Water Journal March 2007

AWA

~

AUSTRALIAN WATER ASSOCIATION

_)

--- -

entley WaterGEMS continues with Haestad Methods' tradition of pioneering research and innovation, advancing the water modelling technology standard once again.

B

SHATTER PERFORMANCE BARRIERS Designed to support all-inclusive models, WaterGEMS gets a new power boost to build, run, edit, and map large water distribution models with ease, at record-breaking speeds.

ELIMINATE PLATFORM RESTRICTIONS Being anchored to a sing le platform is costly and risky.

WaterGEMS is the on ly so lution that lets everyone work within their favorite platform, sharing a single model file.

EASIER THAN EVERI The legendary ease of use of Haestad Methods solutions is magnified w ith WaterGEMS. Enjoy a fresh new stand-alone interface, an improved ArcGIS integration, and an all-new MicroStation platform with dozens of new features and groundbreaking utilities: • Criticality analysis

> Stand-alone GIS demand allocation

• Network trace queries

• Network segmentation analysis

• Stand-alone Terrain Extraction

• Pressure-dependant demands

• New isolation valve element

> Network navigator

• Variable speed pump batteries

• Centralised water use management

• New hydrant element

> Unit demand engineering li braries

• Dynamic SOL-based queries

• Au x iliary fire flow results

www.bentley.com/AWA

Contact us today for more information:

Free water seminar April 19, 2007, in Brisbane

Toll-free (Australia): 1800 245 005 Toll-free (New Zealand): 0800 444 046 e-mail: anz.marketing@bentley.com

For more information and to register, v isit: www.bentley.com/seminar / AWA

Volume 34 No 2 March 2007

Journal of the Australian Water Association

OPINION AND INDUSTRY NEWS OPINION Passing the Baton Ozwater - The Plaudits My Point of View

DDay, President, AWA CDavis, CEO, AWA John Ruetten, President and Co-founder of Resource Trends Inc

AWA NEWS Includes: Changing of the Guard, IWA Deputy Lured from London, New Industry Award Announced, National Specialist Networks, WaterAid Australia, Climate Change - Water Industry Adaptation and Actions AWA EDUCATION Includes: Water Education Network (WEN), Water Industry Capacity Development (WICD), Awards WORKFORCE ISSUES & TRAINING Includes: AWA's Professional Skills Program, Workforce Planning - The Melbourne Water Perspective CROSSCURRENT National Issues and Policy, States, Clippings, People in the News AWA MEMBERSHIP NEWS New Members

4

5 6 8 24

28 35

42

PROFESSIONAL DEVELOPMENT NATIONAL EVENT CALENDAR

43

Events

44

TECHNICAL FEATURES ([·,] indicates the paper has been refereed ) MEMBRANE TECHNOLOGY Membrane Magic - Desalting By Forward Osmosis Drawing water through the membrane, rather than forcing it li: Process Stability of an Add-on Membrane Bioreactor Results of o pilot trial at Rouse Hill, Sydney

BBo Ito, TTran, MHoang

46

AMacCormick, EBrois

49

KCodee, IWallis

54

CLane

61

ACampbell, AWilson, PAtkinson

64

DKirono, GPodger, WFranklin, RSiebert

68

ODOUR CONTROL ~ Odour Containment and Ventilation at Perth's Major WWTPs Recommended design criteria for near complete capture of odour SLUDGE DRYING Cyclonic Thermal Drying of Biosolids Trials of a newly developed system Thermal Drying: Operations and User Acceptance Successful experience at two rotary drum driers in New Zealand CLIMATE CHANGE [iJ Climate Change Impact on Rous Water Supply The need for a new source is most likely in 2018

WATER BUSINESS 73

NEW PRODUCTS AND BUSINESS INFORMATION - SPECIAL FEATURE: UV DISINFECTION ADVERTISERS' INDEX

84

OUR COVER Perth '.r Subiaco Wastewater Treatment Plant is in the centre ofa thriving inner suburb and deals with warm, high strength sewage. To eliminate odour complaints all equipment has been covered and ventilated to the 50 m stack and scrubbers. Both theory and pragmatism have led to design criteria for the necessary ventilation rates (see paper page 54). Photo cottrtesy of the Water Corporation.

Journal of the Australian Water Association

Water

MARCH 2007

l

'AWA

;:A: AWA CONTACT DETAILS • 'Promoting the sustainable \I' ..~,:.,. management o1 water ,\\ISIU.UII"

I

POSTAL ADDRESS PO Box 388, ARTARMON NSW 1570

Journal of the Australian Water Association

EMAIL info@awo.osn.au WEBSITE http:// www.owa .osn.au PRESIDENT

ISSN 0310-0367

David Barnes - president@owo.osn.au

CHIEF EXECUTIVE OFFICER Chris Davis - cdovis@owo.osn.au

CHIEF OPERATIONS OFFICER Ion Jarmon - ijormon@owo.osn.au

EVENTS Lindo Phillips - 61 2 9495 9914 lphillips@owo.osn.au

MEMBERSHIP INFORMATION AND INQUIRIES Michael Seller - 02 658 1 3483 mseller@awo.osn.au

MEMBERSHIP RENEWALS AND CHANGES Membership Team - 1300 361 426 info@owo.osn.au

MEDIA AND MARKETING Jennifer Sage - jsage@owa.osn.au

SCIENTIFIC AND TECHNICAL INFORMATION Dione W iesner PhD - 61 2 9495 9906 dwiesner@owo .osn .au

WATER EDUCATION NETWORK Corinne Cheeseman - 61 2 9495 9907 ccheesman@awo.osn .au

NATIONAL SPECIALIST NETWORK Loura Evanson -61 2 9495 9917 leva nson@awa. osn. au

AWA BRANCHES: AUSTRALIAN CAPITAL TERRITORY and NEW SOUTH WALES Errin Dryden - 61 2 9495 9908 edryden@owo.osn.au NORTHERN TERRITORY c/o Ion Jarmon - 6 1 2 9495 99 11 ijormon @owo .osn.au SOUTH AUSTRALIA Sarah Corey - 61 8 8267 1783 sobranch@owo.osn.au QUEENSLAND Kathy Bourbon - 61 7 3397 5644 owaq@owa.asn.au TASMANIA & VICTORIA BRANCH c/ o Rochel-onn Mortin - 61 3 9235 1416 tosbronch@owo.osn .au vicbronch@owo.osn.au W ESTERN AUSTRALIA Coth Miller - 0416 289 075 cmiller@awa.osn.au INTERNATIONAL WATER ASSOCIATION, AUST. (IWAA) c/o Chris Davis - cdovis@owo .osn.au

DISCLAIMER Australian Water Association assumes no responsibility for opin ion or statements of facts expressed by contributors or advertisers.

COPYRIGHT AWA Water Journal is subject to copyright and may not be reproduced in any format without written permission of AWA. To seek permission to reproduce Water Journal material email your request to: jsoge@awo.asn.au

2 MARCH 2007

Water

Volume 34 No 2 March 2007

AWA WATER JOURNAL MISSION STATEMENT 'To provide a print iournal that interests and informs on water matters, Australian and international, covering technological, environmental, economic and social aspects, and to provide a repository of useful refereed papers.'

PUBLISH DATES Water Journal is published eight times per year: February, March, May, June, August, September, November and December EDITORIAL BOARD Chairman: FR Bishop BN Anderson, TAnderson, ( Diaper, GFinlayson, AGibson, GA Holder, BLabza, MMuntisov, ( Porter, DPower, FRoddick EDITORIAL SUBMISSIONS Water Journal invites editorial submissions for: Technical Papers and topical articles, Opinion, News, New Products and Business Information.

Acceptance of editorial submissions is subject to editorial board discretion.

Email your submissions to one of the following three categories: 1. TECHNICAL PAPERS AND FEATURES Bob Swinton, Technical Editor, Water Journal: bswinton@bigpond.net .au AND journal@awa.asn.au Papers of 3000-4000 words (allowing for graphics); or topical stories of up to 2,000 words. relating to all areas of the water cycle and water business.

Submissions are tabled at monthly editorial board meetings and where appropriate are assigned to referees. Referee comments will be forwarded to the principal author for further action. See box on page 12 for more details.

2. OPINION, INDUSTRY NEWS, PROFESSIONAL DEVELOPMENT Jennifer Sage, jsage@awa.asn.au Articles of 1000 words or less 3. WATER BUSINESS Brian Rault, National Sales & Advertising Manager, Hallmark Editions brian.rault@halledit.com.au Water Business updates readers on new products and associated business news within the water sector.

ADVERTISING Brian Rault, National Sales & Advertising Manager, Hallmark Editions Tel: 6138534 5014 (direct), 61 3 8534 5000 (switch), brian.rault@halledit.com.au Advertisements ore included as an information service to readers and are reviewed before publication to ensure relevance to the water environment and objectives of AWA. PURCHASING WATER JOURNAL Single issues available @ $12.50 plus postage and handling; email dwiesner@awa.asn.au BACK ISSUES Water Journal bock issues are available to AWA members at www.owa.osn.au PUBLISHER Hallmark Editions, PO BOX 84, HAMPTON, VICTORIA 3188 Tel: 61 3 8534 5000 Fax: 61 3 9530 8911 Email: hallmark.editions@halledit.cam.au

Journal of the Australian Water Association

tecllnical features

MEMBRANE MAGIC DESALTING BY FORWARD OSMOSIS B Soito, T Tran, M Hoang Abstract Personal packs capable of removing all imp urities from water, including sale, have been perfected that use simple, truly low pressure osmosis. T he water fro m the feed sid e of a special hydroph ilic membrane passes to a concentrated sugar solution.

The resulting product water is in the form of an energy drink. Large-scale variations of the techn ique based on sugar recovery by reverse osmosis, or the use o f different solutes that can be recovered from the spent draw solution and recycled have been reported in the literature.

Introduction Forward or d irect osmosis (FO ) relies on osmosis, the natural diffusion of water through a semipermeable memb rane from a low concentration solution to a solution having a higher co ncentration of dissolved material. It has been explored as an alternative to reverse osmosis (RO), which is still an expensive desalination hyperfiltration procedure, both in capital and energy costs, and requires prerreatment of the feed water to minimise the foulin g chat shortens operational time and membrane life. RO utilises pressure applied against osmotic forces, to fo rce water but not salt into a more dilute receiving water; FO has a water fl ow in the opposite direction, from the dilute feed to the co ncentrated draw solut ion . The contrasting processes are rep resented in Figure 1. FO has been the subject of a recent comprehensive review (Cath et al., 2006).

T he osmotic pressure is related to the molar concentration of the solution , the pressure char must be app lied to a solution to prevent transfer of water into the

ROpmc~

FOproo.~ F ~ ~ n r a t e

t

• t

solution through a semipermeable membrane. These fo rces can b e significan tly greater than rhe hyd raulic forces used in RO, leading to the possibility of higher water flux rates and recoveries (McCutcheon et al., 2006). The osmotic pressure for some compounds is given in Table 1. I t depends on the mo lar co ncentration of the dissolved species, so low molecular weight compounds are preferred.

FO has been applied to contaminated waters, converti ng them on the small scale to purified fluids containing sugars and critical electrolytes, suitable in emergency situations and for military and bush walki ng purposes. T he first practical use seems to have been a proposal involving glucose as rhe draw solute, with seawater as the feed and a membrane made of cellulose acetate (Kravach and Davis, 1975). A nutritious drink was rhe result. In emergency life-boar application s glucose could be added to seawater to provide the draw solurion, rhe salt concentration being lowered to an acceptable level by d ilution following FO. Perso nal packs capable of delivering a litre of product water in an hour have now been developed, with 4 ,000 donated to victims of the tsunami disaster in Southern Asia, and 25,000 as a response to the US Gulf Coast hurricane disaster (H ydration Technologies, 2005).

Compound

MW

Concentration, %

Osmotic Pressure, kPa

Sea water

58.5 58.5 180 180 342

3.5 0.16 15 6 6

2800 140 2800 900 540

Fructose Fructose Sucrose

46

MARCH 2007 Water

- - . Freshwiter Inwsdooon

Figure 1. Compa rison of RO a nd FO processes.

Table 1. Osmotic p ressu re of aqueous so lutio ns.

Brackish water

:

Journal of the Australian Water Association

Drawing water through the membrane, rather than forcing it.

The contaminated feed can include urine, as in space travel situations. FO is also used to concentrate fruit juices (Beaudry and Lampi , 1990) and to hydrate dry rat ion foods and beverages (SoldierTech, 2004). Large-scale desalination systems rhar have a recyclable draw solution have been proposed (McCurcheon et al. , 2005).

Membranes Th e membrane employed is a cellulose ester with rhe same selectivity as RO membranes . However, RO membranes are not ideal for FO because of their low water flux (McCurcheon et al., 2006). The pore size is 0.3-0.5 n m, much smaller than bacteria (200-50 ,000 nm) or virus (5-100 nm). As a diffusion process, water transfer through the memb rane is slow, so FO membranes need robe very ch in (10 Âľm ) and supported on the surface of a porous microfiltration (MF) mem b rane (H ydration Technologies, 2003). The filte rs are claimed to be tolerant to very muddy water, up to 1000 NTU. Because there is no high p ressure, particles are not fo rced into membrane pores, th e performance in muddy water being essentially the same as ch ar for clear water.

T his is in contrast with p ressure driven MF (pore size ~200 n m), which does nor protect against viruses or toxins, and may plug up rapidly when the feed is of high turbidity.

Ir has been shown that present day RO membranes of rhe polyamide and cellu lose acetate rype, bo th having thick fabri c backing layers to provid e mechanical support, are nor suitable for FO because of

tecnn1ca1 reatures

low fl uxes, despite their having similar water permeability to cellulose criacetate FO membranes (McCuccheon et al., 2005). T his has been attrib uted to severe internal concentration polarisation in the porous support and fab ric layers of the RO membranes. The FO membrane, on the ocher hand, is made up of two layers, between which is em bedded a polyester mesh to give reinforcement, el iminating the need for a chick polymer support layer and a chick fab ric layer.

The application of FO as a pretreatment for RO in wastewater reclamation in space, reseed us ing a draw so lu tion of 1.7 M NaCl, two cell ulose diacecace/rriacetate and two thin film polyamide composite RO membranes gave fluxes of 1.90, 1.97, 0.54 an d 0.66 L/ m2 h respectively, versus 17.4 L/ m2 h for a cellulose triacecace FO memb rane (Cath et al., 2005a) . T he low flux of the RO membranes was again ascribed to internal concentration polarisation (Loeb et al., I 997). As a pretreatment module FO could efficiently recover more than 90% of the wastewater and was highly effective in rejecting sa le and organic co mpounds as well as in providi ng a high fl ux. However, trace amounts of surfactants were fou nd to diffuse through the system, so tighter membranes are to be tested. T he FO membrane, like most RO membranes, perfor med very badly in rejecting urea.

Complete urea rejection can be ach ieved by a dual FO/membrane distillation process which can successfu lly treat a complex feed chat could not be addressed by the separate processes (Cach et al., 20056) .

Draw Solutions for Large-Scale Operation

Find ing draw solutions that are readily recoverab le is a challenge. The ideal d raw solution must have high water solub ili ty and a low molecular weight in order to

generate a high osmotic pressure, and should not be consumed in the process. It muse be non-tox ic and palatable.

Compatibili ty with the membrane is essential. As the product water is not prima rily used as a nutrient provider, there must be a si mple and economical method of recoveri ng and recycling the solute. A number of early patents quoted in a recent article (McCurcheon et al., 2005) describe systems chat use volatile solutes such as sulphur dioxide or aliphatic alcohols chat are recyclab le. Another specifies a precipitable salt such as alumini um sul phate, wh ich after the transfer stage is treated with calcium hydroxide to produce Al(OH)) and CaSO4.

Fluxes were always higher at 50°C than at 30°C. T wo flat sheet RO me mbranes, one an asymmetric cell ulose acetate membrane, the ocher a polyamide composite membrane, were studied, together with an FO membrane made from a cell ulose ester supported on an MF membrane.

Decomposition was evident with 6 M NH4HCO3 even at 35°C. When glucose or fructose was tested, the flux for 0.5 M NaCl and 4 M fructose was consistently higher, by some 50%, than that for I M NaCl and 4 M gl ucose. Fructose is more soluble than glucose, but creates the same osmotic pressure at the same molar concentration. Sale rejection of> 97% was achieved.

Am monia and carbon dioxide also ful fi l the req uiremen ts, so a novel desalting system based on FO has been proposed that uti lises a concentrated ammonium bicarbonate solu tio n as the draw sol ution to extract water fro m a sal ine source (McCutcheon et al., 2005). Separation of fresh water fro m the FO product is simply achieved by moderate heating (~60°C), when the ammonium bicarbonate decomposes in to ammonia and carbon dioxide. The gases can be removed by lowpressu re disti llation with a relatively low energy requirement. Laboratory-scale experi ments employed a feed of 0.5 M NaCl (seawater concentration) and a draw solu tion of 6 M N H4H CO 3 at 50°C. A high water flux with ve ry little salt passage resulted (> 95% rejection). However, fl uxes were considered to be too low and a sale concentration of l M was desirable for economic and operational reasons.

Fu rther studies of the system made use of different mole ratios of ammonia and carbon dioxide, with higher ratios needed for more co ncentrated sol utions, varying from 1.2 far a l. I M draw so lu tion to 1.4 for a 6 M solu tion (McCurcheon et al., 2006). T he higher ratios favour formation of am monium carbamate, NH 2CO2 NH4, which is very water soluble. The membrane used was made fro m a hydrophilic cellulosic polymer. It had a th ickness of less than 50 µm and was q uite different to RO membranes in chat it lacked the porous polyme r and fabr ic support layer. The feed, 0.05 to 2 M NaCl, and draw liquors were held at 50°C. Resul ts showed chat increasing the draw solution concentration led to less utilisation of the osmotic driving force, because of an increase in concentration polarisatio n, as had been foun d in earl ier stud ies of osmosis. Concentration polarisation can be of two types: external and internal. External co ncentration polarisation, governed by solute properties, membra ne properties and hydrodynamics, is a build up of solute at the active membrane layer surface on the feed side. A red uced permeate fl ux resul cs because of an increased osmotic pressure and possible

Lacer workers fo und that they could achieve less than half chis fl ux at 50°C with 0.5 M NaCl and 4 M NH4HCO3, in tests char made use of ammonium bicarbonate, glucose and fructose as draw solutio ns (Ng et al., 2006). The flux differe nce was attributed to che different membrane and experimental condi tions.

Table 2. Draw solutions for FO treatment o f saline water.

Compound

FO Product

Recycling of Draw Solute

References

Glucose

Energy drink

No; seawater os solvent for glucose

Krovath & Davis, 1975

Fructose

Energy drink

No

Stache, 1989

Glucose

Energy drink

No

Hydration Technologies, 2003

Sulphur dioxide, aliphatic alcohols

Drinking woter

Heating or air stripping

Batchelder, 1965; Glew, 1965

Aluminium sulphate

Drinking woter

Ca(OH)i precipitates Al(OHb & CaS04

Frank, 1972

Sucrose (low osmotic pressure)

Drinki ng woter

Loose RO to remove sucrose

Yael i, 1992

Potassium nitrate, hot; S02 in step 2 Ammonium bicarbonate

Dri nking water

Cooling precipitates some KN03 in step 1

McGinnis, 2002

Drinking water

Heating to -60°C forms NH3 & CO2

McCutcheon el al., 2005, 2006

Ammonium bicarbonate, sugars

Drinking water

Regeneroble by heating

Ng el o/., 2006

Journal of the Australian Water Association

Water

MARCH 2007 47

technical features

osmotic d e-swelling of the membrane.

They were shown to have negligible effects on performance. In internal concentration polarisation, the draw solution within the porous backing substructure becomes diluted as water permeates the active layer, resulting in a significant decrease in the net driving force and water flux. Draw solutions of higher concentrati on increased the severity of the internal co ncen tration po larisation. Salt reject ion was good and feed concentratio ns were high, simulating high water recovery. The less than optimal performance was ascribed to the membrane not being designed for desalination of chis type; the design of better salt rejection systems having minimal internal co ncentration polarisation is under way.

The draw solutio ns that have been looked at so far are listed in Table 2 .

O pportun ities exist for the design of improved hydrophi lic FO membranes, as well as m ore efficient and eco nomical recyclable draw solutions .

Summary and Conclusions RO requires high appl ied pressure, leading to a high energy requirement and therefore high operational cost. The high applied pressu re also enhances the accumulation of major foulants on che RO membrane surface, leading to a deterioration in water flux, and costly membrane replacement.

FO is a technology that may potentially be used to treat wastewater and salt water sources with reduced energy usage, higher water flux and recovery, and less pressuredriven fo uling than RO.

The Authors Dr Brian Bolto, Dr Thuy Tran and Dr Manh Hoang (email : brian. bolto@ csiro.au; thuy.rran@csiro.au; manh.hoang@csiro .au) work for CSI RO Manufacturing and Materials Technology, Clayton , Victoria (postal address: Private Bag 33, Clayton MDC, Vic 3169).

References Batchelder, G. W. (1965). Process for rhe demineralisation of warer. US Patent 3, 17 l ,799.

.... .

.

Filtration Media

.......' ... .

Sand, Gravel, Coal, Garnet, Manganese Greensand .

Invest In accurately graded, durable media from your complete filter media professionals.

..' ... . .. .... ...

. .

..

. ... .. . ..

..

We are pleased to offer you: • Decades of mineral processing experience • Media produced to the AWWA B100-89 Standard • An extensive product ra nge at competitive prices • Manufacturing in accordance with Quality System AS/NZS ISO 9001:2000 • Packaging alternatives to suit any requ irement • Proficient technical assistance and support

·.<•:

\t::::' .

:···, .,

.. .... ... ....... . .. ..

• Prompt delivery Australia-wide and Overseas So, tap into our extensive experience and helpful service when next you require filter media.

---•--PTY

RIVER SANDS LTD

683 Beenleigh-Redland Bay Road, Carbrook Queensland 4130 International +6 1 7 3287 6444 Freeca ll 1800 0 77 744 Facsmile 07 3287 6445

48 MARCH 2007

Water

Journal of the Australian Water Association

{0

QUALIT Y MANAGEMENT SYSTEM

ISO 9001 NATA CERTIFIED

.

..

!Ill, ... . .. .. ... ... ..... ... ... .

.. .. .. ... .

Beaudry and Lampi, 1990). Membrane rech nology for direct-osmosis concentration offruit juices. Food Technol. 44, 121.

Cath, T. Y., Gormly, S., Beaudry, E. G., Flynn, M. T., Adams, V. D. and Childress, A. E. (2005a). Membrane contactor process for wastewater reclamation in space. Part l.

D irect osmotic concentration as pretreatment for reverse osmosis. J Membrane Sci. 257, 85-98.

Cath, T. Y., Adams, V. D . and Childress, A. E. (20056). Membrane comactor process for wastewater reclamat ion in space. Part ll.

Comb ined direct osmosis, osmotic distillation and membrane distillation for treatment of metabolic wastewater.].

Membrane Sci. 257, 111-119.

Cath, T . Y., Chi ldress, A. E. and El imelech, M. (2006). Forward osmosis: Principles, applications and recent developments./ Membrane Sci. 281 , 70-87.

Frank, B. S. (1972) . Desalination of seawater.

US Patent 3,670,897.

Glew, 0 . N. ( l 965) . Process for liquid recovery and solution concentrat ion. US Patent 3,216,930.

H ydration Technologies (2003). Osmotic water purification devices. Osmotic White Pape1; www.HydrationTech.com, H yd ration Technologies Inc., Albany, Oregon.

H ydration Technologies (2005). Hydration Technologies' life-sustaining warer fil rration bags deployed to hurricane Katrina victims.

HT! News, 14 Sep 2005 . www.H ydrarionT ech.com/derai1.php?lD=2 3, H ydrarion Technologies Inc., Albany, Oregon .

Kravarh, R. E. and Davis, J. A. (1975) .

Desalinarion of seawater by direct osmosis.

Desalination 16 , 151-15 5 .

McCurcheon, J. R., McGinnis, R. L., Elimelech, M . A. (2005) . A novel ammoniacarbon dioxide osmosis desalination process .

Desalination 174, 1-11.

McCutcheon, J. R., McGinnis, R. L., Elimelech, M.A. (2006) . Desalination by ammonia-carbon dioxide forward osmosis: Influence of draw and feed solution concentrations on process performance. J. Membrane Sci. 278, I I 4-123 .

McGinnis, R. L. (2002). Osmot ic desalination process. US Patent 6,391,205 BI. Ng, H. Y. , Tang, W. and Wong, W. S. (2006) .

Performance of forward (di rect) osmosis p rocess: membrane rransport and transport phenomenon. Environ. Sci. Technol. 40, 2408-2413 .

SoldierTech (2004). Warer, warer everywhere: self-hydrating pouch . www.milirary.com/ soldiertech/0, l 4632,SoldierTech_Hydrat ing Pouch, Military.com.

Srache, K. (1989). Apparatus for transforming seawater, brackish water, polluted water or t he like into a nurritious drink by means of osmosis. US Patent 4,879,030 .

Yaeli, J. (1992). Method and apparatus for processing liquid solutions of suspensions particularly useful in the desalination of saline water. US Patent 5,098,575.

.fereed paper

PROCESS STABILITY OF AN ADD-ON MEMBRANE BIOREACTOR A MacCormick, E Brois Summary Membrane Bioreactor (MBR) tech nology is a potential alternative for conventional post biological treatment and it could make decentralised sewage treatment for urban reuse more technically and economically viable.

This paper describes three months demonstration of a pilot membrane operating system added on co the biological process at the Rouse Hill recycled water plant of Sydney Water Corporatio n ro simulate a membrane bioreacror process.

Over the three month period rhe Memcor tech nology was able ro achieve srable performance under a range of conditions includ ing constant flow, diurnal flow variation, peak flow, low dissolved oxygen concentration , and high mixed liquor suspended solids (MLSS) concentration.

T he simulated MBR treated water was comparable to the plant's current reclaimed water when assessed by physical and microbial parameters and met the NSW Guidelines fo r the Urban and Resid ential Use of Reclaimed Water after chlori nation.

Although a chemical clean was nor necessary during the co urse of the trial, one was done ar rhe end of the three month period and demonstrated full recovery of membrane performance.

Introduction MBR technology uses low pressure membrane fil tration combin ed with biological treatment co produce high quality created effluent. By removing the operational constraint of sludge secrleabiliry in a conventional clarifier the biological activated sludge process can operate at high mixed liquor suspended solids (MLSS) co ncentrations, typically in the 8,000 to 15,000 mg/ I range, and long sludge age.

T hese features lead ro more complete biological treatment of sewage, a decrease of rhe plant footp rint because the MBR membrane operating system requires less space rhan a clarifier, and reduction of the sludge wasti ng volume (Yamamoto et al 1989 and Stephenson et al 2000).

There are two methods for integrating a membrane module into a biological reactor.

The common method is ro suspend the

m

Membrane filter

~-'--11--i Sand filter

Grit removal

Chlorination

Figure 1. Ro use Hill non-potable reclaim ed water p rocess tra in . membranes into the aeration tank, drawing off clear liquor from the MLSS by either gravity or suction. Memcor's "add-on " membrane operating system has the membranes installed in a separate rank, with the aerated MLSS circulated through chis rank, drawing off clear liquor as before.

It is designed ro operate independently of rhe biological reaction and has advantages: • Cleaning the membranes in their separate rank is more cost-effective and environmentally fri endly • Ir is easier ro control and optimise the main biological process, particularly for sludge bulki ng prob lems.

• Retrofit ro an existing biological trea tment plant, ro increase capacity and/or effl uent quality, is simpler to design and instal l.

The current study was designed ro demonstrate, at a pilot scale, process stability over a wide range of operating conditions. Ir also involved analysis for a range of pathogens includi ng naturally occurring viruses. Very little hard data regarding virus rejection by MBR membranes has been published ro dare.

Rouse Hill STP and Pilot Plant Layout Rouse Hill STP The Rouse Hill ST P (Figure 1) is an advanced BNR (biological nutrient removal) facility servicing the expanding North West region of Sydney. The ADWF is 6.3 ML/day serving a current population of 33,000 bur wirh a capacity ro serve 50,000 people. The treatment plant produces tertiary treated effluent for river discharge and reclaimed water supplied to homes via a dual water distribution system for non potable urban reuse (toiler flushing,

Results of a pilot trial at Rouse Hill, Sydney.

garden watering, ere). The non potable water guidelin es set by rhe regulators for internal home use include protozoa, bacteria an d virus removal.

The current process path at Rouse Hill (Figure I) is more complex than conventional serried activated sludge.

Screened sewage is serried and th e solids pumped ro rhe ferm enter stage of rhe biological process. The biological reactor comp rises a pre-fermenter and anaerobi c stage followed by anoxic and aerobic zones.

Mixed liquor from rhe biological process passes ro the clarifier for seed ing. Solids are returned ro the anoxic zone. T he liquid undergoes tertiary treatment involving flocculation, settling, sand filtration, super chlorination and de-chlorination before discharge into wetlands. Effl uent for reuse passes from deep bed sand fil tration through membrane fil tration and super chlorination.

Pilot plant configuration and description T he Memcor pilot plant was designed as an "add on" installation. Ir has an average d ry weather flow (ADWF) capacity of 24m 3 /day and peak flow of 39m 3 /day. T he trial was designed ro simu late che simple process path shown in Figure 7, at the end of this paper, where che clarifier and sand filter are replaced by rhe membrane system, which allows more control of the MLSS. A side-scream of Mixed Liquor was pumped from the Rouse H ill BNR mixed liquor chan nel via a 1-mm pre-screen into an agitated holding rank ro control feed rate and MLSS concentration (Figure 2).

From the holding rank it was pumped into rhe membrane rank where ir was mixed with air before entering the membrane modules. Filtrate was drawn through the modules under suction at a controlled rates ranging from less chat 5 ro 40 LMH (litres per square metre of membrane area per

Journal of the Australian Water Association

Water

MARCH 2007 49

technical features refereed paper

hour - see Figure 3) . Membrane rank overflow was either returned to the biological process or partially to the hold ing rank for MLSS concentration control. Filtered water after sampling was discharged to the Rouse H ill plant.

Membrane tank and membrane

Return to increase concentration

, Chlorine dosing

Q

Screen Turbidlmeter Overflow Pressure TMP

The pilot plant was sufficiently automated to allow unattended operation during most of the day and over weekends. The plant was equipped with PLC control to manage all ro uti ne functions and Mem log® for continuous data logging of temperature, pressures, flow rares and turb idity.

From bioreactor

SP

Sample points were located for mixed liquor ahead of the membrane rank and for filtrate before and after chlorination (Figure 2) .

@ Membrane air supply

Sample points

Fully treated effluent

Blower

Membrane cleaning

MBR processes typically require two types of clean ing - maintenance cleans and more rigorous Clean in Place (CIP).

Maintenance cleaning is needed as a preventative measure to disinfect and control bio-regrowch in the fil trate pipework and membrane fibre lumens; and co reduce and/or control deposition of foul ing materials on membrane surfaces.

Th is was done every 7 to IO days and involves back washing the membrane with sodium hypochlori ce solution followed by a I O minute soaking period and then repeating the process. Draining the membrane rank is nor requi red. A more intensive C lean in Place was not required duri ng the three month period bur was carried our at rhe concl usion of the trial to assess the effectiveness of membrane recovery.

Figure 2. Pilot Plant and Trial Configuration .

quality and MBR process stability was affected by variations in diurnal and peak flow, dissolved oxygen concentration, mixed liquor concentration, stop/scare operation, and to measure performance at the limits of the un ir's design parameters.

Stable performance was maintained throughout the 3 month period (Figure 3) with little increase in trans membrane pressure (T MP) apart from the fi nal two weeks when MLSS was increased above the maximum design level to evaluate how well the system recovered. A peak flow test (flux increased from 25 to 40 LM H - Point A in Figure 3) was run successfully for 16 hours.

There were two occasions (po int B in Figure 3 bur seen more clearly in the expanded scale of Figure 6) when a PLC fau lt and blockage in the wasting line caused the sludge co concentrate to 25g/L

Hydraulic Performance The srudy involved a series of trial runs over three months to quantify how filtrate

(about rwice the design level), the resulting runaway TMP rise triggering a plant shut down . The membrane performance, however, recovered after the system rerurned to normal condition .

A Clean in Place was done on 15 May point C in Figure 3.

The impact on sysrem performance resulting from changes in the operating parameters and the Clean in Place are described in rhe fo llowing subsections.

Baseline (constant flow) performance

The fi rst month was co establish baseline performance. The plant was run at a constant flux of 25 LMH (litres per square metre of membrane area per hour). T MP rose marginally from 12.3 to 13.3 kPa (Figure 3). During this period MLSS in the membrane tank was relatively constant at 8 co 9 g/L. Low dissolved oxygen

45 40

.,."' !l.

• Flux. LMH

...

.

30

.,

25

:E

20

>< ::s

15

.

.

o!!

:i::

a

35

!l.

:E

a TMP • kPa

......,..........

..J

U::

-....... -

a

I

a

'

8

10

B

o:z::a:m.

5 Conatant flow operation

0

•

:.:...

T here are lirerarure references co poor membrane filrerabiliry on anoxic sludge or MLSS with low dissolved oxygen. To evaluate the effect on the Memcor membranes the feed was pumped from the RAS channel (from the settlers) during the latter half of the constant flow rest period.

The low dissolved oxygen appeared co have little if any impact on membrane performance and was left low for the remainder of the study. Membrane performance is measured in terms of permeabiliry and is described in the next section.

Diurnal flow operation

_J__:~==;::=:::;::::::=~~=::;::=;::::::::::::;::=~~:;::::::::::;::::::c:::;::~

25/2

4/3

11/3

18/3

25/3

1/4

8/4

15/4

22/4

29/4

6/5

Time - day & month

Figure 3. W hole Period Hydraulic Performance.

50 MARCH 2007

Water

Journal of the Australian Water Association

13/5

20/5

27/5

Diurnal flow operation

During the second month of the srudy the unit was run to simulate actual diurnal flow co nditions experienced at Rouse Hill.

The Rouse Hill sewage catchment is almost

membrane technology entirely domestic and as a result sewage in flow drops to near zero in the early hours of the morning. It was thus necessary to rest operation of the MBR pilot plant under these wide diurnal flow co nditions.

The pilot plant was stopped for 2 hours each day and membrane feed adjusted in two hourly increments to mimic the actual diurnal flow. Flux (the measure of flow expressed as litres per square metre of membrane area per hour or LMH) is controlled by adjusting T MP. It is therefore more meani ngful to look at membrane performance in terms of permeability, namely fl ux per unit of trans membrane pressure (LMH/Bar).

Membrane permeability remained relatively constant (Figure 4) notwithsta nding wide flu x variation from IO to 31 MLH thus demonstrating the membrane's capacity to handle diurnal flow fil tration satisfactori ly.

ll

refereed paper

45

'5° al

175

40

170

35

..

z ::E

Q.

d. 165

30 ~

(.)

Q.

.,,g,

::E

o

160

f

155

"'

. e

.Q

I25 olS

:c ::E :::..

20 >< ::,

Li:

~ 150 ·

15

145

10

140 - ---,- - - - , - - - - - - - , - ---,-- -- - - - , - -- - - - - - - - - - '·· 5 21:36 0:00 2:24 4:48 7:12 9:36 12:00 14:24 16:48 19:12 21:36 0:00 Time (hours)

Figure 4. Example of Diurnal Flow Performa nce over one day.

When the unit was retu rned to constant flow operation prior to the high MLSS phase the TM P had risen from 13.3 to 15.6 kPa (F igure 3) . It is likely that the rise in TMP was due more to the steady rise in MLSS co ncentration over the diurnal flow period (from around 9 g/L to I Sg/L) than to fou ling (Figure 5).

220

(

Diurnal flow operation ) (

High MLSS Cone

22,000

>

200

· 20,000

180

18,000

16,000

Stop/ start operation

cl,

.§.

en en

A shut down study was performed part way through the diurnal flow phase to simulate a 2 hour equipment/power failure by shutting down all equipment and air supply. After restart no blockage was fou nd and a marginal permeability in crease was observed. A comparison for three days before and th ree days after the rest sh owed no permeability decline.

High MLSS In any MBR system, membrane performa nces will drop and filtration modules will need more air as mixed li quor co ncentration in creases in the membrane tank. There is a trade off therefore, between rhe biological benefits of high MLSS concentrations and the capital and operating cost of the memb rane system; namely, increased air consumption and more membrane area due to lower flux.

T he high MLSS test was run during the last month of the study. Ir was designed to operate the system at the limit of its operating window and to quantify the impact of a higher mixed liquor concentration on membrane perfor mance.

The rest was run in two stages. In the firs t srage MLSS concentration was co ntrolled by reducing the Mixed Liquor d ischarge rare, in order to concentrate MLSS in the membrane tank. Concentration was

:,

..J 14,000 :Ii

.

120

Permeabllity (LMH/Bar)

12,000

Maintenance Clean CIP 100

0- MLSS merrbrane tank 80

10 ,000

MLSS concentration

- - --.----,-....-----------i------+---,--------,-__._ 8 ,000

19-Apr

23-Apr

27-Apr

1-May

5-May

9-May

13-May

17-May

21-May

25-May

Figure 5. M LSS impact on permeab ility.

200

50

.

180

-~ . .

160

-.:-

" ~

a

:Ii

:::!. ~

• .

Cl ,-..

•

•

140

~

\

-

a

• "1 ,.__ "-•

120

:c...

61!131-

_,

45

40

:c

35

;:;:

30

:9.

~ )(

:,

u "' 0,

Q.

25

.

80

20

"-

60

15

~ 100 m

E

"-

~

Q.

~

~

40+--,----,----,---,--.,----,----,----,---,--,----,----,---,--.,--+-,--+10

1/5

2/5

315

4/5

515

6/5

7/5

8/5

9/5

10/5 11/5 1215 13/5 1415 15/5 16/5 17/5

Figure 6. Peak Flow & Clean in Place.

Journal of the Australian Water Association

Water

MARCH 2007 51

refereed paper

maintained between 10 and 12g/L, i.e. close to the design maximum for the system. During chis period there was little change in permeability (Figure 5).

In the second stage, MLSS concentration was raised above the reco mmended maximum and maintained between 13 an d 18 g/L by recirculation mixed liquor to the holding tank. Membrane permeability dropped from around 170 to 140 LMH/Bar when exposed to the higher concentration (MLSS greater than 18 g/L) (Figure 5) .

To test the membrane under peak flow and high MLSS conditions flux was increased from 25 to 40 LMH for 16 hours over the 13th and 14th of May (Figure 6) with the MLSS concentration at 12 co 13 gm/L, i.e. close to the design maximum. The permeability of a membrane under stress conditions usually drops more rapidly than during normal operation and requires extra ti me to recover.

The flux of 40 LMH (achieved by raising TMP) was maintained under stable conditions of steady permeability suggesting the membranes could handle a monthly peak flow without damage or irreversible fo uling. The rapid dips in permeability (points A and B in Figure 6) were due to equipment failure.

Membrane recovery Over time the membrane module becomes increas ingly fouled resulting in a steady increase in TMP. Backwashing and maintenance cleans remove most of rhe fouling bur the system will reach a point, for example TMP greater than 50 kPa, where a more intensive clean is required.

Although permeabili ty during the course of the study did not decline to the point where a chemical clean was necessary it was decided to undertake a clean to demonstrate membrane recoverability. As coagulant is added in the Rouse H ill process for phosphorus removal, a dual chlorine and citric acid clean was used to remove organic and inorganic foul ing, respectively. Permeability after the clean (point C in Figure 6) exceeded the start up permeability, which ranged between 185 and 190 LMH /Bar, showing the membrane's full recoverability with chemical cleaning.

MBR Treated Water Quality T he second purpose of the study was to evaluate the capacity of the process to meet the NSW Guidelines for the Urban and Residential Use of Reclaimed Water and to compare against the existing Rouse Hill process chat included coagulatio n, seeding,

Water

MBR - before Rouse Hill recycled water with disinfection chlorination

MBR - after chlorination

Parameter

Measure

Recycle Guidelines

Cryptosporidium Giardia

cyst/50L

<1

<1

<0.05

<0.05

cyst/50L

<1

<1

<0.05

<0.05

adenovirus

org/50L

<2

<1

<1

<1

enterovirus

org/50L

<2

<1

<1

<1

reovi rus

org/50L

<2

<1

<1

<1

norovirus

org/50L

<2

Negative

Negative

Negative

hepatitis A

org/50L

<2

Negative

Negative

Negative

org/50L

<2

Negative

Negative

Negative

Total Coliforms

cfu/l00ml

<10

<1

Max 320

<1

Faecal Coliforms

Cfu/l00ml

<l

<0.1

Max 1.2

<0.1

rotavirus

Peak flow

52 MARCH 2007

Table 1. Summary of the MBR water quality com pared to the Reclaimed Water Guidelines a nd the Rouse Hill recycled water quality.

Colour

TCU

<15

Highest 8

35

5

Turbidity

NTU

<2 & 95% <5

<0.1

<0.1 max 0.2

<0.1

sand filtration , membrane filtration and super chlorination.

Turbidity, pH, dissolved oxygen and temperature were measured using standard on-line instruments. Analyses for mixed liquor concentration, suspended solids, ammonia, nitrogen, phosphorus, and co lour were undertaken on grab samples by Sydney Water.

The presence of viruses, protozoa and bacteria, as well as turbidity confirmation, was determined by Australian Laboratory Services. Total coliforms and faecal coliforms were tested using a standard membrane filtration method (Australian Standard Method 4276.5 and 4276.7, respectively). Virus detection involved water concentration by ultrafiltration fo llowed by polyethylene glycol precipitation. The presence of viable adenovi rus, enterovirus and reovirus was determined by cell culture assay using an in-house method . Virus concentrations were ascertained by determining the sample dilution char produced 50 per cent of infection of inoculated cell culture (TCIDsol- The presence of norovirus, hepatitis A virus and rocavirus was determined using an in-house PCR method, as these viruses could not be readily isolated by cell culture. Detection of Cryptosporidium and Giardia involved water co ncen tration using ulcrafilcracion followed by additional membrane fil tration, and immunomagnecic separation.

Cryptosporidium oocysts and Giardia cysts were stained with fluorescendy labelled specific monoclonal antibodies, and visualised and enumerated by epifluorescence microscopy.

Microbiological Although pathogen analysis was limited to five feed and nine filtrate samples between

Journal of the Australian Water Association

14 March and 24 May the results were consistent.

No Cryptosporidium or Giardia were detected in any of the MBR filtrate samp les compared to a maximum 186 and 655 cyscs/50L, respectively in the feed.

Similarly none of the six viruses were detected in the filtrate although the feed was positive for all viruses and contained co ncentrations as high as 5.7 x J05 adenovirus per SOL in one sample. This demonstrated a 5.7 log virus removal although only with chis one sample and Vll"US type.

The Memcor membrane used in the pilot trial has a nominal pore size of 0.1 micron and would be expected to remove all particles around this size and above. The fact that particles smaller than the membrane pore size were rejected by the membrane is explained by the propensity of viruses and other particles present in a natural feed stream to agglomerate in to larger particles and become part of the dynamic sludge layer on the membrane surface.

Testing for microorganism removal in drinking water applications using clean and fo uled membranes shows a virus rej ection difference of 1 to 2 logs. However, anecdotal evidence from MBR and secondary sewage applications is that the expected reduction in virus rejection fo llowing cleaning may be less than that for drinking water. In this project there were insufficient samples to confirm chis.

Total rejection of faecal coliforms was observed (feed samples around J05 cfu/ l 00ml) except for a period during the Clean in Place when a test result showed 1.2 cfu/l O0ml in the MBR filtrate (Table 1). The probable cause was cross contamination resulting from a valve that separated the mixed liquor line from the filtrate line remaining open during the CIP.

Table 2. Physical & Chemical Properties.

Parameter all mg/ Litre

Effluent from plant ofter

MBR filtrate

After secondary clarifier

Average

Range

Samples

Average

<2 3.09

*

16 39 39 39

NA 4.14 0.042

0-0.19

1.59 6.47

6.5 - 4.8 1.6 - 13.6

BOD Nitrate Ammonia Ortho Phosphate Suspended solids

0.023 0.094 0.21

0.6 · 8.7 0-0.18 0. 13 - 3.33 0 to 0.7

53

Range

2.2 - 6.2

coagulation & sand filter

Samples

Average

Range

Samples

0

<2 4.11

<2 • 4 2.2 - 6.5 0 - 0. 93 0.16- 0.32 0. 1 - 7.7

15 62 62

63 58 59 11

0.065 0.105 1.19

67 67

* All samples were <2 except one at 26 mg/ L. However the COD for this sample was 74 compared to the mean of 86 mg/ L Total coliforms in che MBR filtrate ranged between 13 and 320 cfu /100ml, which was higher than expected (Table 1). Ir is possible char biofilm regrowth and cross contamination in the filtrate lines (durin g CIP and sampling) may have been the cause. The potential for aerosol cross contamination during sampling was also taken into account. Indeed an exposure plate placed near rhe sample point during sampli ngs indicated the presence of airborne coliforms (7 cfu per plate). The airborne coliforms can be explained by the presence of a low speed aerator adjacent co bu r nor part of the MBR pilot.

Chlo rination of the MBR filcrare, consistent with rhe Reclaimed Water G uidelines, eliminated any remaining coral co liforms (Table I).

Physical and chemical

T urbidi ty was well below the reuse guideli nes. Over che 3 months trial it ranged from 0.05 co 0. 15 NTU with one reading of 2 NTU. Suspended solids were lower fo r the MBR filtrate than the current treatment afte r coagulation and sand fi ltration stages (0.2 1 and 1.19mg/L respectively Table 2).

As expected, without the coagulation process, MBR removed little true colour (Table ! ). Colour, however was reduced co below che Guideline value fo llowing chlori nation (Table I).

As rhe mixed liquor concentration fo r the whole bi ological process co uld nor opera te at rhe "higher MBR level" the objective was co demonstrate cha t the fi ltrate quality was at least no wo rse th an the recycle plant effluent from a chemi cal perspective.

T his was the case as mean values fo r BOD, nitrate, ammonia, orthophosphate and suspended solids were co mparable between rhe MBR and the cu rrent treatment process (Table 2). T he one exception was when arcificial conditions were imposed on the MBR operating parameters such as rapidly increasing che MLSS concentration.

Conclusion

D uring the course of rhe three month study rhe Memco r Membrane Operating System demonstrated stable perfo rmance under co nstant and variable operating conditions and required minim um maintenance. The reclaimed water quality with chlorination was comparable to the existing Rouse Hill plant and well within the NSW Guidelines for the Urban and Residential Use of Reclaimed W ater.

T he study confi rmed growing evidence chat MBR technologies have become very competitive in recent rimes because they produce treated water comparable to che best available from conve ntional technology with less capi tal outlay and space required (Figure 7). Because they eliminate problems associated with sludge bulking they are proving eas ier co operate.

As such they will play an important role in addressing rhe urgent need fo r more sustai nable water man agem ent.

Acknowledgments

The authors have written chis paper based on the original technical report prepared by Etienne Brois with contributors from Roger Ph elps, Fufan g Z ha, T ina Nguyen and Kevin Gabbert. ln addition to rhe original writers rhe author is indebted co Mark Angles of Sydney Water for his review and numerous improvements in rhe manuscript.

The project team gracefully acknowledges the co ntributions of Anica Pangesru (Sydney Water) fo r her work and support throughout the project; Tung Nguyen (Sydney Water), Nigel Barrett (Syd ney Water) and Michael Boake (V eolia W ater Australia) fo r their techn ical s upport; Ian McNee (Sydney Water) for h elp in facilitating the water analyses ; and the Rouse Hill Recycle Water Pl a nt staff fo r making the proj ect poss ible o pera tionally.

T he project was fund ed by Sydney Water Corporation and Veolia Wate r Ausrralia.

The Authors Tony MacCormick is involved in business

development with Memcor A ustralia Pry Ltd , the membrane filtration a rm of Siemens Water T echnologies. Email an tony. maccorm ick@si emens. com; Etienne Brois, at the time of chis srudy, was Development Enginee r with Memcor Australia Pry Ltd. Following V eo lia's sale of US Fi leer and Memcor co Siem ens, Etienne remained with Veolia and is now with OTV. References Ya rnarnoro K, Hiasa M, et al, Water Science and Technology, 2 I (4), 43, 1989.

Stephenson T, Judd S, et al, Membrane Bioreactors for Wastewater Tream,ent, IW A Publishing, London, 2000.

Current non potable reclaimed water process train

Membrane filter

m

,r.,mn,mn ; ~

Sand filter

Grit removal

Chlorination

Possible process water process train using MBR technology

m ~·,,;;,';i;;; { ;;. ·:,__ .,, _

Screens

Grit

Chlorination

,---·! l~- -,,r~ ~~~e

:5.v_ a'__.....,__

___

T

Biological process

~

water

Membrane system

Figure 7. Potential process simplification with MBR .

Journal of the Australian Water Association

Water

MARCH 2007 53

efereed paper

ODOUR CONTAINMENT AND VENTILATION AT PERTH'S MAJOR WWTPs K Codee, I Wallis Abstract Experience at con troll ing odours by retrofit of covers and ventilation systems at the three major wastewater treatment plants in Pe rth has led to recommended d esign criteria for near complete capture of odours from a high sulphide, high temperature wastewater.

Introduction Odours from wastewater treatmen t plants in Perth have caused complaints from nearby residents for many years. Recen tly the Water Corporation made significant investments at its th ree largest wastewater treatment plants to achieve a major improvement in odour control at these plants by retrofitting covers, ventilation and scrubbing systems.

T hese odour control works at the plants were constructed in stages and incorporated different design app roaches enabling the effectiveness of different approaches to be co mpared . As a result, a range of air extraction rates has been exam ined and it has been established that the effectiveness of covers to contain odours de pends on the capture velo city which is a function of the negative pressure developed under the covers. T his, in turn, is a functio n of the venti lation flux rate based on the surface area covered.

Generalising these results, design criteria are recommended to achieve near complete odour capture with the best examples of covering system s used at Perth treatment plants on recent odour control p rojects.

Odour containment has been successfully achieved at other treatment pla nts using lower ventilatio n rates than recommended in this paper. H owever, based on observation at many other plants, this success common ly reflects either lower ambient temperatures or substantial ly lower sulphide levels in th e influen t. Perth has a warm climate and very high sulph ide levels in the wastewater; thus a very high level of capture of odours is essential. This is more

Recommended design criteria for near complete capture of odour.

54

MARCH 2007

Water

Figure 1. Subiaco Wastewater Treatment Pl ant.

difficult to achieve where odour con trol is a retrofit rather than considered as part of the initial design.

Details of Covers and Odour Control Facilities Woodman Point Wastewater Treatment Plant (WWTP) T he Woodman Point WWTP serves Perth's southern suburbs. Wastewater is pumped to the plant and is septic, with hydrogen sulphide levels in the gas space of the inlet channel being typically in the range of 150 to 30 0 ppm.

The treatm ent fac ilities were upgraded in 2001/02 incorporating extensive covering and odour co ntrol. The screens and channels were covered with traffi cable aluminium covers, while the grit and p rimary sedimentation tanks were covered with purpose built aluminium covers. All covered areas were ventilated to two , single staged scrub bers employing hypochlorite/caustic and Odorgardâ&#x201E;˘ reactors. Afte r scrubbing, gases were d ischarged via a common 22 m high stack.

Journal of the Australian Water Association

Beenyup WWTP The Beenyup Po int WWTP serves Perth 's northern sub urbs. About half the wastewater is from large gravity sewers and the oth er half is pumped. H ydrogen su lphide levels in the gas space of the inlet channel are typically in the range of 100 to 300 pp m.

The odour control facilities at Beenyup WWTP were upgraded in three phases between 2003 and 2005.

In P hase 1, the inlet chan nels, screens, grit removal and primary sed imentation tanks were covered with fibre reinforced plastic (FRP) covers. Foul air was treated in two sets of two staged scrubbers employing acid and hypochlorite/caustic with Odorgardâ&#x201E;˘ reactors.

In Phase 2, the aeration tanks and dissolved air flotatio n thickeners (DAFT) were covered. Centrifuges were also installed to dewater d igested sludge. Foul air from the centrifuges, co nveyors and sludge hop per was co llected and ducted to the Phase 2 scrubbers. The ventilation rates from the inlet sewers and screening facilities were

refereed paper

significantly increased. Only rhe secondary clarifiers remai n uncovered.

control faciliries. O nly the secondary clarifiers remain uncovered (Figure 1).

the internal and external temperatures. T his is commonly known as rhe ch imney effect.

Foul air collected from rhe Phase 2 works was rreared in rwo staged scrubbers employing hypochlorire/causric and Odorgard reactors. A 50 m rail srack was constructed to disperse rhe air from rhe Phase 1 and Phase 2 scrubbers.

Ventilation rates from the pre-treatmen t facility were increased, and the building was vent ilated ro the scrubbers ro achieve seco ndary co ntainment. Foul air from the DAFT, sl udge blending ranks, centrifuges, and lime ame nded sl ud ge process was collected and vent il ated ro the scrubbers.

The pressure difference due to the chimney effect can be expressed as:

In Phase 3, the Phase l and Phase 2 ducts were interconnected and rhe venrilarion rares were in creased. The firs t srage of rhe Phase l scrubbers was also modified to operate on a caustic solution rather rhan acid solution and pressure sensors were installed to mon itor the negative pressures under the covers. The observed variations in negative pressure under the covers provi ded an insight into the complexity of extracting air from large covered areas wirh mul tiple tanks and ducts.

Subiaco WWTP The Subiaco WWTP serves rhe central suburbs of Perth. Most of the wastewater reaches the plant by gravity sewers bu t about a third is pumped to the plant.

Hydrogen su lph ide levels in the gas space of the inlet channel are typically in the range of 100 to 200 ppm.

In 2003/04, the primary sedimentation ra nks, aeration tanks and associated chann els and chambers were covered with FRP covers and ventilated to the odour

Foul air from the inlet sewer, pretreatment, primary and sludge trearmenr facilities was com bined and treated by two staged scrubbers consisting of acid and caustic to remove ammonia and hydrogen su lphide respectively. T hese gases were rhen combined with the gases extracted from the aeration ranks and treated by further rwo staged scrubbers employing hypochlorire/ caustic and Odorgardâ&#x201E;˘ reactors. After scrubbing, all gases were discharged via a 50 m high stack (Figure 2).

Theoretical Considerations for Mechanical Ventilation

Two forces must be negated by mechanical ven til ation to achieve a high degree of odour containment. The firs t is thermal buoyancy caused by the difference between

(1)

For most odour containment a pplications, tiP, is relatively small, however, the chi mney effect can cause a significant circulation through buildings h ousi ng large thermal loads.

The most important force for odour conta inment is usuall y the pressure difference created by air movement over the surface of the cover.

(2)

The pressure P0 on the ou tside surface of the cover may be calcul ated as:

p o

=p

_ Cpp[K,.,VJ

a

2

2

(3)

The effect of wind may be represen ted by the global wind pressure coefficient (Cw0 ), defined as the fraction of the dynamic pressure of rhe free wind acting on the outside of rhe structure. Thus:

Figure 2. Sc rubbers and Stack at Subiaco Wastewater Treatment Plant.

Journal of the Australian Water Association

Water

MARCH 2007 55

p _ CwopV,..'

2

ti." -

Table 1. Topography and Height Factors.

ti.P

(4)

The velocity across an opening may be derived from Bernoulli's equation: V

=Cc1 [2M]½ p

(~

Topogrophy

K

a

Open Country - Few Windbreaks

0.68

0.17

Open Country - Many Windbreaks

0.52

0.20

Rough Country/Outskirts Small Town

0.35

0.25

City Centre

0.21

0.33

Source: C/BSE Guide A2-31, Tobie A2. 11 ( 1999)

Combining equations (4) and (5), the velocity across an opening that will negate the effect of wind action on the cover (ti.Pw = 0) is given by: (6)

The effect of height and topography on the wi nd velocity acting on the covers may be estimated using in formation from Table l , where Kw = Kza Where the maximum height of the covers (z) is 1.5 m above ground, and the topography may be characterised as Open Country - Many W indbreaks, and assuming Cd= 0.6, and CP = 0.9 (AS/NZ 1170.2, 2002),

Cc1Cwo 112 = 0.6 (0.9(0.52 X 1.5°·2)] 112 = 0.32 Capture Velocities

Recommended capture velocities for a range of indoor industrial operations, in the absence of sign ificant wind effects, are listed in T able 2.

USEPA regulations for the control of hazardous volatile organic compounds (CFR 40, Chapter 1, Part 52, Appendix B) assume a capture efficiency of 100 % for a captu re velocity of lm/s. In the field of fire protection, pressures of -25 Pa (capture velocity of 3.9 mis) under all conditions (NFPA 820) are recommended to contain potentially explosive atmospheres. Pressures of -20 Pa (capture velocity of 3.5 mis) are recommended to control the spread of smoke in buildings (AS 1668. 1:1998).

Degremont (2005) states chat pressures of at least -7 Pa (capture velocity of2 mis), will generally be adequate to prevent the escape of odours under low wind speeds.

Capture velocities of2 mis have also been demonstrated to achieve near complete capture of vehicle emissions in road tunnels (CEE, 2001).

For practical purposes, near complete capture of odour may be assumed where a captu re velocity of at least 2 mis fo r large odour sources, and 1 m/s for small to moderate odour sources, are maintained at all times. While this is relatively straight forward for small, indoor enclosures, it is more difficult to achieve when covering large process ranks at wastewater treatment plants. At the Subiaco plant, for examp le, there are 8.8 km of seals between covers.

Careful attention to derail is imporranc in the design, installation and operation of cover systems to ensure that the leakage area is as low as possible to minimise the ventilation rates req ui red to achieve adequate capture velocities.

Negative Pressure and Odour Containment

W ind effects have a profound effect on odour release from a liquid surface (Schulz et al. , 1996). Even extremely low velocities may have a significant effect on odour release. Passive covering techniques such as addi ng a layer of straw or tarpaulins can reduce odour emissions by as much as 80 to 90% Qacobson et al.) . This was confirmed during the installation of covers in Perch, where significant reductions in odours was noticeable when the covers were in place, bur before the ventilation systems were com missioned.

Where the area of openings is known , the pressure required to achieve a particular velocity through the opening may be determined by rearranging equation (5).

pV2

M= - -2 2Cc1

(7)

The recommended captu re velocities fo r odour containment in indoor situations are commonly in the range 0.5 - 1.0 mis.

Assuming = 1.2 kg/m 3 and Cd= 0.6, it may be calculated from equation (7) that the negative pressure will be 1.7 Pa fo r a capture velocity of 1.0 mis. If pressure fluctuations under the covers are greater than this negative pressure, the capture velocity must be increased.

The preceding discussion assumes that wind velocities are constanc vectors. Actual winds are high ly variable. Examples of this vari abi lity are shown in Figure 3, which are based on wind measuremencs at the Subiaco WWTP. Variations in wind velocities resu lt in fluctuations in pressures underneath the covers.

While low wind speeds are usually associated with most odour complaincs, ventilation systems sho uld be designed to concain odours under all but the most extreme wind conditions.

Measurements of the pressures underneath the covers at Subiaco WWTP indicate char pressure fluctuations of up to 10 Pa occur.

By targeting a static pressure of - 17 Pa (capture velocity of 3.2 mis) negative pressures of at least -7 Pa (capture velocity of 2 mis) can be achieved at all times.

Experience shows that it is difficult to achieve strong negative pressures under all covers at all times where multiple ranks are ventilated from a co mmon manifold, and where the ai r supplied to different aeration tanks is uneven. In these situations, the static pressure may need to be -20 Pa or even -30 Pa to ensure effective odour capture.

T he capture efficiencies listed in Table 3 are based on rhe experience, observation and judgemenc of the authors for covers subject to wind action in Perth. These are conservative estimates of capture efficiencies. The capture efficiency at Subiaco prelimi nary treatment area was

Table 2. Recommended Capture Velocities (ACGIH, 2001 }.

Condition of Containment

Example

Capture Velocity

Release with Practically No Velocity into Quiet Air

Evaporation from Tanks, Deg reasers, etc

0.25 - 0.5 m/ s

Release al Low Velocity into Moderately Still Air

Paint Spray Booths, Intermittent Conta iner Filling, Low Speed

0.5 - 1.0 m/s

Active Generation into Zone of Rapid Air Motion

Spray Pa inting in Shallow Booths, Barrel fill ing

1.0- 2.0 m/s

Release at High Initial Velocity into Zone of Rapid Air Motion

Grinding, Abrasive Blasting, Tumbling

2.5 - 10 m/s

56 MARCH 2007

Water

Journal of the Australian Water Association

refereed paper

Table 3. Capture Efficiency for Covers S ub ject to Wind Fo rces in Perth.

0 .90

080

Static Negative Pressure (· Pa)

Capture Efficiency 11 )

<5 5- 10 10 - 15 15 - 30

95 95 -99 99 - 99.9 99.9- 100

(%)

070

~ 0 ,60

loso '!1

"i 040 03-0

It may be seen char the venti lation flu x rare q, and the equivalent unit lea kage area a, are the two most importa nt factors governing the effectiveness of odour containment systems.

0 20

0.10

000+-~ ~ - ~ ~ ~ ~ _ , _ .._,~

~ ~ - ~ ~ ~ ~ ~ ~ - ~ ~ . . . . . , . - - . . . . , _ - ~~

15/12 1$,112 15/12 15112 15/12 15112 15112 15/12 15112 15112 15112 15112 15112 15/12 15112 15/12 15/1 2 15/12 15/12 15112 15112 15/12

00:00 00:1• 00-2e 00-43 oo s1 01:12 01 2a 01 ., 0 01 ss 02-09 02:2-4 02 38 02 s2 0301 oJ.21 oJ:36 oJ.so

°" °" ~·,; 04 33 04 , a os 02

Tim•

Dilution of Odours and Other Contaminants

1.00

II

I

0 90

The concentration of contam inants underneath covers may be reduced by diluting foul air wirh clea n air. High odour co ncen rracions under covers r esult in fugitive emissions being of similar high concentrations.

000 070

f

l ~

'i

~~

080 O SO

o.,o 0 30

020 010

000

II

\,

~~

For most wastewater creacmen t processes, the rate of contami nant released is proportional to the area of rh e liquid surface and may be characterised by a co ntaminant flu x rare Z. H en ce the concentration of contami nan c in the air under the cover may be expressed as:

I

~

~

i

~

1-4112 1-4/12 1-4/12 14112 1-4/12 14112 14/12 1-4/12 14/12 1-4/ 12 1-4/12 14/12 14(12 1-4/12 1<4(12 14/12 14/12 14/12 1,4/12 14/12 14112 14/12

oo·s1

0112 012s 01 :40 01 :ss 0200 022o1 02-311 02s2 0307 03.21 03:Je OJ.so 04-04 0.19 0433 o,ue oso2

os.,s

os31 os•s 0600

Time

Figure 3. Variability of Wind Speed Measured a t Subiaco WWTP.

(1 1) For most situations, A1 = Ac, h ence:

measured ar 99. 7% in rhe absence of wind effects (CEE, 2006). Where fugitive emissions based on these capture efficiencies co ntribu te to a significant proportion of the offsice impacts, direct measurement or assessment of the actual leakage of fug itive emissions is reco mmended.

Ir is nor possible to directly calcul ate the leakage area fo r most cover systems with any precision, as most of the leakage occurs at join ts and seals. The equivalent leakage area (A) may be calculated fo r di ffe rent types of covers fro m the ventilation rate and differential pressure:

C

g

=Iq

(12)

(8) Dividing by the rotal plan area of the covers

(Ac) gives rhe ventilation flux race (q):

[2/1?]½ -A

Q =q=Cd - Ac p

_g

Ac

(9)

Rearranging,

11?-_p__ [ !l..]

- 2c/

a

2

(lO)

Nore that Z is not a co nsta nt and may depend on a number of facto r s including the venti lation rate and turbulence in the li quid phase.

For some locations, such as weirs, drops and ocher points of high turbulence, the rate of contam inant release may be assumed to be proportional ro the wastewater flow race (assuming char the weir heigh t, temperature, pH and pressure are constant) .

Hence:

Journal of the Australian Water Association

Water

MARCH 2007 57

Table 4. Design Ventilation Rates at Woodma n Po int WWTP 111 (2002 Upgrade)

C =XQ, g Q

(13)

g

Equation (13) may be rewritten in terms of the gas and liquid flow rates as: C=~ g GIL

(14)

The contaminan t release factor Xis governed by mass transfer laws. It is proportional to the contam in ant co ncentration in the liquid phase, as well as other factors including temperature, pressure, pH and the turbulence in the liquid phase. Where these may be considered constant, the constant of proportionality is the stripping factor 0 .

Covered Odour Source

Screens and Grit Chambers Primary Sedimentation Tanks

Air Changes Per Hour

Ventilation Flux Rotelli (m 3/ m2 h)

G/L Ratio (m3/m3)12l

16.9 4.4

12.7 3.1

0.7 l.0

(1) The maximum ventilation rates actually achieved were only 80% of the design values.

(2) Based on an average dry weather flow of 160 Ml/d (6667 m3/h) and a peaking factor of 2.0.

Table 5 . Design Ventilation Rates at Beenyup WWTP (2003 Upgrade).

Covered Odour Source

Screens and Grit Chambers Pri mary Sedimentation Tanks

Air Changes Per Hour

Ventilation Flux Rate (m 3/ m2 h)

G/L Ratio (m3/ m3)!11

5.5 5.8

6.5 4.1

0.5 l.3

/1) Based on an average dry weather flow of 120 Ml/d (5000 m3/h) and a peaking factor of 2.0.

(15) The stripping facto r depends on residence time, turbulence and the degree of mixing or overturning of the contents of tanks.

Air Changes Per Hour Air changes per hour (AC/h) have been used historically to determine the ventilation requirements in a wide variety of situations to control the concentration of contaminants in the atmosphere.

For odou r co ntrol at wastewater treatment plants, equations (12) and (14) are more appropriate than AC/h. This is consiste nt with the Australian Standard for the design of mechanical ventilation (AS1668.22002), where the recommended approach is based on the dilution index, with generic requirements based on airflow rates per square metre of floor area rather than AC/h.

Ventilation Rates The design ventilation rates for the 2002 upgrade at Woodman Point WWTP are summarised in Table 4. These facilities fai led to meet the odour co ntrol objectives for the plant as they resulted in minimal negative pressures and unacceptably high levels of fu gitive emissions. Community complai nts and concerns about odour extend to 1.5 km from the plant (CEE, 2005). A future upgrade of the odour control facilities is planned for 2007.

The design ventilation rates for the 2003 upgrade (Phase 1) at the Beenyup WWTP are summarised in Table 5. At these ventilation rates, only minor negative pressures were developed underneath the covers resulting in unacceptably high levels of fugitive emissions. Community complaints and concerns about odo ur extended to 1.5 km from the plane (CEE, 2006). As a result of the failure to ach ieve 58 MARCH 2007

Water

Table 6. Design Venti lation Rates a t Beenyup WWTP (2005 Upgrade).

Covered Odour Source

Air Changes Per Hour

Inlet Gravity Sewers Screenings Plant (including channels) Grit Tanks (including inlet channels to PSTs) Primary Sedimentation Tanks (including d/s channels) Aeration Tanks 121 and Mixed Liquor Channels Sludge Treatment (Various Sources)

n/a 30 30 10 n/a n/a

Ventilation Flux Rate (m3 /m 2 h)

G/ LRatio (m3/m3)lll

n/a 40 20 7 n/ a n/a

0.5 l.9 0.45 2.3 9.7 0.45

( 1/ Based on an average dry weather flow of 120 Ml/d (5000 m3/h/ and a peaking factor of 2. 0. /2/ Based on 20% air extraction above peak aeration air flow.

Table 7. Desi gn Ventilatio n Rates at Subiaco WWTP (2003 Upgrade) .

Covered Odour Source

Inlet Gravity Sewers Screens and Grit Chambers Screenings and Grit Conveyors and Associated Equipment Pre-treatment Bu ild ing (31 Channels (u/s of PST) Primary Sedimentation Tanks Channels (d/s of PST) Aeration Tanks 141and Mixed Liquor Channels Sludge Treatment (Various Sources)

Air Changes Per Hour

Ventilation Flux Rate (m 3/m 2 h)

G/ LRatio (m3/m3)1l)

n/a 30

n/a 40

0.4 2.3

60 12 10 12

n/a 60121

0.55 7.0 0.4 2.6 0.35 11.3 1.3

ll

n/a n/a

7 12 13 n/a n/a

/1 / Based on an average dry weather flow of 61 .4 Ml/d /2558 m3/h/ and a peaking factor of 2.0. /2/ Based on floor areo. /3) Actual operating ventilation rates ore less than half of these values. /4/ Based on 20% air extraction above peak aeration air flow.

acceptable levels of odour control, further works were undertaken in 2005 (Phases 2 and 3) to cover additional process units and increase the ventilation rates to the levels shown in Table 6. This resulted in a dramatic improvement in odour control, with increased negative pressures and far lower levels of fugitive emissions from the preliminary and pri mary treatment areas,

Journal of the Australian Water Association

which are the major sources of odour emissions at Beenyup. Now, community complaints about odour have been considerably reduced to virtually nil beyond about 750m from rhe plane (CEE, 2006).

The odour control facilities at Subiaco WWTP were installed in 2003 with the

refereed paper

Table 8. Recom mended Design Criteria for N ear Complete Odour Containment at Wastew ater Treatment Plants in Perth Covered Odour Source

Recommended Design Criteria for Near Control Odour Capture

Inlet Gravity Sewers

The ventilation rote should be sufficient to negate pressure fluctua tions in the sewer and maintain negative pressures at all times. A ventilation rote of 40-50% of the peak wastewater flow rate (m 3/ h) may be satisfactory based on experience at Beenyup and Subiaco WWTPs. The covers on screens generally contain many openings (up to 1% of the total covered area) and it is not practical to achieve strong negative pressures underneath these covers. The ventilation rate (not for personnel entry) should be the greater of 36 m3/ m2h or 30 AC/ h, to ensure> 99% odour capture.

For well fitted covers designed to ach ieve near complete odour capture (> 99 %1the ventilation rate (not for personnel entry) should be 12 m3/ m2h 11 I. For covers with many openings (not greater than 1 % of total covered areal, and not designed to achieve strong negative pressures, the ventilation rote (not for personnel entry) should be the greater of 36 m3/m2h or 30 AC/ h to ensure> 99% odour capture.

For well fitted covers designed to achieve complete near odour capture (> 99%) the ventila tion rate (not for personnel entry) should be 12 m3/m2h Pl. For well fitted covers, designed to achieve near complete odour capture (> 99%) the ventilation rote (not for personnel entry) should be 20% above the peak aeration air flow rate.

The ventilation rote should be 6 m3/ m2h, for well fitted covers (not for personnel entry) designed to achieve near complete odour capture (> 98%).

The ventilation rote (not for personnel entry) for covered tonks, silos and major items of process equipment should be up to 30 AC/h 121 based on head space volume under normal operations.

The ventilation rote for small items of equipment such as conveyors, and other equipment containing sludge or sludge related material should be up to 60 AC/ h 121 based on empty volume.

The ventilation rote for dra ins and partially filled pipes conta ining sludge or sludge related material should be> 0.75 m/s calculated on on empty pipe.

Where the primary sources of odour and H2S ore covered and contained by appropriate ventilation, the building ventilation rote may be operated at 18 m3/ m2h (floor areal and achieve> 95% odour capture with all windows and doors closed. Capacity should be provided to increase the ve ntilation rate to 36 m3/m 2h (floor area) when covers need to be removed for maintenance. Safe working procedures must address the hazards of H2S during maintenance activities.

Where the primary sources of odour and H2S inside the bu ilding ore not effectively controlled, specific investigations ore requ ired to ensure that the atmosphere inside the building is safe for opera tors, maintenance personnel and visitors.

Screens and Grit Chambers

Channels (Unsettled and Settled Wastewater)

Primary Sedimentation Tonks Aeration Tonks Mixed Liquor Channels Sludge Handl ing Facilities

Secondary Containment Buildings (Pre-treatment and Primary Treatment)

Note ( 1): These ventilation flux rates are recommended for use with we// maintained, tightly fitted covers with few openings such as the covers installed at the Subiaco WWTP. Ventilation flux roles in this range should ensure that static pressures of at /east -11 Pa are achieved underneath these covers. Higher ar lower rates may be necessary to achieve the desired negative pressures with other cover systems.

Note /2): Very tight covers and enclosures with few openings, may achieve the recommended capture velocities of >2m/ s at lower ventilation rates.

des ign ventilation rates summarised in T able 7. T hese ventilation rates achieve static negative pressures under the primary tank covers of at least -15 Pa and generally up to -20 Pa. Fugitive emissions are very low, and the degree of odour containment is estimated to be better than 99%. T here have been no odour complaints in over 2 years and telephone surveys show a high level of community satisfaction with the odour upgrade.

Recommendations The ultimate test of odour control works is the level of co mm uni ty satisfaction with the ou tcome. In this respect, the Woodman Point and Beenyup Phase l upgrades were not successful and fu rther work was required. The Subiaco upgrade, which was designed to achieve better odour co ntainment and included a reliable scrubber system with an emphasis on achieving 99.9% availabili ty, was a success.

Based on our experience in Perch, the recommended des ign criteria is to achieve a capture velocity of at least 2m/s under all conditions. This will provide near complete odour cap cure, which is essen ci al for large

Water Advertising To reach the decision-makers in the water field, you should consider advertising in Water Journal, the official journal of Australian Water Association.

For information on advertising rates, please contact Brian Rault at Hallmark Editions, Tel (03) 8534 5000 or email brault.rault@halledit.com.au

treatment plants with high incoming odour levels. T he vencilacion flux races are high er than commonly used in North America and European practice. T his is believed to be the result of the particularly high odour levels experienced at large plants in Perch and che need to retrofit odour control facilities in to planes not specifi cally designed for odour containmen t.

T he degree of odour nuisa nce d epends on the total odour emissions from che plant.

For small plants, or plants with low incoming sulphide levels, a lower level o f capcu re (say 80 to 90%) may be acceptable.

For large planes in Perch, it is n ecessary to aim fo r 98 to 99% odour capture.

Lower ventilation rates may be used where odour levels are low and thus a lower degree of odour capture is acceptable, or where the design, installation and operation of cover systems enables the recommended negative pressures and capture velocities to be achieved at the

Journal of the Australian Water Association

Water

MARCH 2007 59

refereed paper

lower venrilacion races. This can be achieved only if the equivalent unit leakage area is less than the best examples of covers used in recent odour control projects in Perch. The authors caution against using ventilation rates less than those recommended in Table 8 without the strongest of evidence as to the effectiveness of the proposed cover system and meaningful performance guarantees linked to the achievement of a static negative pressure chat will ensure a cap ture velocity of at least 2m/s is achieved at all times.

Because of the high temperatures and high sulphide levels in Perth wastewater, best practice capture (listed in Table 8) must be used to control odour emissions in Perch.

References AS/N ZS 1170.2: 2002 Structural Design Actions - Wind Action.

AS 1668.1: 1998 Use of Ventilation to Control Fire and Smoke.

AS 1668.2: 2002, The use ofventilation and air conditioning in buildings - Ventilation Design far Indoor Air Contamination Control.

American Conference of Govern ment Industrial Hygienists (ACG IH). fodustrial Ventilation: A Manual of Recommended Pmctice, 24th Edition, 200 l.

Chartered lnsrirurion of Building Services Engineers (CIBSE) Guide A: Environmental Design, 1999.

Consulting Environmental Engineers. Air Quality Assessment of T1mnel Emissions for Eastern Freeway. Report ro VicRoads. 2001 Consulting Envi ronmental Engineers. Beenyup WWTP- Results of Odour Monitoring and Modelling Program. 2006 Consulting Environmental Engineers. Subiaco WWTP- Results of Odour Monitoring and Modelling Program and Recommended Buffer Zone. 2006 Consul ting Environmental Engineers. Buffer Zone far Woodman Point WWTP. 2005 Degremont Handbook, 1Orh Edition (in French), 2005.

Jacobso n L , Lorimer J., Bicudo J and Schmidt D. Livestock and Poultry Environmental Stewardship (LPES) Curriculum, Lesson 43, Emission Control Strategies for Manure Storage Faciliries (Covers) . W\Vw.lpes.org NFPA 820. Standard fo r Fire Protection in Wastewater Treatment and Collection Facilities, 2003 Edi tion.

Schulz, T , Jiang, J and Bliss, P. The development of a sampling system fo r rhe determinarion of odour emission rares from areal somces. J Air and Waste Management Assoc., 1996.

USEPA, 40CFR - Chapter! - Parr 52, Appendix B

The Authors

Keith Codee is General Manager Water T echnologies Division, W ater Corporation, Leederville, Western Australia 600 7, Email keith.cadee@wacercorporacion.com.au.

Ion Wallis is Principal Environmental Engineer, Consulting Environmental Engineers, specialising in air quality, environmental studies and strategic planning. Address is PO Box 20 I , Richmond Victoria 3121, Email wallis@cee.com.au

Definitions

c,

C&SBRAND AUSTRALIAN FILTER COAL FOR DEEP BED COARSE DUAL MEDIA FILTRATION "More UFRVs for your money, and better quality water"

C&S BRAND GRANULAR & POWDERED ACTNATED CARBONS

X

z

z a

JAMES CUMMING & SONS PTY LTD 319 Parramatta Rd AUBURN NSW 2144 Phone: (02) 9748 2309 Fax: (02) 9648 4887

Email: jamescumming@jamescumming.com.au

60 MARCH 2007

Water

Journal of the Australian Water Association

QUALITY ENDORSED COMPANY AS/NZS ISO 9001 STANDARDS AUSTRALIA Licence no: 1628

t.P t.P,

t. T p IIJ

Coefficient Totol leakoge area (m 2) Plan area of cover (m 2) Plan a rea of liquid surface (m 2) Coefficient of discharge. Generally assumed to be in the range 0.60-0.65 for shorp edged inlets and turbulent flow.

Concentration of contaminant in air (mass or odour units/m3) Concentration of contaminant in wastewater (mass/m 3) External pressure coefficient Global wind pressure coefficient Acceleration due to gravity Gas/liquid rotio (Q/01} Separation height between inlet and outlet (m) Coefficient Velocity coefficient relating to topography and the height of the structure Atmospheric (Barometric) pressure (Pa) Inside pressure (Po) Outside pressure (Po) Ventilation flux rate (m 3/ s per m2 of pion oreo of cover) Ventilation rate (m3/s) Wastewater flow rote (m 3/ s) Outside temperature (°K) Velocity across opening (m/ s) Velocity across opening that will negate the effects of wind action Free wind speed measured at the standard height of l Om (m/ s) Contaminant release factor (mass or odour units released per m3 of wastewater) Height of covers above ground level (m) Contaminant flux rote (mass/s or odour units/ s per m2 of liquid surface area) Equivalent unit leakage area (m 2 per m2 of pion area of cover) Pressure differe ntial across opening due to ventilation (P0 -P;)(Pa) Pressure differential due to thermal buoyancy (Po) Pressure differential due to wind action (Po) Temperature differential (0C) Air density (kg/ m3) Stripping factor

technical fea u es

sludge drying

CYCLONIC THERMAL DRYING OF BIOSOLIDS C Lane Abstract

Recovery O rganization. (www.grro.net)

Floe Water 1525%

Biosolids have been a problem for The primary pathway of water Capillary operarors of sewage creacmenc planes Water1-2% removal is the mechanical removal of because che rradirional on-sire the liquid. Cyclonic thermal drying Bound Water dewatering approach does not utilises the effect of high volum es of 1-2% produce a market-ready product. A high velocity air, in centrifuga l Matrix Water new system, Cyclo nic Thermal rotation, flowing ove r multitudes of 1-10% D rying, is being trialled by Logan exposed material surfaces, created by City Council ro determine its high races of collisions and impacts. potential ro reduce volume, weight, T he cyclonic process and equipment odours and removal cost. A pilot trial causes separation of the various is producing results that meet chis material compo nents via differences objective as well as rapid processing, 75% in their specific gravity, particle size â&#x20AC;˘ compact des ign and portabil ity. and/or particle shape.

Some local authorities may see these Figure l . Biosolids water distri bution. characteristics as offering advantages The system removes water from rhe for particu lar situations though material via several pathways, but rhe operating costs and biosolids majoriry of the water chat is removed reduce volu me, weight, odours and reduce characteristics may be li miting fo r some rhe cost of removal of biosolids from the does not change phase, i. e. liquid to applications.

LWPCC. T he key objectives of the project vapour, but remains as a liquid which is are to: atomised. Atomisation of the water creates Background very fine (around 10 - 200 microns in I. Dete rm ine the bioso lids quality char can Logan has for the past 25 years been on e of diameter) spherical liqu id water droplets be produced by rhe system the fastes t growing areas in Queensland, which are separated via particle size rather 2. Identify rhe operation al costs of the with a cu rrent popul ation of 170 000. The than specifi c gravity and carried away with system Loganholme Water Pollution Control the air stream. Thus the major pathway for 3. Assess the characteristics of the dried Centre (LWPCC) was built in the 1980s water removal is mechanical (non-p hase bioso lids which will impact on its ab ili ty with the capability of expansion ro meet th e chan ge) with minor water removal via to be beneficially used sewerage needs of the developi ng city. Each evaporation (phase change via air vapour day there is an average of 40 million li tres extraction ).

Process Description flow of sewage through the plant, generated Shear and impact forces act to bring the Logan Water has been investigating by residential and commercial consumers. water ro the surface of the particles. dewacering best practice and have Logan Water is the branch responsible for T hrough agitation or splitting of the conside red a number of options. In sewage treatment for Logan City Council particle, more surface area is exposed and October 2002 a new invention called (LCC). Logan Water biosolids are currently more water is liberated from the surface.

"Cyclonic T hermal Drying" was patented produced by dewarering BNR sewage in the USA by the Globa l Resource The centrifugal force physica lly separates sludge by belr fi lter ro a concentration of the water fro m the solids due to about 15% solids and loading some differences in their specific gravities.

70 tonnes per day of cake into a The heavier mass will rotate at a sludge hopper where it waits to be di fferent speed and cause sheer transported off site for fur ther separatio n. The liberated water will processing. continue to rotate and fractu re into A successful application fo r an smaller and smaller particulates and Advanced Wastewater Treatment become atomised. A minor amount of Technologies grant from che water is removed via air - water vapour Queensland State Govern ment has interactions and the implications of the enabled LCC to assess in detail a psychometric chart and evaporative proposal to supply and operate pilot cooling. scale equipment ro dry biosolids ro

r

J

Trials ofa newly developed system.

Figure 2. Current Belt Filter, producing 70 tonnes per day of 15% cake.

As the solid material continues through the process the surface area to volume ratio increases and the materi al is agitated to expose the water. Air velocity, material temperature, material

Journal of the Australian Water Association

Water

MARCH 2007 61

technical features

sludge drying surface area/volume ratio, vapour pressures, system pressures, water conscicuency and ambient conditions are parameters chat must be considered in determining che extent of water removed using the drier.

E55 Tempest Dryer

Separation Cyclone

Duct

The approximate percentages by weight of rhe various moisture constituents in biosolids are shown in Figure 1.

The moisture associated with rhe material is in part free and separable by gravity; in part trapped in the interstices of floe particles and separable by mechanical dewacering; in part held by capillary action, and separable by compaction; in part chemically bound within the bacterial cell, and separable only by destruction of che cell; and in pare chemically bound within a synthetic matrix (polymers), and separable only by destruction of the matrix. T he water char is chemically bound in a synthetic matrix is extremely difficult co remove wich low heat processes and may prove co be the upper limit of water removal via this drying process.

We t Scrubber" "

Electric Motor

""-

Air Compressor on Motor unit

Control Panel

Figure 3. Cyclonic Thermal Dryer general layout (diagram courtesy of Global . Resource Recovery O rganization).

Pilot Trial Logan Water began resting a rrailermounced "T empest" pilot plant in September 2006 using undigested, unscabilised BNR sludge cake which has been dewacered co 15% solids cake by belt filter, currencly produced ac che rate of 3 r/hr throughout a 24 hour period.

The pilot drying unit was the smallest model available and the feed-ra re chosen for che trial was initially 1 r/hr of cake (Table 1). Higher feed races up to 4 c/h r were tested co determine the effect on product quality.

T he belc-dried biosolid cake (15% solids) is fed by an auger from che hopper where ic is introduced inco che high velocity air-stream educcor chen circulated through a precyclone before entering the main drying cyclone for approximately two minutes.

Dried biosolid granules are extracted from the base of the cyclone through an output auger. The air stream passes through a wet scrubber where the water droplets coalesce along with any fin e dust and are drained off

Figure 4. Trai ler-mounted Cyclonic Thermal Dryer.

62

MARCH 2007

Water

and returned to the treatment plant. There is no noticeable odour because of the very large dilution and dissipation by the input air stream and the passage through the wet scrubber.

Figure 2 shows the existing belt fi lter.

Figu re 3 is a diagram of the process and Figure 4 is a photograph of the containerised pilot plant.

Figure 5 compares the consistency of the dried product with the belt filter cake.

Results Pilot trials are still proceeding but preliminary results are summarised in Table 1 cogether with estimates of the operational coses.

Improvement on Current Processes â&#x20AC;˘ The drying process successfully removes large quantities of water from biosolids without expending large quanciries of energy.

Table 1. Prelimi nary results of Pi lot Trials.

Dried Biosolids from dryer- solids

15% 85%

Feed Biosolids cake from belt press¡ solids Dried Biosolids Pathogens after Output Auger

900 cfu/g dry wt

Dried Biasolids Pathogens after 2 weeks storage

< 100 cfu/g dry wt

Temperature Reached

1os c

Process Detention Time

<2 minutes

0

Dimensions (without trailer)

12.2m x 2.4m x 2.6m

Air velocity

440 kph

Operation Cost (labour, maintenance and fuel)

$30.60/wet tonne feed Biosolids cake

Figure

5. Feed biosolids (left) and the dried product (right).

Journal of the Australian Water Association

technical features

sludge drying • A large volume and weighc reduccion occurs.

• T he process produces a dry, friable, scable end produce wich no pachogen regrowch.

• T he process is fasc.

• T he Piloc Plane equipment fies inside a scandard shipping concainer which is easily portable.

• Secup rime is fasc as ic cakes I person less than 30 minutes co produce dried material from a cold scare.

• An addicional significanc improvemenc is che enclosed processing allowing a clean acmosphere for workers instead of the odorous environmencs in which so me prese ncly have co work.

Financial Benefit Escimaced ongoing financial benefics resulcing from che drying syscem are: • Reduced power consumption because of high efficiency drying compared co chermal dryers.

• Reduced scorage areas because of volume reduccion.

• Reduced removal cosc from site because of weighc reduction. Less truck movemencs.

Because che dried material can be scored longer, removal can occu r ac more conven 1enc mnes.

• Ease of handling and cransporc co mpared co biosoli ds cakes. No wacer is able co drip from cransporc vehicles and if there was a spill ic is easier co clean up dried gran ules than sloppy, muddy cakes.

• Reduced maincenance and repair costs th ro ugh protection of exposed surfaces of plant and eq uipment from moiscure and H2S accack.

Implications The average price for biosolids disposal in soucheast Queensland for councils with similar sicuations co Logan is in the vicinity of $34 per wet conne of belt filter cake.

The pilot unit that we tested is a superseded model and the newer versions are improved.

We have esti mated that installation of a ful l-scale plane co deal with our 70 tonne/day (with repaymenc of principal over IO years) is $33.60/c cake feed. N o dollar value for che dried product is included.)

The dried biosolids means more beneficial uses beco me available for consideration. For example, the Queensland Environmencal Protection Agency has approved in principle che use of che LCC biosolids for agriculcure appl ication. A curf farm business has been developed by Logan Water as an accivity co uti lise recycled water for irrigation, supply turf to the local area and more proficably utilise LCC's land assets.

As turf crops are harvested, there is a req uiremenc co fe rtil ise and replace organic topsoil. Dried bioso lids is an odour free so il additive well suited for this application.

Dried biosolids are likely to be successful as a soil conditioner fo r agricultural and landscaping applications. Sale of product may be possible because of the need for recycled organic products for agriculcure.

So utheast Queensland so ils are moscly carbon depleted and with higher organic concent soils being more moisture retentive, then there could be a profitable demand for a recycled organic produce.

Anocher potencial use which could be generated by an efficienc and eco nomic dryer would be co dry biosolids and other sludges from external sources. O ther loca l authorities and businesses face the same costs as Logan Water does fo r sludge

Community Benefit • Large sections of the co mmunity in cicies and cowns are currently subjected co odours emanating from sewage treatmenc planes, evidenced by the consumer complainrs received. A very dry biosolid is vircually odour free.

• A granulated product may be more conveniently used for transport and application.

• Product with reduced weight and volume will mean fewer truck movements.

By eliminating a primary so urce of odour release and thereafcer being ab le co identi fy and treat any secondary sources, odour-free treatment plants will evencually be achieved.

Water Advertising To reach the decision-makers in the water field, you should consider advertising in Water Journal, the official journal of Australian Water Association.

For information on advertising rates, please contact Brian Rault at Hallmark Editions, Tel !03) 8534 5000 or email brault.rault@halledit.com.au

dewatering and biosolids removal so there wo uld be opporcunities co reduce dryi ng and cransporc costs by using a porcable dryer. Cyclonic Thermal Drying has the advancage of being compact and fasc. Local participating producers would benefic by reducing their disposal costs without capital outlay.

Wider Applications T his AWTT / LCC projecc was to supply and install piloc scale equipmenc co dry biosolids cakes co reduce vo lume, weighc and odours. So far che manufaccurers in che USA (GRRO In c) have successfully applied che machines co: • Paper sludge being dried co reduce carcage and reduced land fill coses • Paper sludge being dried then pressed inco pellecs for furn ace fuel.

• Bioso lids and paper sludge being dried then used in composcing operations • Biosolids drying • Food drying and chem ica l drying • Piggery Manure

Conclusions T he primary aim of the piloc project was co demonscrate che feasibilicy of an onsice, compact biosolids drying plane which can produce a reduced volume and weighc, low pachogen, granular, free flowing product from the poinc of tradicional dewacered cake discharge. This approach wou ld eliminate the need for excensive on-sire scorage prior co scabilisation creacments and thereby eliminace issues associaced wich generacion of odours from stockpiled biosolids. It is also expecced char the drying and granularisation approach wo uld dramatically reduce the difficulty of handling biosolid cakes ch rough being able co use conventional so il handling equipmenc and cechniques.

Logan City Co uncil believes char che projecc is innovacive in ics approach co biosolids drying which will directly benefic State and Local Govemmenc in Queensland.

A furcher paper dealing wirh invescigacions on marketing and applicacions for che product is in preparaci on.

The Author

Chris Lane is che Execucive Technologies Administracor wich Logan Cicy Council and was Project Direccor for che crial.

Email chrislane@logan. qld.gov.au References Global Resource Recovery Organisacion Inc. www.grro.net

Journal of the Australian Water Association

Water

MARCH 2007 63

technical features

sludge drying

THERMAL DRYING: OPERATIONS AND USER ACCEPTANCE A Campbell, A Wilson, P Atkinson Abstract This paper discusses the issues associated with the use of thermal d rying to stabil ise and dewater biosolids from municipal wastewater treatment plan ts, with reference to two plants in New Zealand, New P lymouth and Seaview, both of which have been operating for over five years. T he New Ply mouth Plant has registered its thermally dried produce and now sells it as an organic fertiliser.

Keywords Thermal drying, biosolids, experience, reuse

Introduction Thermal drying is now an established technique for stabilising and dewatering biosolids from munici pal wastewater treatment plants. Ir has a number of obvious advantages: • Provides a high degree of dewarering thus mi nimising the volume of product and transport costs .

• Provides a high degree of pathogen reduction.

• Produces a stable, useable produce with negligible odour or vector attraction.

However, thermal d rying is perceived by some as being capital and maintenance intensive with h igh energy consum ption.

Background New Plymouth is a city of 70,0 00 people located on the west coast of central North Island , New Zealand. The New P lymouth WWTP is an activated sludge plant comprising 3mm gap inlet screens, gri t removal, aeration and chlori nation. Waste sludge is dewate red using belt presses and then thermally dried. The original plant was comm iss ioned in 1984 and th e thermal d rying in 1999.

The Seaview WWTP serves the cities of Lower Hutt and Upper H urt near Wellington, New Zealand. The Seaview plant comprises 1mm gap screens, primary sedimentation, contact stabilisation activated sludge and UV d isinfection.

Successful experience at two rotary drum driers in New Zealand.

64

MARCH 2007

Water

Seaview WWTP thermal drying facility.

Primary and waste activated sludge is dewarered using centrifuges and then thermally dried. The screen ing faci lity was comm issioned in 1984. The ocher fac ilities were commissioned in 200 1.

The thermal d ryers for both plants were manufactured in New Zealand by Flo- D ry Engineering, usi ng technology originally d eveloped for p rotein sterilisation in the meat and rendering industries.

Process Selection T he two plants arrived at the decision to use therm al drying through quite different routes, bur ulrimarely fo r similar reasons.

The New Plymouth plant originally dewarered the sludge using gravity chickening and a centri fuge with rhe dewarered sludge landfilled. Difficulties with the characteristics of the centrifuged material and th e p hysical instability in the landfill resulted in the centrifuge being replaced with gravity d rainage decks and belt fil ter presses. Disposal was by careful land application and shallow soil incorporation on controlled sites. The land in the New Plymouth region is h igh value agricultural (dairy) land. The lack of suitable sites for incorporation meant char a more sustainable system was requ ired. Key considerations fo r New Plymouth were:

Journal of the Australian Water Association

• Decouple surplus sludge wasting from sludge disposal • Process sludge for the further reduction of pathogens • Improve biological stability of the waste sludge • Further dewarer sludge for p hysical stability, reduced rransporration costs and ease of storage • H ave a high confid ence of odour and vecto r control • A strong preference for beneficial reuse (and sale) if possible The Seaview WWTP was delivered as a design , build and operate facility. The contract has a 20 year operating period and faci lities were evaluated within a commercial framewo rk. Thermal drying was selected by the DBO contractor as the preferred means of biosolids treatment and disposal on the basis of: • W ho le of life costs • Land application areas some 50 - l 00 km from the plant • Additional charges imposed by the landfill on (non-dried ) sludge cake as it does n ot ach ieve USEPA C lass A • Accuracy and reliability of treatmen t and disposal cost estimates • Commercial risk assessment.

technical features

sludge drying Process Description Thermal drying is rhe use of hear to evaporate water from rhe residual wasrewarer solids. Hearing can be either direct or indi rect.

Direct dryers have hor air (o r gas) fl owing through a process vessel and co ming in di rect contact wirh rhe wer solids. The pred o minant method of hear transfer is convection. Increasing rhe temperature of the wee solids evaporates the water.

Indirect dryers have a solid metal wall separating rhe wet solids from the heating medium (steam, hot warer or oil). The predominant method of heat transfer is conduction. T he hear is transferred to rhe wer solids via rhe metal wall. The wet so lids do not co me in co ntact with the heati ng medium.

With in each category there are a number of arrangements (as described in WEF, 2004) including: • Direct type, rotary drum dryers • Di rect type, flash dryers • Di rect type, belt dryers • Indirect rype, tray drye rs • Indirect type, padd le, disc or auger dryers • Indirect rype, fluidised bed dryers

Seaview dryer barrel (drive sprocket in foreground) . has an internal spiral which moves rhe product progressively to the discharge end.

The air flow through rhe dryer causes the finer, lighter dried material to be swept out of the drier more quickly than rhe wet material which needs more drying rime.

4. Hood Cyclone and Exit Screw - The dried material (biosolids) are dropped our of the hood and into the exit screw. Th e exhaust gases are passed to th e cyclone where dust and fine material entrained in the air drop out. All rhe biosolids from rhe

• Dewatering dryers • El ectric dryers New Plymouth and Seaview are both direct rotary drum dryers. T hese dryers have four ma in process areas: • Material silos, mixing and product preparation • Drying • Dried biosolids hand ling

-Ideal for long term water level mon,tortng. or fast sampling for pump/slug tests -Ultra rugged, all stainless steel or 111an1um design -Huge memory capacity -3 ves1ons availabl~ to suit every application

• Ancillary equipm ent Th e dryer itself has five main sections: 1. The Combustion Chamber - A gas

fired burner hears the ai r which flows through th e dryer. The 'air' is a mixture of recycled combustion gases and fresh makeup air.

2. lnfeed Screw and Cone Section - The pre-m ixed material is transported via a water-coo led infeed screw into the inlet section of rhe dryer. The inlet section is a cone, which aces as a transition between the co mbustion chamber and the dryer barrel.

3. Dryer Barrel - As the dryer barrel rotates, the wet material is continually li fted and cascaded th rough rhe hoc air scream. A large proportion of rhe air is recycled thus reducing the volume of exhaust gases and maintaining low oxygen levels in the dryer barrel to prevent combustion of organic material. The dryer

-Conductivity, Level, Temperature, Data logging -Easy installation & low maintenance -Ultra rugged and corrosion ·resistant titanium housing

-Measures Temp, pH, ORP. DO.EC, Depth, Barometric Pressure.Turbidity, Nitrate, Ammonia & Chloride -Up to 9 sensors in a 4.7cm OD (Clarke membrane or optical sensor)

T: (03) %46 4190

E: info@enviroequip.com

MELBOURNE SYDNEY/ BRISBANE PERTH AUCKLAND

Journal of the Australian Water Association

Water

MARCH 2007 65

technical features

sludge drying hood and cyclone are collected in the exit screw.

5. Condenser and Condenser Fan - The condenser fan blows air through the dryer system. The air is passed through rhe condenser where the evaporated moisture (condensate) is removed and rhe air temperature is reduced from over 110°C to approximately 40°C. 70-75% of the condensed gases are recycled to the combustion chamber co provide the n ecessary transport air for the process. The excess gases are d irected co a biofilrer for odour control.

The temperature in the dryer is low enough co p revent oxidation (burning) of the organ ic matter. T hus most of the organic matter is preserved in the dried material.

T he biosolids exiting the dryer range in size from dust co over 6mm in d iameter. The b iosolids are then classified th rough vibrating screens into "overs", "accepts" and "unders". T he "overs" and "unders" are passed through a crusher to reduce them ro 1mm minimu m in diameter and recycled for mixing with wet raw material. If insufficient crushed material is availab le, "accep ts" are used co make up the recycle volume.

The recycle of d ried material is a key feature. The recycle scream is mixed w ith the incoming wee slud ge co give a feed to the drier with greater than 70% DS. Th is avoids the "sticky phase" which can be encountered with sludge at around 60 65% DS. Recycle and mixing results in the dry granu les being coated with wet sludge so char only surface drying is needed.

Odour Control Considerable attentio n has been paid co odour control at both sires . The New Plymouth facil ity is located next co a prestigious golf course and the Seaview plant is surrounded by industrial properties.

T he two main odour sources co be created are the dryer exhaust gases and fugitive process odours .

D ryer exhaust gases are passed into the manifold from the d ryer system via an air break. To contain any fugitive gases, the dryer building at both plants is designed co operate under negative pressure, being fully enclosed. In addition there are extractions from poi nt sources with high likelihood of odour such as the raw material silo and rhe mixi ng area. Dryer gas and air from point sources is treated in b iofilrers .

Safety A co ncern often voiced regarding thermal dryers is the risk of fires or explosions in

66

MARCH 2007

Water

Table l Parameter

Population equivalent Design ADWF Feed rate jmax) Dewatered sludge feed jtyp) Product dryness jtyp) Electricity use per tonne DS Gas use per tonne DS Gas use per tonne water evaporated

New Plymouth

Seaview

80,000 19,000 m3/doy 3.0 tonne jwet)/hr 14% DS 95% DS* 410 kWhr 24 GJ 3.8 GJ

180,000 50,000 m3/day 5.0 tonne jwet)/hr 22% DS 92% DS 210 kWhr 13 GJ 3.8 GJ

* more recently 92% DS the drier or the storage silo. Neither has occurred at the New Plymouth or the Seaview plants after more than 10 years of combined operation.

Safety protocols were developed to prevent explosions during the drying process. This included formal HAZOP studies char co nsidered both hazardous gases and dust.

The oxygen level in air flowing through rhe dryer is maintained below combustible levels whilst there is product with in the drier. T his control is achieved by balancing oxygen depleted combustion gases and recycle air. On start up, the drier and ancillary equipment is heated and the oxygen levels in the dryer depleted before material is fed in co the drier. Before the plant is fu lly shut down, rhe equipment keeps running for a p rescribed rim e co purge the equipment of product and co allow the equipment to cool. In add ition , there are a number of ocher siruarions where the p lant will automatically stop feedi ng material or shut down, primarily as a safety precaution. These include: • Cooling water fai lure • No wee or d ry material being fed • Blockages • High temperatures • Posi rive pressure in the d ryer • High oxygen level d u ring operation All electrical and instrumentation equipment is classified co meet rhe relevant h azardous area zone requirements. In addition, the fac ilities have n itrogen fire sup pressio n systems. Nitrogen is supplied via an exchangeable bottle pack feedi ng into a receiver. The system requ ires chat any o ne of the dried material storage silos or the d ryer can be flooded with nitrogen within seconds to suppress any fires. The nit rogen suppression system is activated automatically, but only if a series of ocher safety features have fa iled to successfully remedy a potentially hazardous situation.

Journal of the Australian Water Association

Process Performance For both plants, rhe thermal d rying equipment was procured by means of a performance contract. The contracts included design, supply, insrallario n, commissioning, performance proving and operator training. Boch contracts rhe vendor guaranteed: • Throughput • Gas usage • Electriciry usage • Product solids content and granular size range.

At the rime, in the absence of a recognised NZ standard defining biosolids quality, th e USEPA Parr 503 Biosolids R ule was used .

The thermal driers were required co achieve C lass A biosolids. Extensive testing was undertaken to confirm compliance and the driers at both plants achieved the vendor performance guarantees.

Operations Experience Operating experience at both plants has either met or exceeded expectations, bur both p lants have experien ced higher rhan anticipated main tenance needs. T he maintenance particularly relates to the abrasive namre of the dried produce and consequent wear on the materials handling equipment. Sludge degricring has been retrofitted at Seaview to reduce wear. Use of wear areas, liners and harden ing of faces is recommended to m inimise wear. Bisalloy and UHMPE liners have proven most effective, plus maximising the size of screw co nveyors and operating these at m inimum speed.

The preferred operating regime is 24 hours per day. Preventive maintenance is u ndertaken on non-operational days each week. Continuous operation 24 h ours per day rather than for shorter periods has the fo llowing advantages : • Reduced elecrrici ry and gas usage as each start up requires time to deplete rhe oxygen levels and bring the plant up to

technical features

sludge drying temperature, and each shut down time to ensure all product is removed from the process train.

• Reduced thermal cycling of the equi pment and improved thermal efficiency.

• Increased opportunity for preventive maintenance T he New Plymouth facility operates with pare-time manni ng. An operator is in attendance at start up to conduct check and housekeeping duties. Shu tdown is automatic and unmanned.

The Seaview dryer is manned co nti nuously when in operation. The plant operates 120 hours per week to process the sludge inventory. Full ti me ma nni ng allows operators to take corrective action and avo id "unnecessa ry" automatic sh utdowns.

Due to its 12 to 14 day sludge age, the waste activated sludge feed to the New Plymouth facility varies li ttle on a day-today basis. The Seaview facility processes a mixture of primary and secondary sl udge and the ratio of rhese ca n vary on a day-today basis. ln addition, the natu re of the seco ndary sludge from rhe contact stabilisation process also varies as ir has a sludge age of only 3 to 5 days. Operation steps taken to ad dress this issue are storage of the sludge and blending of primary and secondary sludges. Even with va riable and shore sludge ages the plant is able to maintain a relatively constant feed to the drier of 20 to 24%DS . The drier feed is mon itored every fo ur hours and operation adj usted as requ ired to maintain prod uct qual ity.

End User Acceptance Although the statutory framework in NZ differs fu ndamentally fro m Australia, the Guidelines for the Safe Application of Biosolids to Land in New Zealand, 2003 were developed along philosophically sim ilar lines to the Victoria and New South Wales guidelines. T he guidelines were prepared with input from a range of governmental and industry stakeholders.

Biosolids quality is defi ned in terms of a stab il isation/treatment grade and a chemical/contaminant grade.

The adopted limits fo r metal concentrations in soils and biosolids are based on the (precautionary) European approach - the so called "LOAEC approach" which involves setting soi l limits at the lowest observed adverse effects concentrations. Hence the concentrations are similar to the Australian guideline numbers and so are considerab ly lower than in the USEPA Rule 503.

Biosolids reuse typically focuses on application to agricultural land. The New Zealand agricultural industry relies heavily on export markets and so has been cautious in its approach to use of both biosolids and treated wastewater.

In the absence of any national biosolids guidelines in 1999, New Plymouth successfu lly sought registration of the thermally dried product as an organic fe rtiliser. The product is trademark protected as Taranaki Bioboost 6-3-0TM the nu mbers representi ng the N-P-K percentages (to the nea rest whole percentage point) as required by the Fertiliser Act (Wilson 2003). The plant has a formal quality management system to maintain product qual ity and consistency.

New Plymouth made the decision to restrict marketing of Bioboost to non- food linked and non-pastoral activities and has entered into a long term relationship with a sales agent. The agent is respo nsible for marketing and distribution. T he contract is for fi ve years with three rights of renewal each fo r a further fi ve yea rs.

The outcome has been a very successful co mmercial partnersh ip. The marketing has been particula rly successful in the niche areas of sports turf an d ornamental horticulture. Ta ranaki Bioboost 6-3-0™ is now the fertilise r of choice for golf cl ubs throughout the North Island of NZ. The production cost ofB ioboost (1250 to nnes per year) is rough ly th e same as for landfi ll ing dewatered sludge (10,000 tonnes per year) but with rhe added advantage of a sustainable distributio n of product for beneficial re-use. The public can purchase rhe prod uct in 25 kg bags for $NZ14.95. It is sold also in 500 kg bags for $NZ2 I 5.00 and as bulk product for $NZ61.90 per tonne. T he distributor is presently organising the sale of the product in 8kg bags for the home gardener. T he plan is to eventually sell all product in bag form which will give a better return. T here is a profit shari ng clause in the contract to provide a fair return to all parties.

New Plymouth has recently gained acceptance in principle from the dairy industry for agricultural reuse - provided rhat the biosolids can not be ingested by the cartle. However demand from the turf ind ustry exceeds supply and agricultural reuse has not commenced.

Seaview has been disposi ng if its biosolids to land fi ll. Discussions regarding reuse have commenced with fo restry and fuels bei ng rhe most likely routes.

Ongoing Development Flo-Dry has reported on development of a dual stage dryer, with the objectives of reducing energy usage and avoiding the recycle step. It comprises a rotary drum dryer (as at New Plymouth and Seaview} bur fo llowed by a belt dryer (Browne et al, 2006). The first stage dries to 40 - 45% OS and the second to over 90% OS. A working demonstration plant commenced operation in J une 2006 at the Mangere WWTP, Auckland. The gas usage was measured at between 16 and 22% lower than a conventional drier. The avoidance of the recycle step ca n be expected to reduce wear and maintenance.

Conclusions 1. Thermal d ryi ng can be a cosr effective treatment method.

2. The thermal drying plant itself can be successfully operated withi n a wastewater rrearment fac ility.

3. T he thermally dried product can be marketed to the point where demand exceeds supply.

The Authors

Allan Campbell is T echnical Manager, Wastewater, Beca Pty Ltd (emai l allan.campbell@beca.com), Anthony Wilson is General Manager, Co mmunity Assets, New Plymo uth District Council and Paddy Atkinson is General Manager, Hutt Valley Water Services.

Acknowledgments Seaview WWTP: Owner: Hutt City Cou ncil Design/Construction/Operation: CH2M Beca, Bovis Lend Lease, H utt Valley Warer Services (Austral ian Water Services, OMI Beca}

References Browne, OW, Kurvink, MR and Fernando, T (2006). Two is Better than One: A More Efficient Drying System. European Biosolids and Organ ic Residuals Conference, Wakefield, Yorkshire (United Kingdom) Novembe r 2006 WEF (2004) Thermal Drying ofWascewacer Solids, White Paper. Water Environment

Federation Residuals and Biosolids Committee Bioenergy Technology Subcommittee Wilson, AE (2003) . T hermal Drying of Biosolids - the New Zealand Experience.

Proceedings Sch Eu ropean Biosolids and O rganic Residuals Conference, Wakefield, Yorkshire (United Kingdom) November 2003

Journal of the Australia n W ater Association

Water

MARCH 2007 67

fereed paper

CLIMATE CHANGE IMPACT ON ROUS WATER SUPPLY D Kirono, G Podger, W Franklin, R Siebert Abstract Climate change poses significant risks to the security of water resources in many parts of Australia. This study investigated the implications for secure yields of a regional water supplier in NSW. The aim of the study was to estimate future secure yields, using the Integrated Quantity and Qual ity Model (IQQM) and a range of climate change scenarios, to identify the most appropriate time horizon for making new investments in infrast ructure.

The assessment co ncluded chat, when taking into consideration impacts of glo bal warming, the need fo r a new so urce will be most likely after 2018, consequently plan ning needs to commence in 2008.

B\'11011

USMOllE

• restrictions of any k ind should not be applied for more than 5% of the time(> 1862 days); • restrictions of any k ind should not be imposed more than one year in ten on average (> 10 years); and • the system should be ab le to supply 80% of normal demand (i.e. 20% reduction in consumption) through a repeat o f the worst drought on record.

Introduction There is an increasing body of research that supports a picture of a warming world with significant changes in regional Figure 1. Rous Wa ter scheme (Rous Water, 2006). cli mate systems (IPCC, 200 1; IPCC, 20 07). This has significant implications for the (NSW) . The third sect ion describes the reliab ili ty of water supplies across Australia framewo rk built to assess climate change Qones and Preston, 2006). Ir is important impacts on Rous Water's su pply. Section to consider the risk that climate change four presents and discusses the assessment poses to catchment yield withi n the context procedure and subsequent results, and of changing demand. section five summarises the main This p roject investigated the implications conclusions of the study. that climate change may have on Rous Water's regional water supplies. T he aim of the study was to help identify the most appropriate time horizon for making new investments in infrastructure.

The first section provides an overview of Rous Water's regional water supply scheme, the second d iscusses the projected climate change profi le for New South Wales

The need for a new source is most likely in 2018: planning should start in 2008.

68

MARC H 2007

Water

Under this scheme the secure yield is defined as the annual demand that can be supplied from the headworks over the 103 year h istoric record and wh ich satisfies the 5/ l 0/20 rule, i.e.:

Overview of Rous Water Regional Water Supply Rous County is located in northeastern NSW and is part of the Wilsons River catch men t, from Lofts Pinnacle in the west, along Nightcap and Koonyum ranges near the coast (Figure 1). Rous Water is the regional water supply authority providing water in bulk to a number of Councils, from Lismore to Ballina, with an app roximate population of93,000. T he supply system was designed based on the forme r NSW Department of Public Works and Services definition of'secure yield'.

Journal of the Australian Water Association

Water p resently comes from cwo main supply storages: Rocky Creek Dam and Emigran t Creek Dam. The former has a storage capacity of 13,956 ML and a safe yield of about 9,600 ML/annum (DIPNR, 2004) while the latter has a capacity of 820 ML with a safe yield of about 1,100 ML/an num . There are also some ocher small sources under Counci l control with a combined safe yield of about 900 ML/annum.

Based on population p rojections and other considerations, GeoLINK (2005) est imated chat the best estimate of the likely fu ture demand from the Rous Water scheme in the year 2030 is around 18,000 ML/annum. This equates to an increase of 43% over the current demand of 12,600 ML/annum. To manage th is growth in demand, Rous Water adopted a water managemen t strategy in 1995 chat was amended in 2004. The strategy provides a range of options to meet water requirements. One key objective of Rous Water's strategy is to implement effective demand management, the target being a minimum 10% reduction in per capita demand by the year 2011, relative to 2005.

Recognising future demand may outstrip existing sustainable yield, the strategy

technical features

climate change identified two additional supplies, Lismore source and Dunoon Dam. Lismore source is a medium-term solution which is able co assist in meeting the high demand projection up to 2024. le consists of a pu mping station capable of abstracting up co 30 ML/day fro m the upper reaches of the tidal pool in che Wilsons River, which is on ly utilised when Rocky Creek Dam is below 95% capacity and is subject co abstraction licence constraints. The proposed Dunoon Dam is located downstream of Rocky Creek Dam and captures local inflows as well as Rocky C reek dam spills (CMPS&F, 1995).

refereed paper

6 Global Climate Modelo

Demand Projections

(GCMa) Slmulatlon1

c$

l

Three Climate Scenarios for each year

20, 2030)

Assessment

•Rainfall •Evaporation

f=:J t...:::j

I

•Rolriall . .

•Evaporati:.J

7 fI~ .11

Ory

Secure Yield VS Demand

·Evaporation

l

High Demand

Secure Yields •Baseline •Under cilmate change

An Overview of Climate Change in

NSW During the 20th century, the globally averaged surface temperature increased by 0.6 ± 0.2°C with the warmest year being 1998, followed by 2005 (WM O, 2005) . In NSW, che temperature has also been steadily increasing over che last fifty years (BOM, 2006). In northeastern NSW, cemperacures have increased ac the race of approximately 0.4°C per decade. The BOM (2006) has also shown chat rainfall has bee n declining in the non summer months at a race of roughly -20 to -50 mm per decade.

In su mmer, the rainfall has been increasing at a race of approximately+ IO co +20 mm per decade. Lismore rainfall records show chat the annual rainfall has been decreas ing sligh tly at a race of -6 mm per decade from the late 1880s to present. Winter rainfall has been decreas ing at a rate of -5 mm per decad e, whereas the summer rainfall has been increasing at a rate of+ I mm per decade. However it should be noted chat che trends in Lismore's rainfall are not scaciscically significant at a 95% co nfidence level.

The climare system is highly complex, and therefore it is inappropriate co simply excrapolace past trends co predict future conditio ns. T o estimate fu ture climate

Dally flow and Supply

Sacramonto model

•Ba,ehne

•Undef clomatt change

2005

Time

2030

Figure 2. Framework for climate change impact assessment for Rous Water Scheme.

change, scien tists have developed climate scenarios from global climate models (GCMs). Best estimates for globally average surface air warm ing are expected to range between l .8°C to 4.0°C at 2090-2099 relative to 1980-1999 (IPCC, 2007). In Australia, CSIRO uses both global and regional climate models in the development of regional climate change projections.

According co Hennessy et al (2004), the models tend to simulate decreasing annualaverage rain fall over NSW, particularly in winter and spring. In autumn the di rection of the change is uncertain, while in sum mer there is a tendency for increases in the north-ease. Annual-average potential evaporation is projected co increase across NSW. The largest changes are proj ected in winter with the smallest changes in summer. Compared to changes in the other areas of NSW, the projected changes in the north-ease, where Rous Water is located, are relatively small.

Framework for Impact Assessment Impact and risk assessment is one stage in a larger risk management fra mework. Ideally, risk management invo lves all related stakeholders. The decision-making process is commonly ci rcular co allow the performance of chosen decisions to be reviewed and revisited as new information on climate change and its impacts are available.

Mose research on the hydrologic impact of cli mate change uses a predictive approach. le begins with generating cli mate change scenarios. Cli mate information is then fed into hydrologic models and/or watermanagement systems to evaluate the differences in system performan ce under different climate scenarios. Adaptations can chen be designed to manage chose changes.

Through consul tations with stakeholders, the assessment fra mework for cl imate change im pacts and adaptation fo r Rous

Journal of the Australian Water Association

Water

MARCH 2007 69

Water has been established and is presented in Figure 2. D etails about the main steps in chis framework are described as fo llows.

First step: preparing climate change scenarios As described previously, future climate scenarios are commonly developed through GCMs. C urrendy, there is a range of available GCMs, each developed by a different scientific group across the world.

These models differ in their approaches co simulating the climate, hen ce different models may project different climate fu tures, even when driven by the same scenario of future emissions. The standard measure co co mpare climate models is their 'sensitivity' defined by how much eventual warming they proj ect when the-p reindustrial atmospheric con centration of CO 2 is doubled from around 270 co 550 parts per million by vol ume (p.p.m) (today's atmospheric CO 2 is about 380 p.p.m). GCMs use a particular emissions scenario as the input to generate a projection of climate change over the next centu ry. The IPCC commissioned a range of scenarios of greenhouse gas and sulfate aerosols emissions up co the year 2 I 00. The scenarios were reported in the Special Report on Emissions Scenarios (SRES, 2000). For example, the SRES-A I scenario depicts a very rapid eco no mic growth, global population that peaks in midcentury and declines thereafter, and the rapid introduction of n ew and more effi cient technologies (IPCC, 200 I).

Fo r this study, scen arios of regional change as a function of global warming (percent change per "C of global warming) for potential evaporation (Ep) and precipiracion (P) were prepared based on six GCM simulations through the use of the CSI RO Climate Scen ario generator, OzClim. OzClim is a PC-based cl imate scenario generator char si m plifi es the process of calculating scenarios from climate change model outputs, applies scenarios co impact models and manages uncertainty (Page and Jo nes, 200 I ). A range of GCM, emission scenarios and climate sensitivities can be harnessed using chis system. For chis assessment, multiple climate change simulations were conducted for the years 2010, 2020, and 2030. These were based on six climate models assuming a range of climate sensi tivities (low, medium and high) and the SRES A I B, A IF and B 1 scenarios. The Al scenarios can be considered as a pessimistic as they give high CO 2 em issions. AlB depicts a balance across all sources, while Alf dep icts a foss il intensive situation. The B 1

70

MARCH 2007

Water

scenario may be considered as a more optimistic future. le describes a convergent world with the global population that peaks in mid-century and declines thereafter, rapid change in economic structures coward a service and information economy, with reductions in material intens ity and the introduction of clean and resource-efficien t technologies (IPCC, 2001). This step p roduced a total of 18 simulations for each year and for every station used in th e Wilsons River IQQM.

Second step: estimating supply under climate change The climate scenarios were applied co the Integrated Quantity Quality (IQQM) hydrologic, river system simulation package (S imons et al. 1996) co generate climate change flow sequences. IQQM consists of the Sacramento rainfall-runoff model and river routing, water demand and allocation routines co simulate river flo w and river regulation. This software has been implemented in most regulated and a large number of unregulated river systems in NSW and Queensland . This study used the Wilsons River IQQM implementation which is described in detai l in DIPNR (2004) .

The n umber of observed rainfall, evaporation, and stream flow stations used in the Wilso ns River model are 8, 1, and 14, respectively. Rainfall data was used to account for soil moisture (which governs the crop water demands of irrigacors) and co apply rainfall co the water surfaces of reservoirs and river reaches. Rain fa ll data was also required for calculating catchment inflows. The evaporation data were used co estimate evapocranspiracion from crops, evaporation from reservoirs and river reaches, and co synthesise srreamflow. The model incorporates existing irrigation development (3,000 ha); system demands (11 urban/rural deman d centres); sources of supply (Rocky Creek Dam, Emigrant C reek Dam, and Borefields); scheme operating rules and access constraints co screamflows at Lismore source.

The published results for the Wilsons River model (DIPNR, 2004) are based on a model scenario known as FX04D. H owever, subsequent co chis report, the Lismore source access rules have been investigated further, resulting in a revised scenario k nown as FX94D, which is as yet unpublished. Based on a d iscussion with Rous Water, chis study has adopted the FX94 D model scenario as the baseline from which co assess the impacts of climate changes . To assess the 20% rule a variation of the FX94D model was created for chis

Journal of the Australian Water Association

study. In chis variation the storages are configured co 55% capacity, which represents the capacity at which restrictions are imposed on the system. This con figuration is used co verify the 20% rule.

T o be able co assess the impacts on fl ow due co changes in climate the rime series inputs used in IQQM were modified. In the FX94D scenario IQQM uses a mixture of observed and predicted flows chat vary between sites. The Sacramento rainfall runoff model was used co extend flows for most of the tributaries. Em igran t Creek D am was based on che Australian Water Balance Model (AWBM) rainfall-runoff model wh ile Rocky C reek dam was based on a monthly correlation w ith a nearby gauge and then disaggregated by a mixtu re of techniques. For chis study all tributaries needed co be modelled using rainfall-runoff models and co be consistent the Sacramento mod el was used for all tri butaries. DIPNR had already developed Sacramento models for Rocky Creek D am and Emigrant Creek D am that had a good match with the short periods of observed data and these were adopted.

Simulations with the Sacramento model generally matched well with observed flows at all sites. However, when the observed daca were replaced with the simu lated data, the result for secure yield was different to that from the o riginal FX94D model. As the model is most sensitive co inflows from Rocky C reek Dam and Emigrant C reek Dam, the Sacramento models for these tributaries were adjusted with an aim co reproduce rhe original secure yield o f 14,900 ML/annum. Unfortunately rhis was not quite possible due co very small differences (<2%) in critical events for che 5% and I 0 % rules. The best that could be achieved was a secure yield of 15,00 0 M L/annum while keeping the d ifference in overall volume at Emigrant Ck and Ro cky C k w ith in 2%. Consequently, fo r ch is study, all secure yields are compared against a baseline secure yield of I 5,000 ML/annum rather than the 14,900 ML/annum obtained by DIPNR in the FX94D scenario.

Estimation of the secure yield for each climate scenario was estimated based upon the minimum yield required co meet the 5/10/20 rule. The estimation of secure yield was determined by an iterative solution chat successively modifies and runs IQQM until the demand was just met in accordance with the 5/10/20 rule. T his process was carried out for each of the 18 climate scenarios, for a coral of 54 scenarios

technical features

climate change (i.e. 18 scenarios for each of the years 2010, 2020 and 2030). T o solve each scenario, IQQM was run approximately 10 times, resulting in approximately 540 runs ro cover all of the climate scenarios and time peri od s.

20000 18000 16000 -

Third step: risk assessment

"O

::::i 14000 -

T he issue of climate change is beset by unce rtainties. T hese uncertainties include the magnitude of global warming, regional changes in rainfall and evaporation, and regional supply sensitivity and cop ing capacity. Quantifyi ng the uncertainties in climate change and its downstream consequences in units of probability or likelihood helps ro identify robust adaptation strategies Qones and Hennessy, 2000). In chis study an event-based probability, where the likelihood of recurring events is estimated, was used to describe the futu re state of cli mate change under che enhanced greenhouse effect. T o do ch is, Mo nte Carlo methods (repeated rando m sampling) were employed ro stochastically generate probabilistic estimates of fu ture climate change and its impacts on Rous Water's secure yields.

~

12000 10000

Iv

r,,,tt'>'cc

Actual demand

8000 2000

2020

2010

2030

2040

Figure 3. Supply projections due to clim a te c ha ng e (CC) versus demand pro jectio n.

Note: the high demand , best estim ate dema nd, a nd low dema nd curves are ta ken a nd modified from G eo LIN K (2005).

demand and its trend, the historical reco rd from 1996 to 2005 and its trend projection are also presented. The demand projections (either based on the GeoLI NK, 2005, or based on the trend analysis of the historic demand) show an obvious increase. The slope of th e actual demand trend is similar to che besc-escimace demand provided by GeoLINK (2005), suggesting that the two are relatively comparable. Demand is likely co increase ro more than I 4, 150 ML (according to the trend analysis) or more than 14,900 ML (accord ing to GeoLI NK, 2005) by 2030.

For assessment purposes, the 5th, 50th , and 95th percentiles of the probabi lity distribution are then considered as the "wet", "medium " and "dry" scenarios as depicted in Figure 3.

An assessment was then performed by comparing the projected futu re supply, after accounting fo r cl imate change, with projected futu re demand. In th is study the demand projections were taken from the resul ts of the GeoU NK (2005) study that were slightly mod ified so that they have the same starting point with the actual demand trend in 2006 (i.e. approximately 11,973 ML). T he trends in actual demand were also estimated based on the available histori cal demand from 1996 to 2005.

Assessment Results and Recommendations Assessment Simulations with fu cure climate scenarios withi n lQQM resu lted in a range of estimated fu tu re secure yields for the Rous Water scheme in 2010, 2020 and 2030.

These results were subsequently used ro generate cumulative probabi lity distributions for projected yields. T he increases in supply are foun d co be very u nlikely (<1 0% probability), whereas the declines in supply are likely (>66% probability) . T his sugges ts char che probability of changes in water supply is skewed towards the "decrease in fu ture supply" scenarios. The best estimate (50% probability) changes in secure yields in 20 10, 2020, and 2030 are 1.7%, -5.8%, and -8.1 % respectively.

refereed paper

• Jn the wet climate change scenarios, over the near term, the secure yield will fal l in 201 0 before increasing co above che present supply in 2020 and will conti nue to rise co more than 16,000 ML in 2030. The slight fall in 201 0 is hypothesised due ro a relative sim ilar rate of change in both rainfall and evaporation (i.e. 2% above current value by 20 I 0). As the models are more sensitive to the change in evaporati on chis causes a slight decrease in yield in 20 10. l n 2020 and 2030, the increase in rai nfall is much higher than the increase in evaporation , hence the yield increases.

• ln the medium climate change scenarios, the secure yield will decl ine to around 14,000 ML in 2030.

• In che dry climate change scenario, the secure yield will decli ne to approximately 11 ,600 ML in 2030.

T he points of co ncern are where a given supply scenario encounters a particular deman d scenario as summarised in Table I. In the combination of "Wet supply and Low demand", "W et supply an d Mediu m deman d", and "Medium supply and Low demand" scenarios, the supply will keep pace with the demand, therefo re no fu rther action would be required.

• In the combination of "Wet supply and High demand" scenario, however, the need for a new source will be in 2018.

T he proj ections of the fu ture demand (esti mated by GeoLINK, 2005) are plotted in Figure 3. To represent the actual

Table 1. Options on the approp riate ti me to ha ve a new source, accord ing to the differe nt su pply a nd demand scenarios.

Supply scenarios Wet

Medium

Dry

2023

2025 2016 2012

Demand scenarios Low

Medium High

2018

2014

• In the combination of "Medium supply an d High demand", and "Mediu m Supply and Medi um demand" scenarios, there is a need for a new source in 201 4 and 2023, respecti vely.

• In che combination of "Dry supply and High demand", "Dry supply and Medium Demand ", and "Dry supply and Low demand", the need for a new source will be in 2012, 2016, and 2025, respectively.

Journal of the Australian Water Association

Water

MARCH 2007 71

technical features

Recommendations The assessment provides several options o n the most approp riate time to build a new water source, accord ing to che d ifferent su pply and demand scenarios. In terms of the future supply, this study has shown that the p robability of changes in water sup ply is skewed towards the "decrease in future supply" (hen ce the "medium" and "dry") scenarios . In terms of the future demand, GeoLINK (2005) has estimated that the probability of che future demand is skewed towards rhe high and medium demand scenarios. This suggests the followings. â&#x20AC;˘ The earliest rime for a new source fo r the Rous Water system is after 2018. T hus, raking into consideration a 10 year planning and construction time from the time of this report there is only two years before commissioning of a new source has to be initiated. â&#x20AC;˘ The medium rime for a new source fo r rhe Rous Water system is 2023. T hus, from rhe rime of this report there is abo ut seven years time before co mmissioning of a new source has to b e initiated.

Within chat rime, it is recommended that actual demand be closely mo nitored. This is because estimates of when a new source is required are highly sen sitive to demand assump tions. In the case of a dry climate change scenario, for instance, GeoLINK's demand projection leads to a relatively similar suggestion with chat of che actual trend of d emand (i.e. only 2 years di fference). In che case of medium climate change scenario, however, chis d ifferential becomes larger (i.e. 5 years) . In addition, given a high demand and a d ry-supply scenario, a crisis may occur in around 2012, by which point supply will be unable to meet the demand yet it will already be coo lace co comm issio n a new source to address the supply/demand gap. Thus, informatio n regarding che demand is valuable in updating the plan as to when a new source has to be built.

The results of the current study indicates chat, ideally, che demand will nor be more than the driest sup ply projection (i.e. about 13,000 ML in around 201 8). This means that the maximum tolerable increase of demand by 20 18 is o nly approximately 10% of rhe current demand. If the wet supply scenario is taken into account, the maximum tolerable increase of demand by 20 18 is about 25% of che current demand.

Therefore, ongoing monitoring of rainfall patterns is also recommended.

To interpret the results appropriately, there is another factor rhar needs co be taken into account, specifically, cl imate variability

72 MARCH 2007

Water

(e.g. p eriods of anomalous drought or rainfall) not represented by che h istorical record or model simulations that could also affect water supply within the time horizon in question. Changes in decadal mean rain fall may occur due to other facto rs other than the greenhouse effect (e.g. long term variations in natural climate), so there may be additional changes and risks above and beyond chose accounted for here.

Conclusions This paper has described the framework and the results of a climate change risk assessment of Rous Water's suppl ies based upon T he W ilsons river IQQM. Results of the risk analysis suggest that decreases in secu re yields are likely in rhe future. The best estimate (5 0% probability) changes in secure yields in 20 10, 2020, and 2030 are - 1.7%, -5.8%, and -8.1 o/o respectively. The assessment suggests the earliest rime to have a new source is 2018. G iven rhar a ten-year rime lead is considered the minimum required to commission a new water source, the plan has to be starred in at least 2008.

Acknowledgments This work was p roduced by CSIRO under contract to rhe Rous Water Regional Water Supply, NSW D epartment of Commerce.

NSW DIPNR is acknowledged for the agreement regarding the use of Wilso ns River IQQM. C hris Ribbons and Richard Cooke are thanked for providing the required fi les and assistance in run ning them. Cher Page and J im Ricketts helped in setting up the OzClim program for the analysis. The study has benefited from constructive comments by Dr Roger Jones and D r Benjamin Preston.

The Authors Dewi Kirono is a research scientist in Climate Change I mpact and Risk Group, CSIRO Marine and Atmospheric Research.

Email: dewi.kirono@csiro.au. Geoff Podger is rhe Principal Research Scientist, River Basin Modeller within CSIRO Land and Water, and Water Managemen t Research Program Leader within eWacer CRC. Email: Geoff.Podger@csiro. au .

Wayne Franklin is the Operational Services Manager within Rous Water.

Email: wayne.frankli n@ro uswarer. nsw. gov.au. Rob Siebert is the Project Manager of rhe Lismore Source Project, Rous Water.

Email: Roberr.Sieberr@commerce.nsw. gov.au

References BOM. 2006. Australian Bureau of Meteorology. www. bom.gov.rw.

Journal of the Australian Water Association

CMPS&F. 1995 . Rous Regional Water Supply Strategy Planning Study, Scheme Options, Final Report. May 1995 .

DIPNR. 2004. Rous Water -Augmentation of

Regional Water Supply Scheme, Hydrologic modelling study using Wilsons River IQQM. Issue 3. NSW Department oflnfrasrrucrure, Planning and Natural Resource GeoLINK. 2005. Dunoon Dam, Population and demand projections. Draft report, 1908-2005.

H ennessy, K. , Page, C., Mclnnes, K. , Jones, R., Bathols, J ., Coll ins, D. and Jones, D. 2004.

Climate change in New South Wales, Part 1: Past climate variability and projected changes in average climate. Consultancy report for t he New South Wales Greenhouse Office.

IPCC. 2001. Climate Change 2001: The scientific basis. Summary for policymakers. In Houghton, J.T., Ding, Y., G riggs., D.J., Noguer, M., Van Der Linden, P.J. and Xioaosu, D (eds) Contribution of working group I to the thi rd assessment report of the Intergovernmental Panel on Climate Change, Cambridge Un iversity Press, Cambridge. !PCC, 2007 . Climate change 2007: The physical science basis. Summary for policymakers.

Contribution of working group I to the fourth assessment report of the Intergovernmental Panel on Climate C hange.

Jones, R.N. and H ennessy, K.J. 2000. Climate change impacts in the Hunter Valley, A risk assessment of hear stress affecting dairy cattle. A research report, CSIRO Atmospheric Research, Victoria.

Jones, R.N. and Preston, B.L. 2006. C limare change impacts, risk and the benefits of mitigation. A report far the Energy Futures Forum, CSIRO Marine and Atmospheric Research, Victoria.

Page, C.M. , and Jones, R.N. 2001. OzClim: the development of a climate scenario generator for Australia. In: MODSJM 2001 :

International Congress on Modelling and Simulation: proceedings, Australian National University, F. Ghassemi, and others (editors) .

Canberra, ACT: Modelling and Simulation Sociery of Australia and New Zealand. p.

667-671.

Rous Water. 2004. Demand management plan, 2004-2009. Rous Water Regional W ater Supply, Adopted Council Meeting March 2004.

Simons, M., Podger, G . and Cooke, R. 1996.

IQQM - A hydrologic modelling cool for water resource and salinity management.

Environmental Software, Vol 11, Nos. 1-3, pp 185-192, 1996 SRES. 2000. Special report on emissions scenarios (SRES) . Special report on rhe lnrergovernmental Panel on Climate Change. N . Nakicenovic and R. Swarr (eds), Cambridge 2000. hrrp://www.ipcc.ch WMO (2005) WMO statement on the status of the global climate in 2005. World Meteorological Organisation. ht tp://www.wmo.ch/web/Press/index. html

The new DR5000 UV-VIS Spectrophotometer Simply more ...

... 240+ Test Parameters ... Pre-packaged Reagents ... Intuitive Operation

The Complete Instrument & Reagent Analysis System • Automatic method detection with TNT plus ' " - Bar code reader yields simpler, faster results • Built-in accuracy with automatic averaging - Eliminates errors from blanks & cuvettes • Touch Screen Operation - Easy to use, fast set up • USB Data Communications Port - Easy data export and upgrade integrity