN 1015-11:2007, and the DPT of the mortar joints extracted from the wallets, according to DIN 18555-9. Both tests are described respectively in section 3.1.4 and section 3.1.6. The comparisons were made considering the prisms manufactured in steel molds, and the joints treated with gypsum before the tests, because these tests are more reliable, having lower experimental dispersion. Steel molds are standardly manufactured, and each specimen loses moisture in the same way, while in the wood molds each piece of wood could have different characteristics.

150

5 – Discussion of results As regards the results from DPT, they were considered only specimens regularised with gypsum, because the load is uniformly distributed and the failure mode was the same for all specimens, so the average compressive strengths obtained are very realistic. Remember that in the first experimental campaign the square specimens were treated with gypsum paste, and in the second campaign (only for kerakoll mortar M5) the specimens were treated with gypsum powder. In table 5.1.1 they are shown comparisons between compressive strengths from standard tests on prisms and from DPT, for all types of mortar in the first experimental campaign. Table 5.1.1 – Comparison between compressive strentghs averages fcST,steel VS fcDPT,gy; a) Hydraulic mortar NHL 3,5_1:3; b) Kerakoll mortar M5 “Biocalce muro”; c) Aerial mortar CL90_1:3 “Ciaries”; d) Aerial mortar CL90_1:3 “Segarra”

151

5 – Discussion of results  fcST,steel is the standard compressive strength of the mortar prisms molded in steel;  fcDPT,gy is the compressive strength of the mortar joints tested applyng gypsum on the loading zones of the square samples. For hydraulic mortar NHL3.5_1:3 (Table 4.1.1.a), in the considered period the compressive strengths didn’t have a big variation, and it was very difficult to find a satisfactory correlation. Observing the average ratio between the compressive strength of the joints and the compressive strength of the prisms, it can be assumed that for this type of mortar the joints have a compressive strength 65-70% higher than the prisms, and it is interesting to observe if this trend is the same for mortar with similar mechanical properties. For hydraulic mortar Kerakoll “Biocalce muro” (Table 4.1.1.b), the compressive strength of the joints is much higher than the compressive strength of the prisms, but the ratio between the resistances in each age is very different. Joint and prisms mature differently, have different shapes and their properties are hardly comparable. For the prisms, the maximum compressive strength exceeds early the value of 5 MPa (guaranteed by the product data sheet) and it remains more or less constant after 28 days. Instead in the joint resistance evolution there are strange changes between two consecutive ages, with values that before increase and then decrease. For aerial mortar CL90_1:3_”Ciaries” (Table 4.1.1.c), the compressive strength of the joints is more or less equal to the compressive strength of the prisms, but for aerial mortar CL90_1:3_”Segarra” (Table 4.1.1.d) the compressive strengths of the prisms are higher than the compressive strengths of the joints, and this doesn’t happen for the other studied mortar. Perhaps the difficult extracting procedures of the aerial lime mortar joints damaged the specimens and the compressive strengths of the square samples were underestimated. We can conclude that the compressive strengths of the joints are not comparable with the compressive strengths of the prisms. Changing the type of mortar, changes the ratio (fcDPT/fcST) and it is impossible to find a law that correlates the two types of strength, and this can be seen in Figure 5.1.1.

152

5 – Discussion of results

Figure 5.1.1 – Comparison between compressive strengths averages: fcST,steel VS fcDPT,gy; a) Hydraulic mortar NHL 3,5_1:3; b) Kerakoll mortar M5 “Biocalce muro”; c) Aerial mortar CL90_1:3 “Ciaries”; d) Aerial mortar CL90_1:3 “Segarra”

To search for a better correlation between fcDPTgy and fcStsteel, they were putted in the same graph (Figure 5.1.2) all the points regarding the experiments on all types of mortar. They were also added the points (fcStsteel; fcDPTgy) of the second experimental campaign, where it was used kerakoll mortar. Remember that: -

fcStsteel mean standard compressive strengths of prisms made in steel molds;

fcDPTgy mean compressive strengths of square samples obtained from the joints and treated with gypsum after th DPT.

153

5 – Discussion of results

Figure 5.1.2 – Comparison between compressive strengths averages: All mortars.

The correlation seems better. It should be pointed that in the second experimental campaign it was used gypsum powder to level the samples surfaces before the tests. The gypsum powder increases the compressive strengths averages a little less than the gypsum paste, but the values are comparable and usable for the same correlations. Table 5.1.2 – fcDPTgy VS fcSt kerakoll mortar; a) Campaign 1; b) Campaign 2

Observing the results (Table 5.1.2), there is a similar ratio fcDPTgy/ fcStsteel for averagely strong mortar. The mean ratio is the same in the two experimental campaigns, and it is comparable with the case of mortar NHL3,5_1:3, where it was found average (fcDPTgy/ fcStsteel) = 1,65.

154

5 – Discussion of results So it can be assumed that for weak mortars there is an average ratio fcDPTgy/ fcStsteel between 1,6÷2,5. When the mortar has a good workability, the two types of compressive strength are correlational.

5.1.2 Compressive tests VS Flexural Tests To derive a law that correlates the extraction force of the helices and the compressive strength of the mortars, it is also useful to establish a relationship between the tensile and the compressive strength of the mortars studied. In fact according to the “theory of MohrCoulomb” (W.F.Chen) the angle ϕ of internal friction depends to the ratio fc/ft (ft=tensile strength of the mortar) and ft is closely related to fflex (flexural strength). The angle ϕ is a fundamental parameter in the “Theory of the thick pipe” (Belluzzi 1980). This theory has been followed to analytically interpret the result concerning the helifix pull-out test, and to correlate the pull-out force with the mortar compressive strength. The relationship between fc and fflex is observable in Figure 5.1.3 and in Table 5.1.3.

Figure 5.1.3 – fc VS fflex_all types of mortar. Prisms made in Steel molds.

fflex is the flexural strength of the mortar prisms;

fcST is the standard compressive strength.

Observing the results and the trend-line (y=2,516x-0.1394) it can be noted that for the types of studied mortars, averagely fc ≈ 2,5·fflex. This data is very important for the choice of the 155

5 – Discussion of results internal friction angle ϕ. It is sufficient to establish a ratio between ft (tensile strength) and fc. Then it will be possible to establish a friction angle ϕ useful in an analytical model that correlates fc and FEXT. Table 5.1.3 – fc VS fflex a) NHL 3.5_1:3; b) Kerakoll M5 Campaign 1; c) CL90_1:3 “Ciaries”; b) CL90_1:3 “Segarra”

156

5 – Discussion of results

5.1.3 Brazilian tests VS Flexural Tests The Brazilian test was made to have another parameter to describe the tensile strength of the mortars, and to assign a friction angle ϕ which represents the studied mortar. The angle ϕ is related to the ratio between compressive and tensile strength of mortar (W.F.Chen) and it is a fundamental parameter in the analytical model that links the joint compressive strength with the pull-out force. In the section 3.1.5 of this thesis they are shown the results of Brazilian test on mortars, at each age of maturation. The tensile strengths ζ t are very low. In fact weak mortars have low friction angle, insufficient to develop the typical complete splitting mechanism characterized by two wedge regions under the strips and a vertical crack connecting them. This splitting mechanism occurred only for the prisms of the strongest kerakoll mortar (Figure 5.1.4) and the ratio fflex/ ζ t is lower than calculated for the other types of mortar (Table 5.1.4).

Figure 5.1.4 – Brazilian test; a) Test setup; b) Splitting mechanism characterized by two wedge regions under the punches and a vertical connection crack connecting them

157

5 – Discussion of results Table 5.1.4 – Ratio fflex /ζ t for all types of mortar

fflex is the flexural strength of the mortar prisms;

σt is the tensile strength obtained from the brazilian test.

The ratio is very high for weakest mortars, between 3,5 and 4,5, because the splitting mechanism does not develop, and the data are hardly acceptable. As expected kerakoll mortar has the lowest ratio fflex /ζ t =2,89. In this calculation it wasn’t considered the first age of maturation, because ζ t reaches immediately maximum value, then it remains more or less constant, but fflex after 10 days is still low, with ratio fflex /ζ t<<3. In spite of doubts about the validity of the Brazilian tests on weak mortars, putting on the same graph (Figure 5.1.5) all the points (fflex /ζ t) it was found a good correlation between the two types of tensile strength. It was discarded the worst point concerning the test done on kerakol mortar after 10 days of maturation.

158

5 – Discussion of results

Figure 5.1.5 – ζ t VS fflex all types of mortar. Prisms made in Steel molds.

5.1.4 Double Punch Test VS Helifix screw pull-out tests In this section the compressive strengths of the joints and the results from helifix screw pullout tests on the same joints are compared. The data cover the testing done on hydraulic lime mortars NHL3.5_1:3 “Cementos Tigre” and Kerakoll “Biocalce Muro” M5 (experimental campaigns 1 and 2). For aerial lime there were problems and satisfactory results were not found: the pull-out force of the helices gave almost always

but the compressive

strengths were close to the compressive strength of mortar joints NHL3.5_1:3. So this data may hinder a good interpretation of the phenomenon and it was preferable to not consider them. The compressive strengths of the joints and the pull-out force from the same joints in all age of maturation were compared (Table 5.1.5 and Table 5.1.6).

159

5 – Discussion of results Table 5.1.5 – Compressive strengths of the joints VS Helifix screw pull-out: fcDPTgy VS FEXT,hel Experimental Campaign 1__a) Nhl 3,5_1:3 “Cementos Tigre”; b) Kerakoll “Biocalc Muro” M5.

Table 5.1.6 – Compressive strengths of the joints VS Helifix screw pull-out: fcDPTgy VS FEXT,hel Experimental Campaign 2__Joint of mortar Kerakoll “Biocalc Muro” M5.

To get better observable data, they were made averages of FEXT and fcDPT for the whole wallets (Table 5.1.7). In this case the averages are made with more data and perhaps they could better represent the phenomenon and show a clear trend.

160

5 – Discussion of results Table 5.1.7 – fcDPTgy VS FEXT,hel; averages for whole wallets

In Figure 5.1.6 the graphs fcDPT,gy VS FEXThel are shown, where:  fcDPT,gy is the average compressive strength of the square samples obtained from the joints and treated with gypsum before testing (Double Punch Test);  FEXThel is the average pull-out force of the helices from the joints.

Figure 5.1.6 – fcDPT,gy VS FEXThel; a) Averages made for each joint; b) Averages made for whole wallets.

In the graph of Figure 5.1.6.a, each point represents a compression strength average obtained by testing two specimens of a joint, treated with gypsum, and an extraction force average obtained from at least 3 pull-outs on the same joint. In the graph in Figure 5.1.6.a, each point represents a global compressive strength, obtained averaging all compressive strengths of square samples obtained from a wallet and treaded with gypsum (10÷12 tests at all) and an extraction force obtained averaging all pull out tests on the same joints of the same wallets (24÷30 minor destructive tests at all).

161

5 – Discussion of results Observing the graphs, it can be noted that for high compressive strengths sometimes the helifix screw pull-out understimes fc the joints. This is not negative, because an underestimation is precautionary in the analysis of the mortar mechanical properties in a building. For very weak mortars we observe that helifix screw pull-out overestimates the compressive strengths. Anyway these errors can be normal in a big experimental campaign, and the worst data may be discarded. In statistics, the coefficient of determination R2 (5.1) is a number that indicates how well data fit a statistical model (Gujarati, Damodar N. 2009, Kmenta, Jan 1986).

Where: ∑

(5.1.1)

is the total sum of squares (proportional to the data variance); -

(5.1.2)

∑

̅ is the mean of the observed data ∑

(5.1.3)

is the regression sum of squares, also called “explained sum of squares”. ... -

are the predicted values. ∑

(5.1.4)

is the sum of squares of residuals, also called “residual sum of squares”. (5.1.5)

In some cases the total sum of squares equals the sum of the two other sums of squares defined above. R2 varies between 0 and 1. When the model does not explain at all the data, R=0. When the model perfectly explains the data, R=1. So despite the estimation errors described above, the trend line shows a good correlation between the values fcDPT,gy and FEXThel. In fact the coefficient of determination R2 is:  R2=0,644 for the averages made for the experiments on single joint (Figure 5.1.6.a);  R2=0,748 for the averages made for the experiments on all the joints of a wallet (Figure 5.1.6.b);

162

5 – Discussion of results Now it is interesting to see if there is an analytical interpretation that explains the phenomenon of helifix screw pull-out. The aim is to obtain an interpretive model and a simple formula that links the extraction force of the helices to the compressive strength of the joints. We think that it is possible because destructive and minor destructive tests were carried out on the same material (the joints). In fact in a first stage the pull-out tests were performed (section 3.1.7). Then from the same joints, the square samples 5x5cm were cutted and compressed (Double Punch Test, section 3.1.6). THEORY OF THICK PIPE (Belluzzi 1980) Consider a circular crown (disk with hole), of unitary thickness, or a pipe of arbitrary length. The edges are subject to radial and uniform forces

(inner edge) and

(external edge)

(Figure 5.1.7.a). The forces are positive or negative if they are respectively tractional or compressional (usually the forces are compressional). The tensions on

and

dont’t depend

(see Figure 5.1.7.b).

Figure 5.1.7 – Theory of thick pipe; a) Circular crown with internal and external forces

and

. b) Infinitesimal element subject to the forces’ system.

The tensions are (G.Lamè, B.P.E.Clapeyron 1833):

{ In the frequent case of only internal pressure (that better describes the helifix pull-out), and the (5.1.6) becomes:

163

5 – Discussion of results

( (

{ For each r value, compression,

has the same sign as is in traction.

, and

) )

has the opposite sign of

in the internal edge (

. So if

is in

worth (5.1.8):

In the case of Helifix pull-out, the screw extraction causes tensions η in the interface helixmortar, which measures (5.1.9):

Where T is the pull-out force, D in the diameter of the circumference circumscribed to the propeller (D=6mm) and L is the part of helix driven into the mortar joint (L=30mm). In the extraction step, the part of mortar atteched to the helix slides and has a dilatnacy effect on the surrounding mortar, due to a relative movement of the mortar particulas. So the radial tensions

develop (Figure 5.1.8.a and 5.1.8.b). According to the “low of associated flow”

(W.F.Chen), the radial tensions (Figure 5.1.8) are:

In the (5.1.10) ϕ is the friction angle that according to the law of associated flow measures (5.1.11):

In the (5.1.11)

is the dilatancy angle. So the (5.1.8) becomes (5.1.12):

164

5 â&#x20AC;&#x201C; Discussion of results In the 5.1.12,

are compression pressures which act to the mortar surrounding the

screw. The rupture occurs when the tensile strength , and

of the mortar is reached. We can put

depends to and to the ratio Ď (5.1.13):

In the analytical model

is the radius at the interface brick-mortar, and

is the radius that

goes from the center of the helix transverse section to the interface helix-mortar. The joints thickness was fairly constant (more or less 15mm), and it can be assumed a representative value

. So in our model

is:

So (5.1.12) becomes (5.1.15):

Figure 5.1.8 â&#x20AC;&#x201C; Analytical model for Helifix pull-out; a) Development of radial stresses

b) Theory of thick pipe in the case of Helifix pull-out; c) Mohr Coulomb diagram and associated flow.

165

5 – Discussion of results So using the 5.1.15, we can estimate the mortar tensile strength, having the extraction force and assigning to the material an angle ϕ of internal friction. To assign an appropriate ϕ, it is necessary using the experimental results, comparing for example the flexural strengths (related to the tensile strengths) and the compressive strengths of the mortar prisms (section 5.1.2 of this thesis). In fact according to Mohr-Coulomb theory (W.F.Chen), the mortar tensile and compressive strength are related, and both depend on the parameters ϕ and c (cohesion) (5.1.16, 5.1.17): )

(

Then

and

worth respectively (5.1.18 and 5.1.19):

To estimate the compressive strength

of the mortars through the pull-out force average, it is

sufficient to observe that there is a ratio between

and

(5.1.20) according to Mohr-

Coulomb theory:

Finally knowing the mortar tensile strength

(5.1.15) related to the pull-out force,

calculated (5.1.21). It is necessary to assign the internal friction angle represents the tested mortars.

166

can be

, which adequately

5 – Discussion of results In a first step it was made a comparison between the trend lines, obtainable with the experimental data, and three types of analytical curves. The aim of these curves is to estimate the joint compressive strength from the pull-out resistance, without performing destructive tests on mortar. The linear curves are described on the chart fcDPT VS FEXT by the equation:

(

) (

)

Where: FEXT is the pull-out force [N]; D is the diameter of the screw (6 mm); ρ = Re/ Ri; Re is the outer radius of the swept circle; Ri is the inner radius of the swept circle; ϕ is the a internal friction angle of the mortar; L is the length of the screw fixed part. To view the line on the chart is sufficient to assign two extreme values of the pull-out force, and to derive the compressive strength by the formula 5.1.22 (with attention to units of measurement). The terms of the equation 5.1.22 are described in this section (THEORY OF THICK PIPE). In figure 5.1.9.a is shown a comparison between the analytical curves obtained with ϕ = 30°, 35° and 40° and the trendline which linearly interpolates the experimental data. The values concern the averages made for each wallets joint. There are three groups of data, for three types of mortar that provided satisfactory results (kerakoll mortar age 1 and 2, NHL3,5_1:3). In figure 5.1.9.b the data are similar to those of figure 5.19.a, but the averages are made for compressive strengths and pull-out strengths of whole wallet. Observing these two graphics, it can be assumed that the analytical curve made assuming

, fits better the experimental

data because its trend is similar to that observed for the empirical trendline. In our model it would be fair to assume

to better represent the mortars behaviour. So they were

made other two graphs (figure 5.1.9.c and 5.1.9.d). In these two graphs the trendline changes, because the worst data are discarded. The worst data are those concerning Age 2 of kerakoll mortar (campaign 1), which overestimate the joints compressive strength, and the data of Age 4 of NHL3,5_1:3, where the compressive strengths effectively found are too low compared to those obtained from the model. In graphs c and d the trend line is almost the same as the analytical line. In this model it can be assumed

like angle of shear strength. In this way the behaviour of these kind of

mortars is better described (weak mortars and medium resistance mortars).

167

5 – Discussion of results Definitely the model adequately describes the mortar behaviour, and it can be used to estimate the compressive strength if the pull out force obtained from helifix screw pull-out is known. Evidence that certifies the goodness of the adopted model is show in Table 5.1.8. The maximum relative errors that it would commit adopting the model (equation 5.1.22) instead of the trendline are low. The relative errors are calculated with the 5.1.22:

Where:

is the compressive strength calculated by the trendline equation; is the compressive strength calculated with the 5.1.22.

The trendlines used are shown in figure 5.1.9.c and 5.1.9.d: -

Y = 0,0069x-0,0366 (averages for each joint);

Y = 0,0076x-0,5015 (averages for whole wallets);

The model provides

when

, and this can be assumed right. The trendline

was obtained with experimental data, that are affected by several errors, and shows when

. The goodness of the model can be also seen in the graphs of figure 5.1.10.a

and b. Almost all the points in the plane fcDPT VS FEXT are enclosed in the region between the lines [

]

]. The not enclosed

[

points describe compressive strengths higher than 30% than the values estimated by the model (estimation on the safe side). Table 5.1.8 – fcDPT VS FEXT – Relative error εr , Trendline VS Analytical line ϕ35° a) Averages made for each joint; b) Averages made for whole wallet

168

5 – Discussion of results  fc (Y) is the average compressive strength of the mortar joints estimated using the trend line;  fc (ϕ=35°) is the compressive strength of the mortar joints obtained using the analytical model, assuming the angle of internal friction ϕ=35°.

Figure 5.1.9 – fcDPT VS FEXT,hel – Comparison between trendline and analytical curves obtained by theory of thick pipe. a) Analytical curves (assuming ϕ = 30°, 35° and 40°) VS Trendline (averages made for each joint); b) Analytical curves (ϕ = 30°, 35° and 40°) VS Trendline (averages made for whole wallets); c) Analytical curve with ϕ = 35° VS Trendline (averages of each joint, less worse data); d) Analytical curve with ϕ = 35° VS Trendline (averages of whole wallets, less worse data);

169

5 – Discussion of results

Figure 5.1.10 – fcDPT VS FEXT,hel – Analytical curve of thick pipe model: experimental points enclosed in the region [An(ϕ35°)*0,70÷ An(ϕ35°)*1,30] – a) Averages for single joint; b) averages for waal wallets

5.1.5 Compressive standard test VS Helifix screw pull-out tests In this section the aim is too verify if the model described at point 5.1.4 of this thesis well estimates the prisms compressive strength (EN 1015-11:2007). Already we could think that the correlation will be worse than that obtained from double punching test VS helifix screw pull-out test, because in this case destructive and minor destructive tests were performed on different materials (same chemical composition, but different hardening conditions: different water content, degree of compaction, porosity, density). In the table 5.1.9 the results are shown. For compressive strength the average is made for six values (six halves were tested) and for the pull-out force, the average is made for 25÷30 values (a lot of extractions from each wallet were performed). Table 5.1.9 – fcDPTgy VS FEXT; averages for whole wallets

170

5 – Discussion of results In figure 5.1.11.a the graph shows the trendline Y that interpolates the data (FEXT; fc), and the analytical linear curves obtained for

], following the “The theory of thick

[

pipe” (section 5.1.4). As expected, the curves do not accurately fit the experimental results, and the correlation was better for compressive tests on joints. In this case the best analytical curve is that obtained for

(instead for joints

). So comparing with the case of

DPT, if the same analytical curve is used, for standard compressive strength the model overestimates the prisms resistance. In figure 5.1.11.b the graph shows the trendline Y, obtained ignoring the data of age number 4 of NHL3,5_1:3 mortar, were the pull out force excessively underestimates the compressive strength. The trendline is close to the analytical line assuming

Figure 5.1.11 – fc VS FEXT,hel – Comparison between trendline (averages made for whole wallets) and analytical curves obtained by theory of thick pipe; a) Analytical curves (assuming ϕ = 30°, 35° and 40°) VS Trendline Y; b) Analytical curve (ϕ = 30°) VS Trendline Y (averages made ignoring the worse data);

The table 5.1.10 shows a comparison between the relative errors made using analytical model for DPT (5.1.10.a) and for standard compressive test (5.1.10.b). As regards the standard compressions, the model highly overestimates the compressive strength for weak mortars, with a maximum relative error

. Often is normal to find big scattering on weak

mortars and the analytical curve fits quite accurately the data. But for standard compressive strength the best curve is for ϕ=30°, while for DPT the best curve is for ϕ=35°, and here there are always low errors in the whole analyzed range of compressive strength. So the model can be considered more reliable for the joint test interpretation.

171

5 – Discussion of results Table 5.1.10 – fc VS FEXT – Relative error εr , Trendline VS Analytical line a) Double Punch Test VS Helifix screw pull-out, analytical line with ϕ=35°; b) Standard compression test VS Helifix screw pull-out, analytical line with ϕ=30°;

5.1.6 Double Punching Tests VS Windsor Pin tests The experimental results of Windsor pin tests are in table 5.1.11 (averages for each joint) and 5.1.12 (averages for whole wallets). Table 5.1.11 – fcDPT VS P.depth – Averages for each mortar joint

Table 5.1.12 – fcDPT VS P.depth – Averages for whole wallets

172

5 – Discussion of results Is interesting to compare the penetrometer test results with the compressive strengths of the same mortar joints on which the minor destructive tests were performed. In figure 5.1.12 the charts fc VS P.depth are shown. In chart 5.1.12.a it can be observed the trendline which approximates the average data for each mortar joint. The coefficient of determination (R2 = 0,4617) is quite low, there isn’t a good linear correlation between the two types of value. The averages include 5 penetration test and 2 DPT for each joint, and this may not be enough to have a good correlation. If we consider the averages made for all joints of a wallet, the correlation (fc-P.depth) is better, with an higher coefficient of determination (R2 = 0,789).

Figure 5.1.12 – fcDPT VS P.depth– a) Averages made for each joint; b) Averages made for whole wallets; c) Correlation curves (fc VS P.depth) provided by the product data sheets; d) Comparison “tool correlation curves VS trendline Y” obtained from experimental results.

173

5 – Discussion of results It can be assumed that the penetration tests made in the second experimental campaign has been well performed and in the studied resistance range the trendline accurately estimates the joint compressive strengths from the pin penetration values. So for the compressive strength range (2÷10 MPa) of lime mortar joint, the curve (5.1.23) can estimate the compressive strength with low errors (

The instrument tables provide a very different correlation (fc-P.depth), that doesn’t fit the compressive strengths of the joints analyzed in this experimental study. The instrumental curve is piecewise defined. There are several ranges of resistance that not belong to the same line, but belong to different slightly shifted curves. Probably the cement mortar used for the tool calibration had different behaviours for different resistance ranges. The relation between the pin penetration depth and the compressive strength is linear for the experimental values and for the values in the tool tables (figure 5.1.12.c). The tabular values are hardly acceptable for lime mortars. The experimental values have a very different trend (figure 5.1.12.d). For lime mortars in the compressive strength range (2÷10MPa), the curve 5.1.23 can be considered reliable, because obtained from experimental averages made on a large number of tests.

5.1.7 Compressive tests VS Windsor Pin tests The experimental results of Windsor pin test and of standard compressive test are in table 5.1.13. The averages have been made on the following number of test: -

25÷30 penetration tests on the mortar of the wallet;

6 compressions on the six prisms halves. The prisms were made for each wallet, with the same mortar in the joints.

The large number of minor-destructive tests is motivated by the big scattering.

174

5 – Discussion of results Table 5.1.13 – fcDPT VS P.depth – Averages for whole wallets

It is interesting to compare the penetrometer test results with the compressive strengths of the prisms. In figure 5.1.13 the graphics fc VS P.depth are shown. In the graph 5.1.13.a it can be observed the trendline which approximates the average data. The coefficient of determination (R2 = 0,683) is quite high, so the correlation between the joint pin penetrations and the prism compressive strengths is acceptable. In figure 5.1.13.b it can be observed how the curve deduced from the instrumental tables does not fit the experimental data. The trend is completely different.

Figure 5.1.13 – fcST VS P.depth – a) Trendline Y; b) Trendline Y compared with the lines made with the tabular data of the instrument.

For the estimation of the prisms compressive strength (EN 1015-11:2007) the following empirical curve is proposed (5.1.24):

175

5 – Discussion of results This curve is different to than obtained for joints compressive strength. The comparison (fcSTANDARD VS P.depth) is useful to verify if the followed procedure in the minor-destructive tests performing was good (coefficient of determination quite high). This is another aspect which suggests the utility of the experimental correlation found between the joints compressive strengths and the penetration depth into the same joints (equation 5.1.23). This correlation is very useful, because destructive and minor-destructive tests were performed on the same materials, and in a large number of cases it is fundamental to have an instrument that allows us to estimate the mortar compressive strength from minordestructive test data. In fact in the analysis of historical structures it is impossible to take samples for laboratory testing.

5.1.8 Flexural tests VS Helifix screw pull-out tests The experimental results of Pull-out tests and of standard flexural tests are in table 5.1.14. The averages have been made on the following number of test: -

25÷30 screw pull-outs on the wallets mortars;

3 flexural tests on prisms. Only the tests made on prisms molded in steel are considered. Table 5.1.14 – fflex VS FEXT – FEXT averages for whole wallets

In figure 5.1.14 the chart fflex VS FEXT is shown. The graph is interesting because the 13 points regarding averages made for the 13 considered wallets, are well approximated by the trendline, with a high coefficient of determination R2=0,805. Almost all the points in the plane fflex VS FEXT are enclosed in the region between [

]

[

] lines. This

supports the validity of the analytical model assumed (section 5.1.4 of this thesis). The mortar tensile strength is strongly correlated to the flexural strength. The model is based on the mortar tensile strength (equation 5.1.21), mobilized by the screw extraction.

176

5 – Discussion of results

Figure 5.1.14 - fflex VS FEXT – FEXT averages for whole wallets

5.1.9 Flexural tests VS Windsor Pin tests The experimental results of windsor-pin tests and of standard flexural tests are in table 5.1.15. The averages have been made on the following groups of test: -

25÷30 windsor pin tests on the wallets mortars. Five wallets made with kerakoll mortar were considered (experimental campaign 2);

3 flexural tests on prisms. Only the tests done on prisms molded in steel are considered. Table 5.1.15 – fflex VS P.depth – P.depth averages for whole wallets

177

5 â&#x20AC;&#x201C; Discussion of results In figure 5.1.15 it can be noted how the flexural strength and the mortar resistance to the pin penetration are quite related. This is important, because mortar with high flexural strength resists more to the pin penetration.

Figure 5.1.15 â&#x20AC;&#x201C; fflex VS P.depth â&#x20AC;&#x201C; Averages for whole wallets

5.2 Results of in-situ experimental program The in-situ experimental program was very important. It can be observed if the results and our interpretation concerning minor-destructive testing are applicable to a real historical building. Here the materials and the hardening mortar conditions are very different from those present in laboratory. In this workplace it is possible to control and to influence many factor, like the mortar composition and the climatic conditions. In a real case we cannot control these factors. It is very frequent to find very inhomogeneous mortars, and there are very variable experimental results. Anyway is essential to estimate the compressive strength of the materials from minor destructive tests. This is what we try to do in this thesis: give a good method to achieve this result. We have already explained that in a historical building is often not possible to make direct destructive testing, beacause it is necessary to preserve the architectural, aesthetic and historical value of the constructions. The data on non-destructive in-situ testing will be merged with the results from laboratory tests.

178

5 – Discussion of results

5.2.1 Helifix screw pull-out test on mortar joints Now we want to verify if the THEORY OF THICK PIPE (section 5.1.4 of this thesis) is applicable to the results obtained in the in-situ experimental program. The question is: Is possible to derive the mortar compressive strength from the pull-out force FEXT? We can take the graph in Figure 5.1.10.b. The graph shows the curve described by the equation 5.1.22. With this equation it was given an instrument to estimate the mortar compressive strength from the screw pull-out force. The helices were extracted from the mortar joints and the compressive strength of the same joints was evaluated making some Double Punch Tests (chapter 3). In the laboratory some wallets were tested. In these wallets the joint thickness was averagely around 1.5 cm. So we assumed:  Ri = 3 mm : internal radius at the interface helix-mortar;  Re = 7,5 mm : external radius, defined by the interface mortar-brick;  ϕ = 35°: friction angle of the mortar. Following the results of flexural and compressive tests on mortar prisms, we assume ϕ = 35°, which corresponds to a ratio fc/ft = 3,7 (equation 5.1.20) (

) (

)

We can take the graph 5.1.10.b, where the analytical curve fits the points (FEXT; fC,DPT). Each point represents fC,DPT average of ≈12 DPT on the joints of a single wallet, and average of 2430 pull-out tests on the same joints. The points (FEXT; fC,DPT) regarding the in-situ tests were added. Each point represents:  7 DPT on mortar joints (clear and dark mortars, table 4.1.1);  7 helifix screw pull-out (clear and dark mortars, table 4.1.5); We could not make a lot of tests, to avoid damaging of the constructive elements. The analytical curve in figure 5.2.1 is obtained with the data in table 5.2.1, assuming Ri = 3 mm, Re = 7,5 mm, ϕ = 35°.

179

5 – Discussion of results Table 5.2.1 – Parameters for analtycal curve (equation 5.1.22, see figure 5.2.1)

Figure 5.2.1 – FEXT VS fC,DPT; analytical curve, ϕ=35°; Points regarding in-situ tests and laboratory tests

We can observe that the model underestimates the compressive strength of the two types of mortar founded in the battlements of Puig’s house. It is important to point out an important difference between the mortar joints of the wallets made in the laboratory and the joints found in the Puig’s house. The joints of the house are thinner than the joints in the wallets. In fact they have a thickness between 8 and 10 mm. So we can change the parameters Re and ρ = Re/Ri (Table 5.2.2).

180

5 – Discussion of results Table 5.2.2 – Parameters for analytical curve (equation 5.1.22), corrected to describe the mortar joints of Puig i Cadafalch’s house

With the new parameters the compressive strength of the mortar joints can still be estimated from FEXT. The table 5.2.3 is very important: Table 5.2.3 – Comparison between cp. strengths estimated and detected.

 FEXT is the mean of extraction force values (clear mortar and dark mortar);  fc,mean is the compressive strength of the joints, found from Double Punch Tests in the laboratory;  fc,pr(*) is the compressive strength predicted with ϕ=35°, Ri=3mm, Re=7,5mm, ρ=2,5;  fc,pr(**) is the compressive strength predicted with ϕ=35°, Ri=7,5mm, Re=5mm, ρ=2,5. fc,pr(**) can be assumed like the right estimation of the joints compressive strength in the case of our in-situ tests. If we calculate the relative error like:

We can observe that the analytical model underestimates the compressive strengths, 34,0% for the weak mortars and 6,8% for the strong mortars (see Table 5.2.3). 181

5 – Discussion of results For the clear mortar (the weakest one), FEXT=120 N were found two time (table 4.1.5). These results describe two tests which would normally discard in an experimental program in the laboratory, because they were not perfectly performed. We didn’t discard them because it was not possible to perform a greater number of tests. If we had a larger number of results, discarding the worst data we could find a higher average extraction force, so the analytical model could less underestimate the compressive strength of the clear mortar. Finally we can say that the analytical model “THICK PIPE” is an effective instrument to estimate the mortar compressive strength from the pull-out force of a helix previously driven into the mortar joint. It is necessary to make a lot of pull-out tests to achieve this goal.

5.2.2 Windsor pin test on mortar joints Windsor pin tests were performed on the mortars of the wallets (section 4.1.5 of this thesis). It wasn’t possible to perform DPT on the wallet joints, because we couldn’t extract them. Anyway the penetration tests were useful to determine a feedback with the pull-out tests on the same joints. In the mortar where the extraction force is higher, the pin penetration depth should be lower. If this occurs, the results are more reliable. In Table 5.2.4 and Figure 5.2.2 the results regarding the two types of minor-destructive test are compared. Observing the graph in figure 5.2.4 we can observe that the trendline fits very well the points (FEXT – P.depth). Furthermore the coefficient of determination R2, which indicates how well the data fit the trendline, is R2=0,978, really close to 1. The trendline shows a good correlation between the values FEXT and P.depth. When the extraction force increases, the pin penetration depths very coherently decrease. This is another result which enhances the validity of the pull-out test results. From these data we can estimate the joint compressive strength. Table 5.2.4 – FEXT VS P.depth; averages for the mortar of each analyzed hole.

182

5 – Discussion of results

Figure 5.2.2 – FEXT VS P.depth; averages for the mortar of each analyzed hole.

5.2.3 Compressive strength of the mortar predicted from the pull-out force From the pull-out force is now possible to estimate the mortar compressive strength of the tested walls at each floor of the house. At each floor some holes were made on the structural wall, removing the plaster (figure 5.2.3.a). Some points of the exposed mortar were chosen to carry out screw-pull-out tests (figure 5.2.3.b). Following the analytical model of the THICK PIPE, the compressive strengths were estimated. According to the mortar characteristics the parameters showed in table 5.2.2 were chosen (Re=5mm, ϕ=35°).

183

5 – Discussion of results

Figure 5.2.3 – Mortar exposed: a) Holes at the first floor; b) Test set-up in a hole.

The mortar compressive strengths calculated with the following equation are in table 5.2.5:

(

) (

)

Table 5.2.5 – Compressive strength of the mortar in the walls of Puig’s house

184

6. Conclusions 6.1 Summary The present research deals with the development and improvement of minor-destructive techniques on historical bricks and mortars. It is an innovative research, never carried out in this way. The topic dealt is very important because if our results and developed techniques are effective, itâ&#x20AC;&#x2122;s possible to determine the compressive strength of masonry elements more precisely than in the past. In this way the analysis of a historical building, but also of a generic building, can provide a sufficiently accurate estimation of mortars and brick mechanical properties without absolutely affect the architectural value and structural integrity of the building. Thanks to the data provided by the two instruments (Helifix Load Test Unit and Windsor Pin system) they will need less direct compressive tests. Moreover, making the maximum use of the compressive tests on cubic, prismatic and cylindrical specimens cutted from the bricks, it is also possible to take important informations about the brick mechanical properties with little available material. This result could again benefit the preservation of historical building, because it will be sufficient to take few small samples for laboratory testing, correctly determining the mechanical properties. As regards the screw-pull-out test on mortars, in previous researches (Ferguson, 1994 and Vekey 1997) the pull-out force was compared with the compressive strength of mortar cubes. So the non-invasive and the direct compressive tests were made on the same materials but the boundary conditions were different. The existing correlations curve (figure 2.3.12 and 2.3.13) must be improved because until now the data provided by this instrument were very unlikely interpretable and the compressive strength were hardly estimated. So we developed a new method, making the non-invasive and the compressive test on the same mortar joints. To achieve this goal the Double Punching Test was very useful. The method was proposed by Henzel and Karl, 1987, and the importance of this test is due to the fact that the mortar joints samples for punching test better represent the real mortar behaviour inside the masonry. We could have a direct correlation between the non invasive test values (screw pull-out force and pin penetration depth) and the compressive strength of the same material (the mortar joint). Regarding the screw-pull-out test, the proposed analytical model seemed effective to calculate the mortar joints compressive strength but the studies on the bricks need improvements.

185

6 – Conclusions The model is based on the THEORY OF THICK PIPE (Belluzzi 1980), that considers a pipe of arbitrary length subject to radial and uniform internal (

and external (

forces.

6.2 General and specific conclusions Analyzing the obtained results and observations made during the present investigation, it is possible to draw the following conclusions:  As regard the Double Punching Test on mortar, the technique with gypsum powder is better than the technique with gypsum paste for several reasons: the powder is a material that regulates without creating confinement, and it doesn’t affect the resistance. This is not true treating specimens with gypsum paste (DIN 18555-9, Method III).  Comparing the results of Standard compressive test (EN 1015-11) and Double Punching Test (DIN 18555-9), we can conclude that the two methods provide results hardly comparable and to correlate the data from minor destructive tests with the compressive strength is preferable using the Puncing test results;  For aerial lime mortars and hydraulic lime mortars in ratio lime:sand = 1:3, the graphics on the mechanical properties during the hardening process, show a not always linear behaviour in time. This is due to difficulties occurred during the experimental phases. In fact to extract the mortar joints from the walls without damaging them was difficult. Probably the optimal sand was not choosen, and it was necessary to ripen the mortar for a longer time.  The mortar-mix “Kerakoll Biocalce Muro” showed a satisfactory behaviour along the hardening process. The experimental results are consistent for all maturing ages and almost all kinds of test had a satisfactory outcome. For example it was always possible to performe the screw-pull out test, while for the other mortars it was not always possible for several reasons.  As regards the screw pull out test, it is a good method to investigate the mortar compressive strength. The proposed analytical model allows us to determine the mortar joint compressive strength knowing the joint thickness and the mortar friction angle ϕ (theory of Mohr-Coulomb).

186

6 – Conclusions It is possible that different correlation tables depending on the thickness and the angle ϕ (different kinds of mortar) could be proposed to correlate the pull-out force with the compressive strength.  The analytical model gave good results for the mortar strength estimation of the analized historical building and this is the main results of this research.  As regards the Windsor Pin System, the penetration depth in the analized kerakoll mortar decreased over time. The slope of the linear trendline is very different to those described by the instrument correlation table. We can conclude that the strength table of the Windsor pin system WP-2000 is not suitable for determining the compressive strength of lime mortars. The spring force of the Windsor pin system is too high for lime mortars especially during the maturing period. By decreasing force of the penetrometer spring probably more significant result from micrometer will be obtained.  The method is very useful, also to complete the results from the Helifix Load Test Unit and to make them more reliable, as we would see in the in-situ campaign.

6.3 Suggestions for future work  In order to obtain more satisfactory results on lime mortar it is very important to be careful with the choice of the material and during the wallet and prism hardening process.  Sand with higher grain size and without presence of organic particles is advised for the mortar making process. Using river sand instead of broken sand for making mortars similar to ancient mortars is advisable.  In future works on minor destructive instruments, it is recommended to perform the non-invasive tests and the destructive tests on the same materials, with the same maturing and boundary conditions, because our results are good but more researches are required.  As regard the Double Punching Test on mortar, the technique with gypsum powder is the best because the punch confinement effects are almost nullified, the failure mode is always the same and the coefficient of variation is quite low (homogeneous values).  The best way to carry out the screw-pull-out test is:  Cut in half the screw to avoid the mortar damaging if cutting is made after screw insertion. 187

6 – Conclusions  The pilot hole has to be 3 mm in diameter for weak mortars and 4 mm in diameter for mortar stronger than 10 MPa.  A regolable driving tool for the screw insertion is recommended. We made a rudimentary tool (figure 3.1.35.c) to facilitate the perfect screw insertion after cutting. A regolable driving tool could be useful because the deeper the tie is inserted the more homogeneous the material becomes. So any surface effect is reduced.  Design Windsor pin system with weaker spring with lower potential energy especially for ancient mortars. This system should calibrate for mortars with maximum 10MPa compressive strength.  As regards the tensile test on the brick samples, to improve the quality and the accuracy of the data, a load cell of 10 KN is recommended. In these tests it were used a load cell of 450 KN, more suitable for very resistant materials.

188

References «Argentona expropiarà el casal de Puig i Cadafalch». El Punt Avui (en catalán). 08/01/12. Consultado el 8 de enero de 2012. Baronio, G., Binda, L., Saisi, A. (1995). Analisi di malte antiche e comportamento di malte riprodotte in laboratorio, 7° CONVEGNO NAZIONALE, L’ingegneria sicmisa in Italia, Siena 25-28 Settembre 1995. Baronio, G., Binda, L., Saisi, A. (1999) Mechanical and physical behaviour of lime mortars reproduced after the characterisation of historic mortar. RILEM Proceeding PRO 12. Historic Mortars: Characteristics and TestsBelluzzi O. Scienza delle Costruzioni (Volume Terzo). Bologna, Zanichelli, February 1980 Binda, L., Saisi, A. and Tiraboschi, C. 2001a. Application of Sonic Tests to the Diagnosis of Damage and Repaired Structures. Int. Journal Non-Destructive Testing and Evaluation (NDT&E Int.), vol. 34 Binda, L., Papayianni I., Toumbakari E., Hees R.V. (2002). Mechanical tests on mortars and assemblages. Characterization of Old Mortars with Respect to their Repair - Final Report of RILEM TC 167-COM, 57–76. CEN. (1999). BS-EN 1052-1 - Methods of test for masonry - Part 1 : Determination of compressive strength. CEN. (2010). BS EN 459-1:2010 - Building lime Part 1: Definitions, specifications and conformity criteria. CEN. (2010). BS EN 459-2-2010 - Building lime Part 2 : Test methods.

189

References Como M.(2013). Staticsof Historic Masonry Constructions. Departement of Civil Engineering, University of Roma Torvergata. Chen W. F. Plastycity in reinforced concrete. Library of Congress Cataloging in Publication Data. McGraw-Hill Book Company. Christiansen P.D.V (2011). In situ determination of the compressive strength of mortar joints using an x-drill. Danish Technological Institute, DK-8000 Aarhus, Denmark. de Vekey, R.C. (1997). In-situ test for masonry. Buildin Research Establishment (UK) Watford WD2 7JR England. de Vekey, R.C., Non-destructive tests methods for masonry structures, Proc.. 8th IBMAC, V.3, p.1673, 1988. de Vekey R.C., Sasso M. Comparison of non-destructive in-situ mechanical tests on masonry mortars: the PNT-G method and the Helifix method. International brick/block masonry conference Tongji University, Shanghai, China, 14 -16 october 1997. DIN 18555-9 - Prüfung von Mörtel mit mineralischen Bindermitteln, 1999-09 Drdácký M., Mašín D., Mekonone M.D., Slížková Z. (2008). Compression tests on nonstandard historic mortar specimens. Historical Mortars Conference, Lisbon, Portugal. El deteriorament de la històrica casa de Puig i Cadafalch a Argentona (en catalán) 26/12/2010 www.maulets.org. EN 772‐1 (2002). Methods of test for masonry units. Part 1: Determination of the flexural strength. Comité Européen de Normalisation (European standard), Brussels, Belgium. EN 772‐6 (2002). Methods of test for masonry units. Part 6: Determination of the compressive strength. Comité Européen de Normalisation (European standard), Brussels, Belgium.

190

References EN 1015‐11 (1999). Methods of test for mortar masonry. Part 11: Determination of flexural and compressive strength of hardened mortar. Comité Européen de Normalisation (European standard), Brussels, Belgium. Eurocode 6 part 1-1 (1996). Design of masonry structures, part 1-1: General rules for reinforced and unreinforced masonry structures, European standards. Fernandes F.M., Lourenço P.B., Castro F. (2009). Ancient Clay Bricks: Manufacture and Properties. Materials, Technologies and Practice in Historic Heritage Structures. Gambarotta L. (2001). Caratteristiche meccaniche della muratura. Progetto di ricerca “Sicurezza e controllo di ponti in muratura”. DISEG - Research Center on Masonry Bridges, University of Genoa, Italy. Gujarati, Damodar N.; Porter, Dawn C. (2009). Basic Econometrics (Fifth ed.). New York: McGraw-Hill/Irwin. pp. 73–78. ISBN 978-0-07-337577-9. Henzel, J., Sieghart, K. (1987). Determination of strength of mortar in the joints of masonry by compression tests on small specimens. Darmstadt Concrete, 2, 123-136. 8th International seminar on structural masonry, Istambul 05-07 nov 2008. Kmenta, Jan (1986). Elements of Econometrics (Second ed.). New York: Macmillan. pp. 240– 243. ISBN 0-02-365070-2. Lamé G. Clapeyron É. Mémoire sur la puissance motrice de la chaleur. dans Journal de l'École Royale Polytechnique, Paris, Imprimerie Royale, t. XIV, 23e cahier, 1834. Lourenço, P. B., Fernandes, F. M., & Castro, F. (2010). Handmade clay bricks: chemical, physical and mechanical properties. International Journal of Architectural Heritage, 4(1), 38– 58. M. Arandigoyen, J. I. Álvarez. Pore structure and carbonation in blended lime-cement pastes. Materiales de Construcción, Vol 56, No 282 (2006)

191

References M. Bošnjak-Klečina, S. Lozančić (2010). Testing of physicaland mechanicalproperties of bricksand mortar in historic structures. Technical Gazette. Martinez J. L. et Al. Monografia sobre el anàlisis estructural de construcciones històricas de fàbrica. Departamento de Mecànica de los Medios Continuos y Teorìa de Estructuras. Febrero, 2001. McNary W.S., Abrams D.P. (1985). Mechanics of masonry in compression. ASCE, 11 (4): 857–870. Ministerio de Cultura, Patrimonio Histórico. Lourenço, P. B., Joaquim O. B., José C. A. Characterization of masonry under uniaxial tension. Universidade do Minho, Departamento de Engenharia Civil. October 2002. Pelà L, Benedetti A., Marastoni D. (2012). Interpretation of experimental tests on small specimens of historical mortars. 8th International Conference on Structural Analysis of Historical Constructions. Wroclaw, Poland. RILAM TC 127-MS. Test for masonry materials and structures. Materials and Structures, Vol 30, July 1997,pp 323-328. Rocco, C., Guinea, G. V., Planas, J. and Elices, M. Size effect and boundary conditions in the Brazilian

test:

Experimental

verifications.

Materials

and

Structures/Matériaux

Constructions, Vol. 32, April 1999, pp 210-217. Rossi, P.P., 1982. Analysis of mechanical characteristics of brick masonry tested by means of nondestructive in situ tests. Proc. 6th Int. Brick Masonry Conference, Rome, Italy. Technical notes on Brick Construction. Brick Industry association. www.gobrick.com. UNE-EN 772-16. Methodos de ensayo de piezas para fàbrica de albanileriìa – Determinaciòn de las dimensiones.

192

References UNI EN 933:2012. Prove per determinare le caratteristiche geometriche degli aggregati. W. A. Ferguson, J. Skandamoorthy. The screw pull-out test for the in-situ measurement of the strength of masonry materials. Calgary, Canada, July 5-7, 1994. Witt, J. C. (2014). Caracterizaciรณn experimental de la resistencia al corte de en muros de obra de fรกbrica de ladrillo maciz. Dissertation. Technical University of Catalonia.

193