New J-FOR+ and Review of PaperWeek and BIOFOR by PAPTAC

ED IT IO N SP EC IA L

J-FOR A PAPTAC PUBLICATION

JOURNAL OF SCIENCE & TECHNOLOGY FOR FOREST PRODUCTS AND PROCESSES VOL. 5, NO. 2

FEATURING A special section on the official coverage recap of PaperWeek and BIOFOR 2016 PaperWeek Executive Panel Discusses Bio-product Commercialization J-FOR’s Technical Papers with its unique blend of traditional and emerging topics

PAPER WEEK

CANADA

2016

SPECIAL REVIEW of PAPTAC CONFERENCES

BIOFOR International Montréal 2016

FOR THE ADVANCEMENT OF THE FOREST INDUSTRY

News, Stories, interviews contributed by:

Paper Advance

Strong bonds lead to unique possibilities – in business as well as in chemistry. And that’s precisely why we’ve acquired AkzoNobel’s paper chemicals business. Additional expertise. A broader technology base. Greater presence where you need us most. It all adds up to a winning partnership with a shared commitment to innovation. We’ve extended our capabilities so you can extend yours. Let’s work together to build value into paper. www.kemira.com/extend-your-capabilities

News, stories, interviews contributed by: PA

J-FOR

Paper Advance

A PAPTAC PUBLICATION

TABLE OF

CONTENTS 5 EDITORIAL

Greg Hay, J-FOR’s publisher

7 INDUSTRY PULSE CONFERENCE REVIEW SECTION

Executive Opening Breakfast Panel

PaperWeek 2016: Out with the Old, in with the New

PaperWeek Conference Review

19 21

PAPTAC Launches its 100th Anniversary Book

22 24

Keynote Speakers

27 30

BIOFOR International Conference Review

Executive Breakfast Panel Discusses Bio-product Commercialization Bio-products: the tip of the iceberg at Canadian conference Business Luncheons

TECHNICAL PAPERS

60 INDEX OF

ADVERTISERS

Kemira Cascades Solenis AstenJohnson Kadant FPAC FPInnovations

2/3 6 8 12 20 26 28

ABB Prosygma Quattro Buckman

Published by:

PAPTAC

Pulp and Paper Technical Association of Canada

For inquiries, please contact: PAPTAC 740 Notre-Dame St. W., suite 1070 Montreal (Quebec) H3C 3X6 CANADA Phone: (514) 392-0265 4

J-FOR

43 51 55 78

Publisher: Greg Hay, PAPTAC Executive Director Co-editor: Stéphan Desjardins, Paper Advance Co-editor: Carmie Lato, PAPTAC Production Specialist: Thomas Périchaud, PAPTAC

AUTOMATIC SELECTION OF RELEVANT DATA FOR PAPER MACHINE DIAGNOSTICS by H. Hytti, A. Nissinen, M. Lauri, H. Koivisto, H. Ihalainen, R. Ritala

CONDENSER COOLING-WATER CONTROL STRATEGY RECOVERS WASTE HEAT FROM POWER GENERATION by D. Barbour, R. Prontack, J. Boisvert

ASB DREDGING AND ODOUR CONTROL by J. Fowler, D. Davies, W. Teetzel

POST-D0 TOWER BRIGHTNESS – USEFUL OR NOT? by J. Goldman, M. Summerford

3D PRINTING – A REVIEW OF TECHNOLOGIES, MARKETS, AND OPPORTUNITIES FOR THE FOREST INDUSTRY by T. Li, J. Aspler, A. Kingsland, L. M. Cormier, and X. Zou

ENVIRONMENTAL EVALUATION OF DURAPULP BIO-COMPOSITE USING LCA: COMPARISON OF TWO APPLICATIONS by F. Hermansson, M. Janssen, and F. Gellerstedt

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

J-FOR

EDITORIAL

A PAPTAC PUBLICATION

How two conferences combined provide for a truly unique environment PAPTAC hosted two conferences early February with its ﬂagship event, PaperWeek Canada, and the newly launched Montreal BIOFOR International. What a great environment the synergy of these two events created as delegates were able to go from one conference to the other in a unique blend of traditional areas dealing with pulp & paper technology advancements, and world-class presentations on the emerging forest bioeconomy. I truly don’t think one can ﬁnd this type of networking anywhere else. Greg M. Hay, Publisher

PaperWeek Canada 2016: Global Knowledge Connections As per its theme for 2016, the largest gathering for the pulp & paper and forest products sector in Canada presented a platform of presentations of the highest standards around topics that are at the heart of the industry’s challenges and opportunities. Household tracks led the conference with complete programs on Packaging (Techpack), Tissue (Tissus Masters), Safety, Management, Energy and Reliability, as well as new segments on Market Pulp (PulpEx) and the Converting Industry. With an impressive roster of world-renowned keynote speakers, technical sessions and special symposia, PaperWeek once again presented a comprehensive program of 40+ sessions, combined with a with a 3-day tradeshow and a series of business luncheons which have become a true industry rendez-vous. Furthermore, the event was held jointly with another major event – Montreal BIOFOR International – a new conference launched by PAPTAC, dedicated to the emerging Forest Bioeconomy. With tracks on biorefinery, biomaterials, government policy, engineering and financial analysis for the sector’s transformation, this premiere of BIOFOR has certainly set the stage for what will become a growing yearly event. On behalf of the PAPTAC Executive Council, I wish to take this opportunity to thank the many speakers, session moderators and program committee members who provided indispensable support toward the success of these two events. We hope you enjoy this official recap and be sure not to miss next year’s gathering!

J-FOR

LEAD ASSOCIATE EDITORS Martin Fairbank Consultant Patrice Mangin CRML/Université du Québec à Trois-Rivières ASSOCIATE EDITORS Thore Berntsson Chalmers Institute of Technology (SWEDEN) Virginie Chambost EnVertis Inc. (CANADA) Christine Chirat Grenoble INP – Pagora (FRANCE) Jorge Luiz Colodette Federal University of Viçosa (BRAZIL) Ron Crotogino ArboraNano (CANADA) Sophie D’Amours Université Laval (CANADA) Robert Dekker (BRAZIL) Gilles Dorris FPInnovations (CANADA) Paul Earl Paul Earl Consulting Inc. (CANADA) W. James Frederick Table Mountain Consulting (USA) Ramin Farnood University of Toronto (CANADA) Gil Garnier Australian Pulp and Paper Institute (AUSTRALIA) Eldon Gunn Dalhousie University (CANADA) Ali Harlin VTT (FINLAND) Mikko Hupa Åbo Akademi University (FINLAND) Mariya Marinova École Polytechnique de Montréal (CANADA) David McDonald JDMcD Consulting Inc. (CANADA) Glen Murphy Oregon State University (USA) Yonghao Ni University of New Brunswick (CANADA) Ivan Pikulik Consultant (CANADA) Risto Ritala Tampere University of Technology (FINLAND) Reyhaneh Shenassa Metso Power (USA) Paul R. Stuart Ecole Polytechnique (CANADA) Trevor Stuthridge FPInnovations (CANADA) Honghi Tran University of Toronto (CANADA)

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

partenaire de votre quotidien Your Everyday Partner Tournée vers les besoins des consommateurs, Cascades met tout en œuvre pour concevoir et mettre en marché des produits innovants. L’entreprise mise sur une offre qui sait répondre aux besoins du marché : des produits à valeur ajoutée, mieux adaptés aux nouvelles réalités, plus performants et plus verts. / Consumeroriented, Cascades strives to create and market innovative products. The company focuses on solutions that meet market needs—more effective and greener value-added products that are better adapted to new realities.

INDUSTRY

PULSE

Paper Advance featured news on the industry

Organizational overhaul at West Fraser West Fraser leadership is undergoing an extensive reorganization, due to a series of retirements, and as part of its advancement and succession programs. Notably, Peter Rippon, Vice-President of the company’s Pulp and Energy division and Dave Lehane, Vice-President of Woodlands, will retire in April and June, respectively. Some, but not all of the organizational changes happening at West Fraser include; Ray Ferris, currently Vice-President of Wood Products, will take on a new role as Executive Vice-President and Chief Operating Oﬃcer. Chris McIver will move on from his position as Vice-President of Lumber Sales and Corporate Development, to become Vice-President of Sales and Marketing. Sean McLaren, currently the company’s VicePresident of U.S. Lumber Operations, has been appointed Vice-President, U.S. Lumber. Ted Seraphim, President and Chief Executive Oﬃcer, recognized the important contributions made by Rippon and Lehane during their careers and is enthusiastic about the new additions to his company’s executive team.

FPAC gets new CEO The Forest Products Association of Canada (FPAC) has a new leader at its helm. Derek Nighbor, former senior VP of Food and Consumer Products of Canada, is the association’s new CEO. With a background in the oversight of government relationships, development and partnerships, Nighbor is an excellent ﬁt for FPAC, according to the association’s Board of Directors Chair Curt Stevens. “His years of government and association experience will serve us well as FPAC continues to highlight our global leadership in sustainable forest management.” “I am proud to join FPAC and look forward to working with its members and many partners along the value chain so the industry can reach its full potential,” Nighbor said of his new role.

J-FOR

Pöyry predicts the future of pulp and paper A recent study published by Pöyry Management Consulting forecasts global demand for paper and paperboard will grow by 1% a year between now and 2030, with most of the increase in demand being driven by emerging markets in Asia. The study also highlighted the importance of the slowdown in China’s economy and predicts that when coupled with a faltering paper market and the simultaneous introduction of several new pulp lines, the global pulp industry’s value could slide. The Consultancy proposes that investing in low cost areas of the industry before demand heats up, and an inevitable restructuring will help restore equilibrium to the supply/demand balance.

Resolute Forest Products Tracks Fibre Resolute Forest Products is taking a leadership role in ensuring it can account for the source of all fibres that make their way into its final products. By using a fibre tracking system designed to trace products back to their sources and make sure they are responsible and legally-harvested, Resolute is doing its part to ensure the sustainability of the future fibre supply. Resolute employs such mechanisms in all of its North American pulp, paper mills and wood products facilities. The system allows the company to confirm for its customers that they have avoided timber and timber products from unacceptable sources. Going one step further, the tracking systems can be manipulated in such a way as to keep track of the use of genetically modified trees, or violations of indigenous rights, or workers’ rights.

Upgrades at Swedish mill One of the world’s largest paper mills recently undertook a series of projects and capital upgrades. One of these investments will now protect the structures of the Södra Cell Värö

mill, located in Sweden, against all kinds of weather, and the acidity in wood chips. In a separate expansion the mill has altered its production to now produce only chlorinefree pulp. The facility will ramp up to 700,000 tons of high-grade chlorine-free softwood pulp, mainly used for tissue products, later this year. The facility additionally produces electricity and steam both for its own use and for sale to the grid, and supplies the municipal heating network to nearby Varberg.

Belgium Home to World’s Largest Greenfield Plant Belgian Eco Energy is partnering with GE to build the largest greenfield plant in the world. The facility will be powered by wood chips and agro residues and will be 100% biomassfired. Once constructed, the plant will generate roughly 215 megawatts of cleaner energy to be used by both industry and neighbouring households. When operating in cogeneration mode, the plant will reach over 60% efficiency. The new plant will additionally be fitted with a thermal energy heating system. Commercial operation is scheduled for 2019.

iFoodbag Gets Leg Up with European Grant The company behind iFoodbag, the Swedishdesigned water-resistant carrier bag that protects chilled and frozen groceries for up to 24 hours, has been awarded a grant from the European Commission. The grant is part of the Commission’s Horizon 2020 SME instrument program, and will give the company the opportunity to develop a second generation of the bag and work with a €2.5M project budget as it looks to expand outside Sweden and into the rest of Europe. The company says it plans to make the second generation of its carrier bag even more environmentally-friendly, cheaper to produce, and able to keep its contents chilled for a longer period of time.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

THE GENESIS

OF YOUR

SOLUTIONS.

Meet Solenis. A new name for a team with decades of experience. You used to know us as Ashland Water Technologies. Today weâ&#x20AC;&#x2122;re Solenis, the world leader in pulp and paper chemistries. With years of process experience, advanced technology and 3,500 employees worldwide, the Solenis team is ready to deliver the solutions you need. Meet your team at solenis.com

SPECIAL EVENT REVIEW SECTION

Global Knowledge Connections

BIOFOR International Montréal 2016

Recognizing the Forest Fibre’s Value

PAPER WEEK

CANADA

2016

HIGHLIGHTS

PaperWeek 2016: Out with the Old, in with the New

nother successful PaperWeek has come to a close. The annual gathering of Canada’s pulp and paper industry leaders, service providers and researchers marked what many consider a turning point in a long history of technological advancements and partnerships, and it did so with a revamped and innovative approach to its format. Building on past successes designed to increase attendance, PaperWeek 2016 featured business sessions dedicated expressly to individual forest companies, allowing participants to learn directly from experts in their particular field, and to engage in discussions on more streamlined topics. Both PAPTAC Executive Director Greg Hay and PAPTAC Chairman Robert Dufresne were extremely pleased with the new approach, and the outcome. “We were able to attract a very good number of mill participants to each dedicated session – there were at times more than 70 managers in a room,” Hay confirmed. The closeddoor sessions helped create an environment conducive to open and frank discussion and, as Dufresne highlighted, reflected an industry that is ‘very much anchored in the present, but is always looking towards the future.”

Robert Dufresne

Greg Hay

PAPTAC organizers took another step in a new direction and hosted several additional events during the same time as PaperWeek; Montreal BIOFOR International (conference devoted to the forest bioeconomy) PulpEx, (a full-day conference devoted to the market pulp sector), TechPack (insight on the latest updates in market perspectives for the global packaging industry), and Tissue Masters (conference focused on tissue production, operations optimization and technology developments and updates on market trends). The addition of these events underscored the inextricable links forestry shares with a growing number of parallel sectors, and, as Hay noted, reﬂects the importance of knowledge-sharing to the forestry sector in Canada. “The conference remains one of the most respected in the world and that most of the basic principles that motivated its foundation in 1915 are still applicable today: stimulating interest in the science and technology behind forest products and providing a forum to exchange ideas between members.”

While PaperWeek 2016 certainly broke new ground in terms of formatting and sharing the stage with other events, the depth and breadth of programming certainly continued this year, with technical sessions organized along thematic tracks and packed to the brim with information on the latest developments in energy, management, reliability, safety, and many others. Once again, participants were encouraged to converge in the exhibition hall, where lounges provided areas in which to discuss, relax and work in and amongst the approximate 50 exhibitors. The addition of several parallel events enabled delegates with opportunities to make numerous new connections.

J-FOR

PaperWeek has been long renowned for its ability to attract dynamic, high-profile and engaging keynote speakers, and 2016 marked yet another year in continuing this tradition. In keeping with the ‘Global Knowledge Connections’ theme of PaperWeek 2016, members of the Executive Opening Breakfast Panel, comprised of Robert Graham, CEO of Ensyn Corporation, Bruno Marcoccia, Director of R&D at Domtar, Gurminder Minhas, Managing Director, Performance Biofilaments and Rod Albers, Manager - Energy and Bioproduct Development, West Fraser, discussed strategies they have each respectively implemented to seize opportunities presented by the bioeconomy. Shedding light on the government’s perspective on Canadian forestry was Kim Rudd, Parliamentary Secretary to the Minister of Natural Resources. Rudd discussed her view that resource development is strongly linked to the economic benefits associated with a cleaner environment and climate change mitigation.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Paper Advance

“Climate change is the biggest challenge of our generation. Canada’s forest industry is part of the climate change strategy. Ten percent of our Paris commitment to greenhouse gas reducKim Rudd tions will come from the forest products industry,” she said. Rudd’s view that resource development is a ‘nation-building exercise’ was met with wide audience approval.

Suzanne Blanchet

Sébastien Corbeil

The Business Luncheon allowed delegates to switch gears slightly, and tune into insights delivered by keynote speaker Tim Cork, motivational speaker and coach. He urged the audience to “be the best you can be and the best for your company.” This sentiment seemed to resound strongly with delegates.

Tim Cork

As always, the PaperWeek Poster Session was a conference highlight. Featuring posters from nearly two dozen Canadian University students, the 2016 PaperWeek Theo van de Ven Student Community Poster Session was tremendously successful. The first prize was awarded to Leila Jowkarderis of McGill University for her poster entitled “Mesh size analysis of cellulose nano fibril hydrogels with solute exclusion and PFG-NMR spectroscopy.” Second prize went to Abbas Nikbakht of UBC for “The yield stress of refined wood fibre suspension.” The third prize went to Xingye An, student at the University of New Brunswick for “Nanofibrillated cellulose (NFC) as a carrier of TiO2 nanoparticles (TNPs) for photocatalytic hydrogen generation.” Xingye An

Carman Allison

Éric Ashby

Other keynote speakers discussed their company’s recent experiences in a constantly evolving industry. Suzanne Blanchet, Senior Vice President, Corporate Development at Cascades explained how her company is adapting in response to today's environment, and the impacts of the company's new leadership. Sébastien Corbeil, President and CEO of CelluForce announced that significant commercial trials using the company’s cellulose nanocrystals in oil and gas drilling fluids and paper wet end chemistry would materialize later this year. Éric Ashby, Mill Manager and Vice-President at Domtar Windsor gave a talk on ‘Human Performance Improvement,’ and the novel efficiencies it can offer when applied to the pulp and paper sector. Carman Allison, Vice President, Consumer Insights with Nielsen Canada, explored emerging consumer and retail trends and reflected on their effects on the consumer package goods business.

J-FOR

Leila Jowkarderis

Abbas Nikbakht

The closing of PaperWeek 2016 came with yet another announcement by PAPTAC; the launch of a commemorative book in recognition of the industry’s centennial Anniversary. Greg Hay announced the publication of the book, which marks one hundred years of history for an industry that has been, and remains the industrial flagship of Canada. Hay admitted that work on the book continued right up until the last minute, noting that fitting one one years of history into 246 pages is not an easy task! In his words, the book “pays tribute to the most important technological advances, major scientific breakthroughs and highlights the key moments in the evolution of the industry over the past century.” It is also a tribute to a strongly knit community that shaped Canadian industrial development.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

the trade has forever changed. And yet, you’re still here. You’re still here because paper still matters. Because, be it made from eucalyptus, sugar cane, or conventional pulp, paper has the capacity to package our ideas, build tomorrow’s cities and support our daily lives. At AstenJohnson, we’ve seen a lot these past 225 years. Industries have been revolutionized, wars fought, history made. And this product we produce has surely changed. But our passion was, is, and will always be paper. We know the future is full of possibility, and we’re ready to step with you into tomorrow. We’re ready to reconsider every part of this process, ready to recruit fresh talent and make something new. Something great. At AstenJohnson, we’re made of a tougher fiber. And we’re ready to reimagine what paper means to the world. That’s why we’re not going anywhere. And neither are you.

PAPERWEEK REVIEW

PAPER WEEK

CANADA

2016

Pulp EX PaperWeek's Market Pulp Program

Sodium Chlorate Exports on the Rise The export of Canadian sodium chlorate is expected to increase in 2017 from 490,000 Mt to 515,000 Mt. According to HIS Chemicals Stephanie Koenig, this increase in demand will allow the market very little tolerance for unplanned outages. In North America, 80% of the sodium chlorate is controlled by three producers, who churn out 1.6 Mt per year, Koenig confirmed. The U.S. consumes 72% of all the sodium chlorate produced in North America. Of all the sodium chlorate the U.S consumes, Canada produces 61% of it. With electricity representing the lion’s share of the cost in producing sodium chlorate, American producers have benefited from a large price reduction on account of the low Canadian dollar. The reverse is true north of the border, which is experiencing a sharp increase in electricity prices, particularly in the west. HIS Chemicals forecast that while Canadian exports to the U.S. are destined for an increase over the next two years, rebounding after the crash of 2014, high electricity prices may erode profit margins.

The golden rules of bleaching Clean pulp, good pH, timing and temperature, and a balanced chemical use across the plant are the golden rules to bleach plant optimization, according to expert consultant Paul Earl. Clean pulp “You can save a lot of money by improving washing,” says Paul Earl, stating that 15 to 20% of chlorine dioxide (ClO2) is consumed by carryover. Also, there needs to be a low Kappa variability.

Paper Advance

J-FOR

Optimized Conditions The optimal pH is different at each stage of the pulp-making process. Those measures are also different for softwood and hardwood. Being aware of the optimal pH and using the appropriate amount of bleaching chemicals as a result, can allow for significant savings, Earl noted. Since hydrogen atoms are active at high temperature, pH is also temperature sensitive, data taken at 75 °C is misleading. Thus, pulp making mills should build data tables to know what is the real pH of the solution at 25 °C, and use those corrected temperatures, adds the expert. Earl also explains that a 10 °C increase will double the rate of reaction, making it faster, not better. Balancing the Chemicals Balancing chemicals throughout the process can also be very helpful. For example, reducing the amount of ClO2 at the initial stage of bulk delignification can result in a net savings of 3.4 kg/t ClO2, even if you need to increase your use at the residual delignification stage.

TECHPACK

PaperWeek's Packaging Program

Trends in the Recycled Cardboard Industry Lighter, greener, safer, smarter but also more expensive to produce, this is the current trend in the recycled paper industry. “Consumers are looking for more eﬃcient, more eﬀective, and more environmentally friendly products, prompting producers to reduce packaging weight”, says Marco Pescantin, Engineer for Quattro Paper Technologies, a company based in Baden, Ontario. To reduce the weight of the paper, while

producing high performance products, producers will face several challenges. On the one hand, recycled paper and board varies in quality, which can affect productivity. According to Marco Pescantin, producers should consider using different lines for different grades of products. In addition, machines will have to be faster in order to convert the same amount of products and thus be profitable; however, Pescantin, an expert in this field, further notes that at faster speeds particular attention should be paid to breakage. Trends In the coming years, the paper and board recycling industry will turn to high technology. The goal: to produce safe and sustainable packaging that “speaks” to the customer.

Improved Control Systems Would you like to know the weight, moisture, strength and ﬁber orientation at all times? It is what the new and improved control systems allow you to do! In this day and age, data is calculated in real time in order to improve performance and reduce costs. Pulp and paper manufacturers are not immune to this trend. According to Steven Bale (Valmet), a participant at TechPack, new control systems are now faster, more accurate, and they can measure several characteristics. Knowledge of key data on the quality of the pulp facilitates adjustments and improves the yield. For example, the implementation of a quality control system in a recycled paper mill improved weight controls and moisture by 30%. The data collected by such quality checks are then transmitted online, making them readily available to more members of a work team, and thus facilitating data analysis.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

PAPER WEEK

CANADA

2016

This interface also allows for advanced diagnostics or to obtain remote assistance, thus reducing machine downtime. Continuous Improvement

MANAGEMENT SESSION Change Management or Managing Changes?

Using good design, it is possible to follow the evolution of technology, because each scanner generation can be improved with the addition of new modules. Pulp analyzers make it possible to take real-time images of the pulp quality. With the data, it is possible to know the strength of the pulp and make the necessary corrections to resolve critical situations, notes Steven Bale. In short, quality controls improve the stability of production while reducing energy consumption; gains that enable a quick return on investment.

TISSUE

MASTERS

Tissue Masters – another success for the Tissue Segment at PaperWeek For a fourth year in a row, PaperWeek’s tissue conference was a triumph. In fact, this year the tissue conference was successfully kicked oﬀ by keynote speaker Suzanne Blanchet of Cascades. Under the “Tissue Masters” banner, Steven Sage VP Kruger Products, Stéphan Larivière Director at FPInnovations and David Embley VP Irving Consumer Products, put together two solid tissue sessions pertaining to basics of tissue making processes and properties innovations in tissue manufacturing. The value chain was well represented from chemicals suppliers such as Solenis, to suppliers such as Kadant and Voith.

Jacinthe Bergevin

“What if it was time to shelve the ‘change management initiatives’ as we know them? What if change management initiatives in their traditional form were not the solution anymore?” Jacinthe Bergevin’s questions provoked a lot of thought at Thursday’s Management Track Session. Bergevin is an organizational development consultant with experience in the pulp and paper industry; she was a change management leader at Abitibi-Consolidated from 1996-1998 and senior consultant in organizational change at Domtar from 1989 to 1991. Embrace change “Organizations now evolve in a world that is more and more unpredictable and complex. Managers and employees have to deal on a day-to-day basis with shifting markets, cutting-edge technologies, globalization, social media etc,” Bergevin said. “If you try to implement a change management timeline, ‘life’ will always get in the way! The boss leaves, layoﬀs are mandated, a plant’s paper machine shuts down, etc. The common view that every change can be managed in a planned and controlled fashion in order to bring an organization back to a steady state is now a fallacy.” While we all have assumptions about how organizations work, it is rightly expected

J-FOR

of organizations to be structured in such a way that they can deliver predictable outcomes, and that relying on plans and processes is what makes them work. We also often assume that organizations operate in a fairly stable and predictable environment. “Wrong, wrong, wrong! Bergevin said. “Changes are no longer predictable and planned (were they ever?). They are now part of doing business, no matter the type of work environment.” “If managing change in the ways we are used to is not the solution for most of the changes we experience in organization, what can leaders do?’ asked Jean Fuller, of Work Design Inc. Fuller, co-presented with Bergevin to introduce solutions and insight into the type of mind-set and skills required for managers to make real-time changes and to engage team members around the goal of taking action proactively. Essentially, this starts with asking questions and allowing employees to experience their work and organization diﬀerently without having to wait for a plan. The questions asked need to be done so with the goal of ﬁnding opportunities to generate energy and actions, to connect people, to generate ideas and initiatives and to learn from our experiences and those of others. This PaperWeek Canada 2016 session was organized around short workshops for participants to discuss their own situations and to share them with the group. Some were fortunate to discuss their unique situations and receive free advice from the experts!

INCREASE YOUR PokaFORMANCE “Welcome to the social industrial revolution and increase your performance,” Alexandre Leclerc, president of social industrial platform ﬁrm Poka, said in welcome to participants in Thursday’s Management Track session.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Paper Advance

PAPERWEEK REVIEW Poka may very well be the ﬁrst of its kind in the application world. Created in 2012 with $2.5 million in funding from iNovia Capital, SoftTech VC and Nicolas DarveauGarneau from Google, Leclerc underscored his company’s unique approach to business. “Poka allows plant managers to focus on eﬃcient training, knowledge retention and real-time information says Mr. Leclerc, our mission is to provide companies with the ability to use and share their best practices, thoughts and knowledge, anywhere, any time.”

Alexandre Leclerc

Social industrial revolution Concretely, Poka is a social media platform that can be used in pulp and paper plants and in mills via an iPad application. Designed for manufacturers to train staﬀ using videos, other interactive content can be shared in real time through an iPad app. Leclerc claims the technology is not only “cool” and “pleasant” to use, but that it is also helping drive “the social industrial revolution.” Poka is proposing is a collaborative online approach to break down the traditional silos that often divide the workforce, consultants, management and factories. “The concept with Poka basically is that you are part of a team.” The platform could potentially increase productivity while empowering employees with better training and conﬁdence. The possibility to “like” some of the videos and tips by fellow factory workers gives

J-FOR

them a sense of pride in the work they do, explained Leclerc. Troubleshooting is an important feature of the application but it is far from being the only one. Through the Poka application, workers can record improvements by their teams and share them with their firm’s sister plants whether they are in the same city or on another continent. This approach is known as the ‘ecosystem,’ according to Leclerc. Because the tool is not intended to be used solely by a local team, multiple factories from the same company can access and benefit from a post. The application allows users to record a specific problem, to record an associated solution and to store it as long as is required, or at least until a machine or industrial process is replaced! While factory floor video producers and authors are unlikely to be nominated for an Oscar or for the Pulitzer Prize in recognition of their work, they certainly participate in what will undoubtedly become a very efficient management tool.

RELIABILITY SESSION Reliability: Still Plain Good Sense

P. Corriveau

M. Gonzalez

J.R. Claveau

The first session of PaperWeek’s Reliability Track kicked off this morning with presentations from Jean-Robert Claveau of consultancy firm STI Maintenance, and Manuel Gonzalez, from technology provider SKF Canada. Given that many plants experience similar reliability challenges, the Reliability Track has been successfully offered at PaperWeek in previous years. This year’s iteration focused on best practice case studies, as well as discussions of experiences from other industries.

The session was led by Patrick Corriveau, Vice-President of Operations at Resolute. He noted that best practices in reliability are a central concern to industry today. With this in mind, he underscored the importance conference organizers placed on ensuring participants received the most recent advancements and benchmark standards in operational reliability. “What is considered world class when it comes to reliability, and where does the pulp and paper industry stand in relation to the proper assessment of evolution in business culture?” Corriveau queried of the audience. Noting the pulp and paper industry needs reliability in order to remain competitive, Corriveau introduced the two speakers, who proceeded to discuss the experiences of other industry sectors with reliability. In his presentation entitled “The impact of evolution in cultures which lead to best practices in Reliability'', STI Maintenance’s Claveau highlighted the importance of using plain, common sense when operationalizing reliability in a plant. Color-coding for tools and daily operation team meetings were just a few of the logical best practices he suggested. Claveau stressed to the audience the beneﬁts of solid collaboration between operations and maintenance people in a plant. “Common goals should be established once a culture of trust is set,” he advised. SKF Canada’s Gonzalez presented his ﬁrm’s experience with designing machinery for reliability. He focused on optimal solutions for paper machines and fans. Recalling that an important focus of SKF’s technology development is to reduce the environmental impact of an asset during its lifecycle, Gonzalez described the drying Yankee Cylinders and the typical causes for bearing failure. Lubrication is behind 36% of all failures, while contamination accounts for 24%. With respect to recommendations for the protection of bearings during operation,

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

PAPER WEEK

CANADA

2016

Gonzales suggested lubricating seal lips or purging the excess of grease. Reliable production provides an indisputable competitive advantage, and yet remains an ongoing challenge. As such, the role of managers in developing, documenting, communicating and reinforcing better work processes is central. Unfortunately, eﬀective people sometimes work in ineﬀective environments, which is when setting aside time to re-envision the way things are done, makes plain common sense.

Expecting The Unexpected

Jake Zwart

Serge Doucet

When addressed with insight, reliability challenges at the mill level can foster and sustain operational eﬀectiveness. A combination of optimal technical systems and Shuman intervention is required to maintain this eﬀectiveness. Reliability Track Session leader Patrick Corriveau, Vice-president of Operations at Resolute, believes the decision to “manage change” can occur for a number of reasons. While compliance with laws and regulations is often one such reason, the opportunity to generate savings through a reduction in maintenance or operating costs is usually a higher and more rewarding motivation. Jake Zwart from Spectrum Technologies gave a talk entitled “Sensitivity Analysis of Paper Machine Roll Critical Speeds'' He delivered thought-provoking insight on a range of topics, though one area of expertise found particular resonance with the audience. Zwart is a mechanical engineer with over 25 years of experience in research, development and consulting. Throughout his career he gathered and deciphered, as he described it, “data related to vibration including rotating machinery analysis,

J-FOR

modal analysis, operating deflection shape analysis and pressure pulsation analysis.” Zwart undertook vibration testing and deflection shape analysis to address associated problems in a systematic way in order to target the source of specific machinery reliability problems. The vibration and pressure pulsation specialist is more than a troubleshooter; with the data he gleaned he additionally outlined the extent to which vibration analysis and control minimizes long-term maintenance costs and prevents costly failures. The second presentation focused on best practices in maintenance scheduling and planning. Serge Doucet of ProSygma is a planning and scheduling advisor and in his discussion, he explained the methods and concepts of recognized best practices in this area of expertise. Using case studies and audience participation to illustrate collaboration, he expertly fielded questions from the audience on the reasons behind the success or failures of companies who need to implement and maintain these best practices. Doucet ran through a list of misconceptions regarding maintenance tasks, stressing the potentially severe impacts of poor planning. One such misconception, more of a myth, according to Doucet, is that all operations must be planned in excruciating detail, just as a pulp and paper ‘monk’ would be inclined to do. This is misguided, according to Doucet, who said that a welldefined priority system for work orders is an appropriate approach, and allows supervisors the flexibility to deviate from the typical planning schedule to address unexpected events as they arise.

Reliability: “A Journey, not a Destination While plant reliability is often considered one of the greatest money-saving solutions

F. Tremblay

Guy Nollet

in an industrial environment, due to limited resources, many managers resist making the investments that would allow them to increase effectiveness and efficiency of existing resources. This theme was addressed in two of Tuesday’s presentations. Reliability Track Session leader Patrick Corriveau, Vice-president of Operations at Resolute, welcomed participants to the final session of the well-attended Reliability Track of PaperWeek Canada 2016. He introduced Guy Nollet, of Laurentide Controls’s Reliability team. Laurentide is a supplier of automation solutions in Eastern Canada. Nollet delivered a presentation entitled ''Maintenance: First Quartile companies do better. Where are you?'' Quoting Winston S. Churchill, Nollet observed that while reliability is not typically considered a ‘project’ in a mill environment, that perhaps it should be seen as such. “Sometimes doing your best is not good enough. Sometimes you must do what is required.” Churchill would probably agree that a reliability strategy should be launched and managed like a project, “because it is a project”! “Technologies are enablers,” Nollet noted, explaining that once they are in place it is time to engage the people. Nollet suggested establishing a business case from the onset; benchmarking (“knowing who you are”), obtaining management support; instigating an implementation plan and investing wisely in the low hanging fruits. He believes that “reliability is a process, not a destination nor an abstract theoretical goal.” Nollet’s discussion was followed by Frédéric Tremblay from LTI Group, a firm offering expertise in lubrication and oil analysis. Tremblay highlighted optimal approaches

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Paper Advance

PAPERWEEK REVIEW and best practices in lubrication and reliability, stressing the importance of maintenance as witnessed by his own company over the past three decades. “85% of problems in lubrication stem from contamination,” he conﬁrmed. “The product is very seldom the problem itself. Contaminants include solid particles, varnish or water and relative humidity. In light of this, it becomes extremely important to keep oils in the equipment, clean, dry and cold. In order to make sure this is the case, a few approaches are available, including ﬁltration and the proactive monitoring of metals.

ENERGY SESSION Mill Energy Reduction and Eﬃciency

Ben Janvier

The final session of PaperWeek Canada 2016’s Energy Track was a two-part presentation on research by Enero Solutions. The Montreal-based engineering process and automation control solutions firm that assists manufacturing, pulp and paper, oil and gas and power companies bolster their operational control and profitability through advanced engineering technologies, was represented by senior performance consultant Ben Janvier, who described some of the challenges facing today’s pulp and paper sector. Rising energy and fibre costs, high transportation costs, falling demand in developed nations for pulp and paper products and increased competition from emerging economies were noted as paramount obstacles. On a macro level, Janvier noted that energy production to fuel power and pulp mills is the biggest

J-FOR

culprit when it comes to producing greenhouse gas emissions by Canadian pulp and paper manufacturing processes. In response to the high energy costs, Enero Solutions emphasized the benefits of model-based control of steam plants in order to reduce the price tag. The first of Janvier’s presentations discussed the reduction of fossil fuel reliance through the use of dynamic integration and advanced controls. The second was entitled, ‘Energy Efficiency and Profit Optimization Using Advanced Controls in Cogeneration Applications.’ In this subsequent talk, Janvier described an energy dashboard, a tool that “continuously informs the operation team if better control decisions would increase plant profitability.” Dashboards have two main functions: the first to monitor performance of existing advanced controls, and the second to identify malfunction within these existing controls. The latter offers an opportunity to identify room for optimization through increased profits and a “significant reduction in steam header pressure variability and maximum steam availability during a mill island mode operation.”

Energy: Radical Changes Ahead? Reducing energy costs and consumption is a top of priority for mills. In addition to fibre, the primary papermaking process inputs are water, chemicals and energy. PaperWeek Canada 2016 Energy Track showcased recent initiatives undertaken by leading researchers in the area of energy management and efficiency. FPInnovations’ Enrique Mateos-Espejel presented research he is working on to address process operating regimes in energy efficiency studies, and offered insight into the way his research can be applied to the pulp and paper industry.

Enrique Mateos-Espejel

Mateos-Espejel was the recipient of the PAPTAC I.H. Weldon award in 2011 for his work on energy sustainability of the green integrated forest biorefinery. A member of FPInnovations’ Process Engineering program, Mateos-Espejel holds a Ph.D. in Chemical Engineering from École Polytechnique de Montréal. He joined FPInnovations in 2010, and his area of expertise is energy efficiency optimization, process integration and process modeling. He has developed holistic methodologies to improve pulp and paper process efficiencies that give thought to energy, water and technoeconomic aspects. Mateos-Espejel discussed a site-wide approach to identify and correct inefficiencies in complex industrial processes while accounting for individual operation regimes. This work ensued from a global analysis of the entire process, including the interactions between its different parts. This allowed Mateos-Espejel and his team to determine where heat was being used, where it could be recovered and the optimal use for that heat in the facility. In his concluding remarks, Mateos-Espejel noted that, “pulp and paper mills are increasingly becoming facilities with various operating regimes,” making it necessary to “adjust to high variations (e.g. price, demand, grades, etc.) in the context of different modes of operation affecting mass and energy balances.” Thankfully, in turn, this can result in the discovery of energy process inefficiencies.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

PAPER WEEK

CANADA

2016

Dave Tomihiro

The following presentation by Dave Tomihiro of Valmet was entitled ‘Steam Profiler advances - A new beginning.’ Tomihiro’s research applies to steamboxes. He described the characteristics of advanced steam delivery technology and highlighted the specific case of a diffuser plate designed to create a uniform steam blanket in each profiling zone. Through this research, it was discovered that steam condenses inside the sheet, maximizing heat transfer efficiency, and that steam spillage is minimized, which increases profiling capability. Pulp, dry crepe tissue, fine paper and linerboard applications were each described in turn and results indicated improved press dryness and moisture profiles.

Etienne Bernier

Etienne Bernier of CanmetEnergy introduces his presentation as a follow-up on Enrique Mateos-Espejel’s earlier presentation. The last presentation of the session is entitled ''Project Implementation Strategy for Eﬀective Energy Eﬃciency Improvements in the Pulp and Paper Industry.'' Etienne Bernier holds a Ph.D. in chemical engineering from Ecole Polytechnique de Montréal. Since 2012, he works as a process engineer with CanmetENERGY, a research center of Natural Resources Canada.

J-FOR

“CanmetENERGY in collaboration with the Canadian Forest Service, FPInnovations and other partners, is investigating, evaluating and optimizing technology platforms for the integration of biorefinery processes into pulp and paper mills.” Mr. Bernier helps develop software tools to optimize the design and operation of industrial cogeneration systems. Through research collaboration with FPInnovations, he participated in energy efficiency studies in five Canadian pulp and paper mills. Etienne Bernier describes the industrial context in which the research fits and reminds participants “energy efficiency is at the core of Canada’s pulp and paper industry efforts to reduce operating costs and develop a low-carbon economy.” After describing the larger context in more details, Etienne Bernier lays out some of his team’s initial assumptions including this very straightforward one: “operational and retrofit design improvements can provide better return on investment when projects are part of a good long term vision.” Conclusions presented are the following: - “Cross-effects between projects need to be identified and highlighted to facifacilitate the decision for project implementation; - Project modifications should avoid bottlenecking the long term project vision; - Water-energy system interactions should be reviewed at a local and global level; - A long term global vision for project implementation should be developed.” This research was extremely well-received by, and of great interest to, meeting participants. Jean Paris, of École Polytechnique de Montréal and a veteran of the field, praised the work undertaken by Enrique

Mateos-Espejel and Etienne Bernier. Paris said he believe what the two researchers are proposing represents a “radical change” in comparison to traditional approaches and “this should not be described as ‘optimization’ but as genuine improvements in processes.”

SAFETY TRACK Accept the error to be more eﬃcient

Eric Ashby

To improve the performance and safety of its employees, the Domtar (Windsor) mill completely revised its philosophy. However, in order to do so, it first had to face up to its mistakes and improve processes. Result: a new working philosophy takes root. A whole culture of health and safety at work has been built and instilled at the Domtar (Windsor) facility over the past 10 years. "Every employee made five errors per hour. Humans makes mistakes and to err is human”, humbly admits Eric Ashby, Mill Manager of the Windsor mill. The first step is to accept this reality! It is then that the employer realizes that he bears some responsibility in regards to the health and safety of his workers. "Every decision has an impact on the health and safety of workers. We must collectively learn to minimize the risks, " adds Ashby. There are actually four major stages of development within a company on the philosophy on health and safety in the

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Paper Advance

PAPERWEEK REVIEW workplace. First is the natural instinct, where no concrete plan exists. Second, there is the supervision stage, where employees are sanctioned when they break the rules. In the third stage, the worker is sensitized to his well-being and takes charge of it. The ﬁnal stage is that of interdependence, where every worker has the courage to step-up and intervene for his health and well-being as well as that of his colleagues. This change in mentality is part of the development of a company to ensure that safety is rooted in its culture. The aim of the operation is to improve human performance, better known by the acronym HPI (Human Performance ENERGY SESSION Improvement). This philosophy be used not only to improve worker safety, but

also to improve the quality, reliability and environmental aspects”, said Mr. Ashby, "the eﬃciency improvements are enormous because they allows us to do that type of investigation, regardless of the industry." Rather than blaming individuals, Domtar chose to learn from its mistakes and simply its procedures, as over 70% of human errors are due to poor procedures. In order to reduce the risk of serious accidents to a minimum, Domtar took inspiration from the aviation and nuclear industries. known for their high risk factors. "A worker at a nuclear plant cannot justify an error due to fatigue or frustration, because one wrong move could cost the lives of thousands of people," states Mr.

Ashby. Procedures that ensure safety, such as audits and veriﬁcations, are therefore required. Under this principle, the risk factor is greatly reduced if two people perform the same task. It is this kind of process that achieves interdependence, whereby every person in a company plays a key role in ensuring the safety of his co-workers. There is no miracle recipe to ensure safety, it is something that must be worked at at every moment; but, if an operation is dangerous, it is essential to place barriers between the employee and the task at hand, notes Eric Ashby, otherwise the accident is predictable and the company will have failed in its task.

PAPTAC Launches its 100th Anniversary Book! The idea behind this book is to paint a portrait of the industry's evolution over the past century as seen through the eyes of PAPTAC. One should not expect a book that is inward looking about the association but rather a book about the industry; with the unique angle from which PAPTAC has witnessed the key landmarks, milestones and breakthroughs. For 100 years PAPTAC has had a front-row seat. Featuring editorials from world experts outlining key developments in the sector's technologies and processes, the book's first objective is to serve as a branding tool for the industry. We have often been told that the industry needed to be better at telling its story. This is what it's all about. PAPTAC's mission will be to ensure it finds itself in governmental offices, chambers of commerce, corporate offices, forest products customers, allied sectors, our network of international partners, and as many stakeholders as possible. We have had the privilege of working with whom we call our storytellers - those who contributed the articles and many of which are with us today. We thank them for their immense knowledge and passion. One of the key elements behind the concept of the book is to feature the industry's key players – the pulp and paper and forest products companies that play an essential role in the sector's development and in the Canadian economic landscape. This would not have been possible without the participation of these organizations and we thank them for embarking on this journey and for seizing this wonderful opportunity, as they tell their story and vision. This book builds on the past century as the foundation for the next 100 years. Those who would like to place an order to purchase a copy of the book can contact Carmie Lato (clato@paptac.ca / 514-392-6969).

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Executive Opening Breakfast Panel

Executive Breakfast Panel Discusses Bio-product Commercialization

The conference’s ﬁrst session on Tuesday was a panel discussion

of various and diverse bioreﬁning initiatives in the Canadian industry, presided over by moderator Jason Linkewich of LinksEdge Ltd.

The first presenter was Rod Albers of West Fraser. He described the company’s 50+ year history developing downstream products from its core lumber business, including kraft and mechanical pulp, plywood, MDF and LVL. With respect to bio-products he says that “The key to profitable bio-products is that it has to have a strong wood products business as its base”, and that is wood products from its sawmills in West Fraser’s case. “We are looking for bolt-on processes and near-commercial technology that add value and synergy to our existing processes. This process of extracting value comes from FPAC and its Biopathways network with support from FPInnovations and NRCan’s IFIT program.” he adds. Current initiatives include electricity generation sold to the grid from Rankin cycle heat recovery and bio-methane projects, activated carbon and lignin-based products from FPInnovation’s lignin extraction process now being commissioned at the Hinton kraft pulp mill. West Fraser plans to use the product initially as a plywood adhesive but other external applications are possible, Albers says. Next, Gurminder Minhas of Performance BioFilaments, a joint venture of Resolute Forest Products and Mercer International, described the application development of their cellulose filaments (CF) now being produced at a 5 t/d demonstration plant at Kruger’s Trois Rivieres mill. Through functional modification he says that CF is a “very adaptable product” with applications in fiber-reinforced plastic composites, concrete, paints and coatings. Automobile parts are a particular application that he singles out as auto manufacturers are looking at lightweighting their vehicles to boost gasoline mileage. CF is a high strength and low weight substitute for fibreglass.

Robert Graham of Ensyn presented his company’s now commercial Renewable Fuel Oil (RFO) process which turns residual forest biomass into a liquid bio-hydrocarbon using its patented rapid thermal processing technology. He describes the process as being “very simple, with no catalysts or hydrogen.” The liquid product now being produced at its 3 million gal/yr Renfrew, Ontario plant is sold now as a substitute for petroleum-based heating fuel oil and is being demonstrated as a co-processing feedstock at oil refineries for producing gasoline and diesel fuels. A hospital in New Hampshire has now converted entirely to RFO, thereby replacing traditional fuel oil. RFO’s use as a potential refinery drop-in fuel is awaiting USA approval for toxicity and emissions testing. The fuel market is huge according to Graham even with 1% to 2% substitution in refinery feedstock. Finally, Bruno Marcoccia of Domtar described the company’s “bio portfolio” approach in which various initiatives with different investment potentials, time lines and risks are being evaluated.

Jason Linkewich

Rod Albers

Gurminder Minhas

J-FOR

Bruno Marcoccia

Robert Graham

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

PAPER WEEK

BIOFOR

CANADA

International Montréal 2016

2016

KEYNOTE SPEAKERS

Interview with Kim Rudd,

Parliamentary Secretary

Interview with

Sébastien Corbeil, President & CEO CelluForce

Parliamentary Secretary Rudd on Supporting Canadian Forestry

New Start for CelluForce

In an interview with Paper Advance during PaperWeek 2016, Kim Rudd, Parliamentary Secretary to the Minister of Natural Resources, outlined her government’s vision for supporting Canada’s forestry sector. Key to this vision is acceptance of the fact that economic growth and environmental responsibility work in cohesion, Rudd says. ‘We recognize we need to grow our economy and we see the forest sector as having a huge role in this.’ Pillars of the government’s strategy also include assistance for the sector in its efforts to move into clean technology, partnering with science, in particular the Ministry of Natural Resources, to ensure new processes are backed up by solid data, the promotion of Canadian exports through proposed tax credits for investment in clean technology and the development of new trade opportunities to assist companies in their efforts to move their products to a broader range of markets, and significant investment in training and education opportunities for Indigenous people living in communities where forestry companies/processes are based.

President and CEO of CelluForce Inc. Sebastien Corbeil met with Paper Advance during PaperWeek 2016 to discuss his company’s growing pains and return from what he refers to as the “valley of death.” While adaptability was integral to the company’s return to success, after a disappointing start for its newly constructed plant, the company ran out of money in 2013 and was forced to seek and secure a partner. In conjunction with NCC, who became a shareholder in the company, CelluForce applied to STDC for funding and revamped its strategy to focus on six major applications (oil and gas, pulp and paper, adhesives, cement, paints and coatings and resins). Within approximately three years, the company hopes to be self-funding, and expects to commercialize several new applications this year.

Interview with

Rod Albers,

Interview with

Bruno Marcoccia, Director of R&D Domtar

Domtar Casts a Wide Net Domtar’s Director of Research and Development Bruno Marcoccia discussed his company’s new portfolio approach to product development during PaperWeek 2016, explaining that diversifying its product base allows for a certain degree of risk mitigation. While the company has moved recently towards consumer and bio-products, these continue to be arenas Domtar is cautiously, though enthusiastically, exploring. “To this point we haven’t gone in a single, hard direction because the value proposition hasn’t warranted it,” Macoccia said. “So I would underline that in order to get from product development and demonstration to full commercial deployment, we have quite a lot of work ahead of us.”

J-FOR

Manager - Energy & Bioproduct Dev. West Fraser

West Fraser Seeks Opportunities For Rod Albers, Manager of Bioproduct Development for West Fraser, the secular decline in printing and writing grades has pushed his company to take a strong look at what opportunities exist for its feedstock. In discussing with PaperAdvance during PaperWeek 2016, Albers noted that energy generation and lignin extraction are two such opportunities, and with the support of FPInnovations, West Fraser is working on a new venture in the latter. “We have gone through the building and constructing phases, we’ve run water through the plant, so we know the technology for moving water around is there. We will be adding liquor this year.” Albers also noted the company will be its own “ﬁrst customer” with plans to consume roughly 10-20% of its own feedstock, but that the search is on for new partners to develop applications with.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Live keynote interviews on PaperWeek Canada YouTube Channel

Interview with

Robert Graham, CEO, Ensyn

Interview with

Suzanne Blanchet,

Senior VP Corporate Development Cascades

Ensyn's drop in fuels on the cusp

Integration & Innovations at Cascades

Suzanne Blanchet, Senior Vice President of Corporate Development for Cascades Inc., outlined the company’s rise from a Lemaire family waste collection business in 1964 to today’s multi- national paper company within $3.7 billion sales.

Interview with

Gurminder Minhas,

Managing Director, Performance Biofilaments

Interview with

Tim Cork,

Straight A’s Inc.

Cellulose ﬁlaments oﬀer renforcing advantages

Great coaches are givers

Tim Cork of Straight A’s Inc. is a motivational speaker who coaches people and organizations to “be the best you can be and the best for your company.”

More live interviews...

Jean Hamel, FPInnovations

Glenn Mason, NRCan

JF Levasseur, NRCan

... on PaperWeek

Paper Advance

J-FOR

Eric Ashby, Domtar

Channel

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

BIOFOR International Montréal 2016

HIGHLIGHTS

Bio-products: the tip of the iceberg at Canadian conference

It may be a technology of the future, however, its present just as exciting, as attendees were told how close it is to commercialization, in some cases. Others will take time.

The Pulp and Paper Technical Association of Canada (PAPTAC) made a significant and bold change in course in February this year when its annual meeting in Montreal featured the BIOFOR International bio-product and bio-refining conference in addition to the traditional PaperWeek Canada pulp and paper technical agenda. To put these diverse conferences on an equal footing is a sign of the times as bio-products and bio-refining are regarded as the fiber industry’s future. It may be a technology of the future, however, its present just as exciting, as conference attendees were told how close it is to commercialization in some cases. There were a number of important announcements about upcoming trials and some real results. Of course, there is much work to be done and the time lines may be long, as emphasized by some presenters.

“We are looking for bolt-on processes and near-commercial technology that add value and synergy to our existing processes.” he adds. Current initiatives include electricity generation sold to the grid from Rankin cycle heat recovery and bio-methane projects, activated carbon and lignin-based products from FPInnovations’ LignoForce lignin extraction process. That 30 t/d scaled-up process is now being commissioned and going through “water runs” at the Hinton, Alberta kraft pulp mill. West Fraser plans to use the product initially as a plywood adhesive, but other external applications are possible.

BIOFOR panel discussed bio-product successes and initiatives.

Conference news highlights: • CelluForce’s nanocrystals will be tested this year in drilling fluids and in papermachine chemistry applications. • Ensyn’s biomass oil is replacing Robert heating oil and Dufresne Greg Hay awaits approvals as a refinery feedstock. • Kruger’s FiloCell cellulose filaments in publication papers reduced printing press breaks in trials. • Enerlab’s polyurethane foam containing lignin will be included in insulation products this year. • The LignoForce 30 t/d lignin plant is in “water runs” at a West Fraser mill. The output will be used first in plywood.

Upcoming trials, demonstrations The conference’s kickoff session was a panel discussion of various and diverse bio-refining initiatives in the Canadian industry. Rod Albers of West Fraser Timber, says, “The key to a profitable bio-product is that it has to have a strong wood products business as its base.” That is wood products from West Fraser’s sawmills.

J-FOR

Gurminder Minhas of Performance BioFilaments, a joint venture of Resolute Forest Products and Mercer International, described the application development of their cellulose filaments (CF) now being produced at a 5 t/d demonstration plant at Kruger’s Trois Rivières mill. Through functional modification he says that CF is a “very adaptable product” with applications in fiberreinforced plastic composites, concrete, paints and coatings. Automobile parts are a particular application that he singles out as auto manufacturers are looking at lightweighting their vehicles to boost gasoline mileage. Robert Graham of Ensyn presented his company’s now commercial Renewable Fuel Oil (RFO) process which turns residual forest biomass into a liquid bio-hydrocarbon using its patented rapid thermal processing technology. He describes the process as being “very simple, with no catalysts or hydrogen.” The liquid product, now being produced at its 3 million gal/yr Renfrew, Ontario plant, is sold now as a substitute for petroleum-based heating fuel oil and is being demonstrated as a co-processing feedstock at oil refineries for producing gasoline and diesel fuels. A hospital in New Hampshire has now converted entirely to RFO, thereby replacing traditional fuel oil. RFO’s use as a potential refinery drop-in fuel is awaiting USA approval for toxicity and emissions testing. The fuel market is huge according to Graham, even with a modest 1% to 2% substitution in refinery feedstock.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Paper Advance

Bruno Marcoccia of Domtar described the company’s “bio portfolio” approach in which various initiatives with diﬀerent investment potentials, time lines and risks are being evaluated. The company is transforming itself from a primary pulp and paper producers into an innovative ﬁber company. Bio-products may be become “a new pillar” in their operations. Domtar has invested in demonstration plants for cellulose nanocrystals (CNC) at its Windsor, Québec mill and a LignoBoost lignin extraction plant at the Plymouth, NC mill. Other recent and ongoing projects include “ultra-high performance” softwood kraft pulp at its Dryden, ON mill and dried lignin powders at its Windsor mill that could be used in thermoplastic compounding.

A focused CelluForce Later in the week Sébastien Corbeil, CelluForce’s President and CEO, put to rest the industry’s curiosity about what has been happening with the fledgling biotechnology company when he announced that significant commercial trials using the company’s cellulose nanocrystals will happen later this year. It has not been easy for the company, as it faced a critical cash flow wall in 2013 and was facing the dreaded “valley of death” encountered by many start-up companies that miss the mark. Then along came a white knight investor in Schlumberger, the world’s largest provider of oil and gas drilling services, who saw the economic potential of CNC as a viscosity modifier for drilling fluids. The companies have formed a consortium to develop this application and test it later this year. Papermachine wet end chemistry is another application of CNC that will be tested online later this year with the partnership of a chemical supplier.

Cellulose filaments reduce press breaks Balázs Tolnai of Kruger updated the trial performance of its FiloCell trademarked cellulose filaments in papermaking applications. To date, there have been 30 trials in publication papers, specialty grades and tissue. CF’s strength enhancing and lightweighting potential were demonstrated. Tolnai says that CF provides much more than better average strength as the uniformity of strength has improved, resulting in a sheet that is less prone to breaking at the weakest points. For a pressroom trial 2000 rolls of CF-containing publication paper were produced on a production machine. The results in the pressroom were astoundings press break frequency was reduced by 60% at the same average sheet strength. CF has also been tested as a concrete additive at the University of Sherbrooke. At low dosages the flexural strength of concrete doubled. CF has shown its capability to modify liquid cement rheology so it is easier to pump to the top of high buildings.

Kruger’s Balázs Tolnai.

Claire Jahier of the Université de Québec à Trois Rivières presented studies showing the performance of CF in polyethylene (LDPE) composites. Modiﬁed with a chemical agent to reduce moisture absorption, the composite material showed mechanical properties similar to glass ﬁbre composites, was lower density and had low water absorption.

Sébastien Corbeil of CelluForce.

Corbeil says that they have learned lessons about how long it takes to introduce new products, understanding customers’ risk tolerance and some strategic blocking by adversaries. He explains that understanding the target niche submarkets is key. “We have now adopted a leaner start-up model in line with customer expectations. We have trimmed down and focused on our tier 1 applications.” In addition to oil drilling and papemaking these include adhesives, cement, resins, rubbers and composites, CNC enhanced polymers, paints and coatings. With the potential of these applications CelluForce hopes to fully sell the plant capacity of 300 t/y by 2018.

J-FOR

Armand Langlois of Enerlab, made a strong case for using 20% to 25% iso-lignin, as they call it, as a component in polyurethane (PU) foams, thereby displacing petroleum-based polyols. Iso-lignin is produced by catalytically reacting dried lignin with isocyanate. Presently, the PU foams are produced on a pilot line, but production of foam panels is expected later this year. Lignin also offers a cost reduction potential, he says. Since the introduction of bio-products into non-traditional markets often has a long time line and is fraught with some risks, the conference agenda also included sessions on R&D efforts in Canada, engineering and project management of new technologies and project financing opportunities.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Canada’s forest produCts industry

GrowinG a Greener tomorrow A PArt OF the ClimAte ChAnge sOlutiOn

Did you know that the Canadian forest products industry has the best environmental reputation in the world? Learn more at FPAC.CA

@FPAC_APFC ·

/FPAC.APFC · #FutureOFFOrestry

BIOFOR International REVIEW

BIOFOR International Montréal 2016

BIOREFINERY

GOVERNMENT POLICY SESSION

Mariya Marinova

Glenn Mason

Biorefinery kickoff highlights Tuesday morning’s Biorefinery Track, presided over by Mariya Marinova of Polytechnique Montreal, kicked-off with several presentations on the progress of some bio-chemical product developments. Dorothea Stark of BASF emphasized that the chemical products from renewable biomass feedstocks are challenging to produce since the raw materials have multiple heterogeneous components. Their development requires the input and collaboration of a variety of multidisciplinary partners with unique expertise. She described recent BAST developments from bio sources, including succinic acid, renewable BDO (1,4 Butanediol) and FDCA, a monomer for polyesters used in bottles currently using PET and in barrier films. Dr Kouisni of FPInnovations describe the low sulfur compound emissions of the 30 t/d lignin extraction process now being commissioned at a West Fraser kraft pulp mill in Hinton, AB. The process also has advantages in chemical consumption, high filtration rates, and high lignin purity. Dr. Zhang of Lakehead University described a laboratory study converting TMP lignin into a focculant which can be used to remove dyes in the effluent of paper and textile mills, for instance. The COD removal effectiveness could ne over 70%.

Paper Advance

J-FOR

Government support, policies for bio-economy The bioreﬁnery track continued on Wednesday morning with a panel discussion on federal and provincial funding and industry support programs for the emerging forest bio-economy. Ontario’s representative Kathleen McFadden stressed the need for alignment of various departments including energy, environment and natural resources since the new bio-economy touches on all. She hinted that a government-wide policy may be required. Ontario’s CRIBE funding agency, Jobs and Prosperity Fund for the forest sector and Forestry Growth Fund were singled out for their contributions. Ontario is in the process of developing a roadmap for addressing forest industry costs, value of products and aboriginal people’s engagement in the industry. Quebec has contributed to the development of wood construction, energy from biomass and green chemicals. The latter category includes 8 projects involving 20 companies outside the forest products industry.

Glenn Mason of Natural Resources Canada sees that the forest sector has a bright future as, “Canada’s forestry practices are second to none on a world scale and you can make anything from a tree that you can make from a barrel of oil”, he says. He highlights IFIT as a successful program that supports ﬁrst-in technology. The role of the government in changing building codes to allow taller wood frame buildings was highlighted as a successful model that has attracted international interest and provided a market pull for Canadian lumber products.

ENGINEERING SESSION Engineering to commercialize at minimum risk In Wednesday afternoon’s BIOFOR engineering track a panel experts convened to discuss how to bring new processes up to speed and produce the expected results, Risk mitigation was a common theme. Guy Martin of KSH discussed some of the positive outcomes and pitfalls of commercialization and how to avoid them. Jim Wearing of NORAM described his company’s track record bringing pilot scale processes to their full commercial potential. Their portfolio includes the LignoForce process at West Fraser’s Hinton mill. Jim Burgess of API described his company’s track record bringing its AVAP biomass sugar extraction process online. Finally, Virginie Chambot of Invertis described business models and risk management strategies to bring new bioreﬁning processes to market.

Nova Scotia has established an Innovations Hub, combining companies and organizations involved in energy, research and development, and economic development. Projects include a biomass to fuel plant at a former paper mill site, dynamic modeling to reduce ﬁbre costs, substitution of biofuels in building heating. Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Creating business opportunities for companies that see things differently. Helping our members stay ahead of the pack! Next-geNeratioN fibre and paper products.today. from fibre supply to innovative products FPInnovations is a world leader in the development of new products and applications based on nextgeneration fibre, papers and bioproducts. We rely on a team of more than 500 innovators to create new science-based solutions that will increase manufacturing efficiency and reduce operating costs. fpinnovations.ca Follow us

BIOFOR International REVIEW

BIOFOR International Montréal 2016

Federal R&D Activities and Services for Forest Bioreﬁnery Development in Canada

Jean Hamel

R&D spurs BIOFOR transformation On Wednesday afternoon R&D representatives of CANMET Energy, NRC and FPInnovations convened to discuss R&D efforts to promote the bio-product transformation of the industry. Fernando Preto of CANMET Energy presented programs to promote cleaner, more energy efficient biomass conversion. The combustion of low quality fuels like agricultural residue, sludge and bush slash are being investigated for maximum performance in hog fuel boilers. Pyrolysis oils can be a source of high value chemicals. Eric Soucy of CANMET Energy presented a process data analysis and optimization program which identifies process inefficiencies and solutions to reduce energy input. Key process indicators and predictive models are derived. The program can be used to assess the impact of pulp mill biorefinery integration to maximize revenue streams. Minh-Tan Ton-That of NRC discussed research on new bio-product including polymers from lignin that can replace conventional thermoplastics, bio-composites, and emission reduction solution. Minh-Tan Ton-That explained the purpose of their work is to bridge the gap between new products and consumers and to add value to biomass.

J-FOR

Jean Hamel of FPInnovations described the various business models that FPI have promoted to advance bio-product development. These include joint ventures, strategic alliances, licensing and technical support, technology scale-up and international collaborations for safety standards and environmental compliance. FPI is opening up new business opportunities for oil and gas drilling, plastics, building materials, composites and resins. FPI is also fostering new product application development and developing expertise and knowledge for new markets.

paper were produced on a production machine. The results in the pressroom were astounding as press break frequency was reduced by 60% at the same average sheet strength.

BIOMATERIALS SESSION

Next, Claire Jahier of UQTR presented studies showing the performance of CF in polyethylene (LDPE) composites. CF offers high surface areas and reinforcement potential and could be a viable replacement for glass fibre in FRP. Modified with a chemical agent to reduce moisture absorption, the composite material showed mechanical properties similar to glass fibre composites, was lower density and had low water absorption.

Promising CF and lignin applications

The third presenter, Armand Langlois of Enerlab, made a strong case for using 20% to 25% Iso-lignin as a component in polyurethane (PU) foams, thereby displacing petroleum-based polyols. Iso-lignin is produced by catalytically reacting dried lignin with isocyanate. Presently the PU foams are produced on a pilot line, but production of foam panels is expected later this year. Langlois says the potential for PU foams is 15 million tonnes/yr in which 200,000 to 300,000 tonnes/y of lignin could be used. Lignin also oﬀers a cost reduction potential, he says.

At the Thursday afternoon Biorefinery track NRCan’s Jean Francois Levasseur presided over a very informative session that highlighted very promising applications of cellulose filaments (CF) and lignin. The market potential of theses products is getting closer and more well defined. First, Balázs Tolnai of Kruger updated the trial performance of CF in papermaking applications. To date, there have been 30 trials in publication papers, specialty grades and tissue. CF has demonstrated the capability to improve both dry and wet sheet strength. Tolnai says that it is much more than better average strength as the uniformity of strength has improved, resulting in sheet that is less prone to breaking at the weakest points. For a pressroom trial 2000 rolls of CF-containing publication

CF has also been tested as a concrete additive at the University of Sherbrooke. At low dosages the ﬂexural strength of concrete doubled. CF has shown its capability to modify liquid cement rheology so it is easier to pump to the top of high buildings. Once the pumping stops the cement thickens.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

BUSINESS LUNCHEONS

The Business Luncheons are one of the most popular draws of the event. Their unique format features a keynote presentation each day of the conference. Incorporated to the program schedule and included with the registration - they are a true complement of the conference and tradeshow and serve as a great platform for the advancement and gathering of the industry. Tuesday

Wednesday

Motivation to discover yourself and improve teamwork

Resource development is a nation building exercise

Tuesday’s luncheon keynote was given by Tim Cork of Straight A’s Inc. This motivational speaker coaches people and organizations to “be the best you can be and the best for your company.” “Change the me to we and invite others to join you,” he adds. It’s an important part of team building.

Kim Rudd, Parliamentary Secretary on behalf of the Honourable Jim Carr, Canada’s Minister of Natural Resources, gave a positive expression of government policy for the forest industry to a fullhouse luncheon crowd on Wednesday. Her statement “Resource development is viewed as a nation building exercise,” obviously drew favourable feelings. Moreover, she linked resource development and its economic beneﬁts to a cleaner environment and mitigation of climate change. That’s very big picture thinking. “Climate change is the biggest challenge of our generation. Canada’s forest industry is part of the climate change strategy. Ten percent of our Paris commitment to greenhouse gas reductions will come from the forest products industry,” she says.

He encourages the process of self awareness and discovery and to, “love what you are doing and be genuine, since it is contagious.” He stresses that collaboration is essential since people work better as a group.

Tim Cork

Other quotes that people would take home and consider included: “Life is like an echo, it comes back to you” “Listen more than you speak” “The person that asks great questions and then listens, controls the situation” “If you want to be successful, help others to be successful”

Kim Rudd

Continuing on, she says that the government will support innovation and clean technology, with a greater emphasis on scientiﬁc developments. “We will increase our eﬀorts for forest product markets at home and abroad and eliminate barriers to international trade,” she adds.

Food for thought.

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

PAPER WEEK

CANADA

2016

Within our social network she says that, “The engagement within indigenous people will be enhanced to improve their skills and job prospects.” Good thoughts for an industry that already was feeling good about its future. Thursday

Consumer shopping habits: Where are they heading? At Thursday’s luncheon Carman Allison of the AC Nielsen Company gave an entertaining keynote on where consumer shopping habits are going and what are the best bets for growth. How do we spend and where do we shop are critical trends that retailers should know. Right now consumers in Canada are not as optimistic and they once were in better economic times and are lagging behind their American counterparts. Worries over job security, economic uncertainty, high debt ratios and regional recession in the west are factors. Are we in a recession? According to Nielsen polls seven out of ten Canadians think so and thirty per cent of us have adjusted our buying habits to stock up on only the basics. Allison says that shoppers are more value conscious and, as a result, discount stores are flourishing. But there is a polarization as premium stores are doing well and those in the middle are having trouble. Retailing has been redefined. Online shopping, dollar stores and ethnic stores in some areas are doing well. There is less cross bordershopping for obvious reasons. Bathroom tissue is up a bit, keeping pace with population growth. Tissue buyers are value conscious so they are buying in bulk to get the best price, but large purchases do not mean more consumption. He says that twenty percent of tissue products are bought in warehouse stores. Groceries overall are down three percent in the last five years but fresh prepared foods and produce at the store perimeters are doing well. The millennial generation is quickly becoming the engine for consumer product growth as they take over from baby boomers who are now consuming less. Urbanization also has changed the retail landscape as city dwellers, some without cars, are making more shopping trips but buying in smaller quantities. Interesting facts for retailers and product marketers.

Paper Advance

J-FOR

Carman Allison

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

AUTOMATIC SELECTION OF RELEVANT DATA FOR PAPER MACHINE DIAGNOSTICS ABSTRACT

HELI HYTTI, ANTTI NISSINEN, MIKKO LAURI, HANNU KOIVISTO, HEIMO IHALAINEN, RISTO RITALA*

This article presents a method to focus the attention of process operators and experts on the most relevant data in a large data base when searching for the root cause of a detected process disturbance. The method is generic in that it analyzes control loops individually based on their signal triplets of set-point, measurement, and control output. The model of normal operation is a Gaussian mixture model (GMM) for a two-dimensional feature vector identified at regular intervals, e.g., once per day for a two-week period. When a process disturbance occurs, the relevant control loops are chosen based on their poor correspondence to the normal operating model. Data for hundreds of control loops can be reduced to data for 10–30 relevant loops, thus focussing the analysis significantly. The efficiency of GMM modeling methods is discussed. The method is demonstrated with real-life paper production data involving hundreds of control loops.

INTRODUCTION

A paper manufacturing process is operated non-stop, except for planned maintenance breaks and unplanned breaks caused by process disturbances. In paper manufacturing, capital costs form a significant share of total production costs. Hence production assets must be operated efficiently: whenever disturbances occur, their root cause must be identified rapidly. Performance monitoring and diagnostic tools are among the key elements that the plant owner expects from the plant supplier. Process data bases contain the information required for efficient diagnostics, but the challenge is to find the relevant signals among the huge amount of data stored. At present, the process data bases contain

data from essentially all the control loops. Controller status (manual/automatic), setpoints, measured values of the controlled process variables, and controller outputs are logged. At a modern paper mill, up to 50,000 signals can be collected to data bases at intervals of, say, 5 seconds. Of these, 2000–3000 signals are needed for performance monitoring. When data are saved to a data base, it is processed in real time, and its basic statistical characteristics are calculated. Special operating events—start-ups, shutdowns, breaks, and known disturbances— are detected from the data for annotation. The rest of the operating time represents normal operating conditions. Variables critical for good operation

are end-product quality, e.g., dry basis weight or total ash content, and material streams at junction points or entering mixing volumes. During normal operation, the variability in critical variables is low, whereas all abnormalities in them are explainable by deviations from normal operation. This article reports methods that detect which control loops are abnormal during the time frame for abnormality in some critical variable. Thus, analysts are directed to focus on certain control loops and time intervals to seek the root cause of an abnormality in the critical variable. Finding the actual root cause is based on the analyst’s knowledge of the topological structure of the process and on further manual data analysis. In particular, process

HELI HYTTI

ANTTI NISSINEN

MIKKO LAURI

HANNU KOIVISTO

HEIMO IHALAINEN

RISTO RITALA

Tampere University of Technology, Tampere, Finland

Process Automation Systems, Valmet Oyj, Tampere, Finland

Tampere University of Technology, Tampere, Finland

Tampere University of Technology, Tampere, Finland *Contact: risto.ritala@tut.fi

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS experts are good at finding root causes simply by scanning the relevant signals visually for features such as level changes, ramps, changes in oscillation frequency, changes in the shape of oscillation, and changes in variance. Thus, the methods presented in this article speed up the analyst’s work by reducing the number of signals and time intervals to be scanned and analyzed. In industrial process monitoring, abnormal behaviour is identified by first generating a statistical model for normal behaviour and then studying the exceptionality of the current data with respect to this model. For an overview of anomaly detection techniques spanning multiple application domains, the reader is referred to, e.g., [1–5]. Many data-driven methods have been used to construct the normal state for process monitoring [6–8]. Principal components analysis (PCA) [9] and partial least squares (PLS) [10] methods with T2 and squared prediction error statistics have been commonly used. Standard PCA and PLS assume a linear, stationary system with Gaussian distributed data. However, dynamics and non-linearity are important aspects of industrial processes. Furthermore, these processes often run under several normal operating conditions, and hence the normal state model is multimodal. PCA and PLA algorithms have been developed to tackle the dynamics [11–15]. Linear time-series methods have been applied to dynamic fault detection problems [16–19]. Kernel-based PCA and PLS [20–23] and support vector machines [24–25] have been intensively studied for non-linear processes. Fisher discriminant analysis (FDA) and independent components analysis (ICA) [26,27] are multivariate statistical techniques for industrial data. FDA, assuming Gaussian data, can be applied for fault classification and detection [28–31]. ICA extends fault detection to nonGaussian data [32–34]. However, the ICA negentropy statistic does not necessarily reflect the multimodality of the process data. This may degrade the performance of ICA if the normal state is multimodal [35].

J-FOR

Gaussian mixture models (GMMs) decompose the data into multiple Gaussian clusters to represent the multimodality of the normal state [35]. Gaussian mixture modelling [26,36] has been combined with PCA [37] and Bayesian inference [35,38] for non-Gaussian process monitoring [35]. Among these methods, PCA is the fastest to compute, whereas ICA is the most complex of the algorithms [7]. This paper is organized as follows. First the proposed data selection procedure, which consists of a modelling and an analysis stage, is presented. Normal state is represented as a GMM. The applicability of GMM methods for the purposes of this study is then analyzed. Two examples with real data are presented, and finally concluding remarks are provided. DATA SELECTION PROCEDURE

The goal of this study was to develop a method for selecting relevant data and time intervals that would be generic to continuous processes and applicable online to large data bases. Therefore, it is sufficient to detect abnormalities in variables automatically rather than studying their interdependences: the abnormal variables are candidates for root cause analysis. The hypothesis proposed here is that process disturbances are reflected as abnormal operation of control loops. In all processes, a control loop has a signal triplet: set-point, measurement, and control action. These signals are easily found in the tag lists of the automation system: the loop is identified by a unique name, and the signal with a supplement, e.g., ‘spa’, ‘me’, or ‘con’. Control loops fall into four mode categories: 1) ones that merely provide a user interface for actuators, 2) independently operating control loops in manual mode, 3) loops in automatic mode having a local set-point, and 4) cascaded loops following a set-point from an external source, for instance quality controls. The interface mode is recognizable in that both set-point and measurement are constant except in transients, the manual mode in that only the measurement varies over

time, and the independent loop in that the set-point is constant while the other two signals vary. In cascade control, all three signals vary continuously. In an industrial process, the controlled variables (e.g., flow, level, power, consistency, and pressure) have different physical units, ranges, and characteristic waveforms. For the purpose of classifying control loops as normal or abnormal, the instantaneous values of the control-loop signals are less relevant than their statistical interdependencies. Hence, a control loop can be characterized by sliding timewindowed statistical features computed from the signals. The length of the sliding window is adjusted based on which dynamics the characteristics should reflect. The requirement for automatically selecting the relevant loops from large data bases is that the number of loop features must be kept small. In this study, only two features are computed for manual, independent, and cascaded loops: the sliding mean ∆(M) and the sliding relative standard deviation σrel(M) of the difference between the set-points and the measured value x. This choice is simple, effective, and fast to compute as a convolution: (1) (2)

(3)

where the sliding window M = 15 min was found to reflect the process dynamics appropriately. Any changes in interface loops were recorded by detecting the change in mean. The standard deviation of the difference is scaled relative to the mean set-point because a higher set-point commonly results in a larger variation in the measurement. Obviously, if the set-point is given relative to a variable, such as pressure difference, scaling by the set-point should not be used. The detection method for relevant

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

control loops must allow multimodal normal states: the normal state model consists of clusters, so that in the normal state, the feature vector is close to one of the cluster centroids, whereas in abnormal states, the feature vector is far away from all centroids. Transitions between centroids also give valuable information. Clustering is unsupervised, and computation for detection—comparing distances to centroids— is fast. However, clustering performance hinges critically on how effectively the model cluster structure captures the features of the normal state. Furthermore, constructing the cluster model is often computationally complex. GMMs were chosen in this research as normal state models for each control loop. A GMM probability density function (PDF) is a weighted sum of Gaussians. The GMM method can be applied to nonnormalized data. This helps to retain the intrinsic character of normal control-loop operation, in particular that of changing set-points and control functions and their effect on statistical features. GMM model parameters can be estimated efficiently, and classification into normal and abnormal states is simple. The expectation-maximization (EM) algorithm [36,39] finds maximum likelihood solutions for GMM parameters. As a starting point for the data selection method, it has been assumed that the data base pre-processes the data in the background, computes and stores the control-loop feature vectors, and annotates time intervals of special process events so that they can be excluded when constructing the model for the normal state, as done in [40]. The procedure for selecting relevant variables and time intervals for the expert consists of a modelling and an analysis stage. The modelling stage, also running as a background process, constructs the control-loop GMMs based on normal state data. The modelling stage is repeated at specified intervals (e.g., once per day with the normal state data from the previous two weeks) to adapt the models to normal process variation , e.g., the grade spectrum. The analysis stage operates as follows:

J-FOR

A1. Analysis is triggered by a disturbance in a critical variable, for instance, detected by operators. A2. For each loop, the membership of feature vectors to components of the current GMM is computed over a recent history of given length. A3. If the sum of component memberships of a loop is below a threshold for a given period of time, the loop is labelled as abnormal. A4. The process expert is provided with the list of abnormal loops for manual root cause analysis. The memberships of the most relevant control loops are visualized with time/colour graphs. The feature vectors and component memberships can also be computed as a background process, thus providing the list of abnormal loops quickly when a disturbance in critical variables is observed. IMPLEMENTATION OF GAUSSIAN MIXTURE MODELLING Model Structure, ExpectationMaximization Algorithms, and Abnormality Detection

A Gaussian mixture model of order K for feature vector z is a weighted sum of K Gaussians N : (4) The model is parameterized by the means μk, co-variances Σk, and mixing coefficients πk. Given a

set of observed features {zn }N , the GMM n =1

model determines model parameters that maximize the likelihood of the data. The expectation-maximization (EM) algorithm for finding the parameter values is as given in [36], pp. 438–439. After initialization, the expectation (E) step for updating the mixing coefficients and the maximization (M) step for updating the means and covariances alternate. The model fit after the ith iteration is assessed by the log-likelihood function: (5)

The alternation of EM steps continues until a pre-defined convergence criterion for the change in log likelihood is met. The EM algorithm is initialized by guesses for the model parameters. Random initialization generally leads to a long iteration process, and model parameters will vary from initialization to initialization. The log-likelihood function may have multiple local maxima, and hence EM is not guaranteed to converge towards the global maximum. The K-means clustering algorithm, which is much faster than the EM algorithm, is used to find a suitable initialization for the GMM parameters. However, with large data sets, K-means initialization increases the computational burden. In the present case, the data are of low dimension, but the number of observations is large. In such a case, GMM modelling is speeded up by using the binned-EM algorithm [41,42]. The main difference between binned and unbinned EM is that the latter replaces individual data values by bin nominal values. As a result, the summations over individual data values of unbinned EM turn into weighted sums over bins. If the number of bins is much less than the number of individual data values, but the data points are dense enough that the probability densities within a bin can safely be assumed constant, the computing time is reduced by a factor roughly proportional to the ratio of the number of bins to the number of data values. Furthermore, a histogram-based, deterministic initialization of the GMM parameters including model order is here proposed as follows: 1. Approximate the probability density function of the data as a histogram in which the data on each dimension D are divided into L bins of equal width. The proportion of observations represents the probability density in that bin. 2. Find the bin, B, containing the largest number of observations. Select a cube of bins around B, of size RD, where R is determined according to

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS the number of bins per dimension and the expected size of the seed clusters. The mean of the Gaussian component is the mean of the cube, the covariance is estimated by the cube covariance, and the mixing coefficient is the proportion of observations in the cube to the total number of observations. 3. Mark the selected group as excluded from subsequent steps. 4. Repeat step 2 for the remaining part of the histogram until the whole data set is covered. The number of GMM components is obtained by iteration. Bhattacharyya distances between the Gaussian components are computed, and components closer to one another than a threshold DB are combined [43]. The Bhattacharyya distance between two multivariate Gaussians with means (μp, μq) and co-variances (Σp, Σq) is:

Figueiredo and Jain [44] selects the number of components by starting from a given maximum number of components and eliminating through iteration the components not supported by data. With a large number of initial components, this algorithm is less sensitive to initialization, but tends to be slow. The unbinned standard EM algorithm by Chen (see http:// www.mathworks.com/matlabcentral/ fileexchange/26184-em-algorithm-forgaussian-mixture-model) with K-means initialization from Netlab Toolbox is the other reference. Chen’s algorithm also finds the optimal number of components by elimination. Three two-dimensional data examples were considered. The first example was a synthetic sample of GMM data with six clearly distinct components, sample size 106 (Fig. 1a). The second example

consisted of synthetic GMM data with three components, of which two overlapped strongly (had the same means), sample size 106 (Fig. 1b). The third data set consisted of real process data with 993,159 observations and is illustrated in Fig. 2. GMMs for all three datasets were identified using all three methods. The Kullback-Leibler (KL) divergence (Bishop, 2006), (8) between the true PDF q and the PDF estimated by GMM p was used to measure goodness of fit. With real process data, the “true” PDF was constructed as a GMM of 30 components using the binned algorithm with 1000 bins per dimension, as shown in Fig 2b.

(6)

Given a GMM model, the membership Mn,k of a control-loop feature vector zn in a model component k is defined as: (7) The feature vector is abnormal with respect to component k if the membership is below a threshold, here set to be 0.001. The feature vector is abnormal with respect to a GMM model if the sum of memberships is below a threshold. Transitions between components of normal states can be studied through individual component memberships.

Fig. 1 - Probability density functions of example data sets 1 and 2: a) example data set 1, b) example data set 2.

Comparison of Unbinned and Binned EM Algorithms

The binned EM was implemented and compared on synthetic and real data with two unbinned EM algorithms available for MATLAB. The EM method proposed by

J-FOR

Fig. 2 - a) Histogram of example data set 3 (500 bins per dimension); b) reference probability density function computed using a histogram with 1000 bins per dimension and 30 GMM components.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TABLE 1

Computation time and optimal number of components for the three algorithms. Sample size 106.

Data

Elapsed time [s] Proposed, 100x100 bins 0.8 0.3

Example 1 Example 2

Chen

Figueiredo

16 21

404 184

KL divergence [nats] of the estimated PDFs from the true PDF. Sample size 106.

TABLE 2

Example 1 Example 2

Proposed, 100x100 bins 2.7×10-3 3.2×10-5

For examples 1 and 2, a GMM model was identified for one data set sampled from the PDFs. The histogram-based initialization is deterministic, resulting always in the same PDF, whereas the K-means algorithm used by the reference methods initializes randomly, and hence the GMM models for a data set vary. Therefore, identification with the reference methods was repeated ten times for each data set and the results averaged. Tables 1–5 provide the main results.

TABLE 3

Final number of components Proposed, 100x100 bins 6 3

Chen

Figueiredo

131.6×10-3 1.1×10-5

0.02×10-3 2.0×10-5

Figueiredo’s algorithm gave the best model for example 1. The KL divergence between the true and estimated PDF was 2×10-5 nats. With 1002 bins and the Bhattacharyya threshold DB = 0.1, the binned EM resulted in a KL divergence of 3×10-3 nats, whereas with 4002 bins, the value was 2×10-5 nats. For 1002 bins, binned EM used 1/500 and for 4002 bins 1/40 of the time spent by Figueiredo’s algorithm. Chen’s algorithm gave in all respects the worst results for these data.

KL divergence and computation time as functions of the number of bins; example 1, sample size 106.

Number of bins 502 1002 1502 2002 2502 3002 3502 4002

KL divergences from the real PDF [nats] 24.8421×10-3 2.6908×10-3 0.6495×10-3 0.2242×10-3 0.1172×10-3 0.0719×10-3 0.0484×10-3 0.0374×10-3

Time elapsed [s] 0.5 0.8 1.4 2.1 3.0 5.1 7.1 10.2

Chen

Figueiredo

3–6 3–5

6–9 3

For example 2, the PDF estimates resulting from all three methods were similar, and hence the KL divergences were of the same order of magnitude. The Bhattacharyya threshold in the binned algorithm was DB = 0.2. The binned method took 0.3 s, Chen’s 21 s, and Figueiredo’s algorithm 184 s to compute. The unbinned methods performed poorly with example 3. The model was tested with the maximum number of clusters set to 2, 4, and 6, as shown in Table 4. Chen’s algorithm always resulted in one or two clusters with poor results. Figueiredo’s algorithm was not able to capture the shape of the probability density function, as shown in Table 5 and Fig. 3. The binned algorithm with Bhattacharyya threshold DB = 0.2 gave fairly good results. It can be concluded that for large data sets of low dimensionality, the binned EM algorithm is significantly more efficient computationally than the unbinned EM. The modelling accuracy depends on the number of bins per data dimension. For two-dimensional feature vectors, the case relevant to this work, the number of bins can be set to 4002, which gives a good combination of model accuracy and computation time. As the data dimensionality increases, the binned method quickly loses its advantage because the number of bins increases exponentially with the number of dimensions. In this study, the binned EM algorithm was used for the process data.

TABLE 4 Computation time and optimal number of components for the three algorithms, real data, example 3. Max. number of clusters 2 4 6

J-FOR

Time elapsed [s] Proposed, Chen 100x100 bins 0.5 11.7 1.6 2.9

Figueiredo 19.4 50.0 105.9

Final number of components Proposed, 100x100 bins 2 4 6

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Chen

Figueiredo

1–2 1–2 1–2

2 4 6

TRADITIONAL AREA CONTRIBUTIONS

Best Kullback-Leibler divergences [nats] of the TABLE 5 estimated probability density functions from the true probability density function, real data, example 3. Max. number of clusters 2 4 6

Proposed, 100x100 bins 0.0682 0.0210 0.0173

Chen

Figueiredo

0.5119

0.2443 0.1546 0.1433

will then visually compare the recent data with respect to the normal state and study the set-point, measurement, and control signals of each abnormal loop to locate the root cause with the help of his process expertise.

Fig. 3 - Contours of the best estimated probability density functions (in the sense of Kullback-Leibler divergence) compared with the 30-Gaussian reference probability density model: a) histogram-based method, 6 Gaussians, b) Chenâ&#x20AC;&#x2122;s method, 2 Gaussians, c) Figueiredoâ&#x20AC;&#x2122;s method, 6 Gaussians.

TESTS AND IMPLEMENTATION WITH REAL PAPER MACHINE DATA

GMMs provide the normal operating reference when diagnosing abnormalities in control-loop operation. If the current feature vector of a loop has a low membership in all Gaussian components, the loop and hence the unit process related to it is deemed abnormal. When a disturbance in a critical process variable is detected, the control loops and unit processes that are abnormal within a time window are candidates for root causes. With hundreds or even thousands of control loops to analyze, pruning normally behaving loops automatically speeds up the disturbance analysis, even though an operator or a process expert is needed to identify the root cause from the short list provided. Two examples of process disturbance analysis in a paper manufacturing process are described next. The mill has more than 2000 process signals logged to an on-line data base. Of these, 226 control-loop triplets were included in the analysis. The GMM models for

J-FOR

normal operation are constructed once per day using the normal state data from the previous two weeks. As new data become available and a disturbance is detected in one of the critical variables, analysis steps A1-4 are carried out for all loops (or computed as a background process), and the abnormalities are shortlisted for operators or process experts to examine. In both example scenarios, dry basis weight is the critical variable, and the course of actions is as follows: 1) Excessive variation in dry basis weight is detected at a time instant. 2) The current GMM models of all loops for normal operation are retrieved from the data base. 3) The loop data for the last two hours are retrieved, and their memberships to the GMM components of each loop are computed. If the sum of the memberships of a loop is less than 0.001 for at least 15 consecutive minutes, the loop is listed as abnormal. 4) The operator is provided with the list of abnormal loops. The operator

Case 1

The process area of interest in Case 1 is shown in Figure 4. Variation in dry basis weight exceeded a pre-set threshold at 04:45 (Fig. 5). The GMM analysis between 03:00 and 05:00 found that 19 of 226 loops were abnormal. The loop with the lowest membership sum was that for the de-aeration flow to the wire pit. Deaeration flow was much below its set-point for over half an hour (Figs. 6 and 7). This led to increased air content in the wire pit, which maintains a constant level due to overflow. The wire pit water is recirculated for final dilution of machine stock flow before the short circulation (Fig. 8). Because the air content increased and the density decreased in the wire pit, the hydraulic pressure at the bottom of the pit changed. Less dilution water was fed, the consistency of the stock increased, and the dry basis weight increased. Eventually, the basis weight control corrected the basis weight closer to its set-point, but a transient disturbance had been caused.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Case 2

Fig. 4 - Process area of interest in Case 1.

Fig. 5 - Dry basis weight between 03:00 and 06:30, case 1.

Fig. 7 - Set-point (SPA), measurement (ME), and control (CON) of the defoamer flow-control circuit. A significant change in flow is clearly visible around 04:30.

J-FOR

The process area of interest in Case 2 is shown in Figure 9. Variation in dry basis weight exceeded the pre-set threshold at 03:00 and increased further around 04:30 (Fig. 10). GMM membership analysis for 02:00–05:00 found 27 of 226 loops abnormal. Eight of the loops were related to starch processing and dosing in the process. In addition to paper strength properties, starch has a strong effect on fine-particle retention on the wire section. Starch is prepared in a cooking

Fig. 6 - Feature vector of the defoamer flow-control circuit compared with normal operation (clusters 1–3). The operation deviates radically from normal.

Fig. 8 - Feature vector of machine-stock flow-control loop compared with normal operation (clusters 1–3). The loop is normal (within cluster 1) until the defoamer flow problem affects it at 04:30.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS phase and transferred to the dosing tank. Pressure filters after the dosing tanks make sure that starch is uniform and clean. The

cooking process was interrupted between 02:45 and 3:15 and then continued normally (Figs. 11 and 12). However,

Fig. 9 - Process area of interest in case 2.

when the cooking process was re-started, some inconsistencies blocked one of the pressure filters. The blockage can be seen as a variation in starch dosing pressure (Figs. 13 and 14). The disturbance was resolved by stopping the starch flow for a short time and switching the filter. The memberships as functions of time can be visualized in Fig. 15, where the dominant GMM component of the 25 most abnormal control loops is shown for case 2. The eight top rows represent the starch processing and dosing control

Fig. 10 - Dry basis weight between 02:00 and 06:00, case 2.

Fig. 11 - Feature vector of starch flow to starch cooker control loop compared with normal operation (clusters 1â&#x20AC;&#x201C;7). The flow deviates significantly from normal most of the time after 02:45.

Fig. 12 - Set-point (SPA), measurement (ME), and control (CON) of starch flow to the starch cooker loop. The flow doubles at 02:45, is then turned off for a few minutes, and then returns gradually to normal.

Fig. 13 - Feature vector of starch dosing pressure compared with normal operation (cluster 1). The flow deviates significantly from normal between 02:30 and 03:45.

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

loops. The colours indicate the component with the highest membership. The brightness of the colour indicates the degree of membership in the component. Hence, a black colour indicates abnormality. Frequently changing colour indicates that the loop is nonâ&#x20AC;&#x201C;stationary, with transitions among the components. This visualization provides a graphical user interface that can help to choose loops for further analysis. In the live software, clicking on a point in the image shows the data for the control loop over a short time interval before and after that instant in time. DISCUSSION AND CONCLUSIONS

This paper has presented a tool for focussing attention when troubleshooting based on large industrial data bases such as those available for paper manufacturing processes. The need for troubleshooting is assumed to be triggered by monitoring a few critical process variables. Then the proposed algorithm finds the related abnormal control loops and relevant time periods for further inspection by a loopby-loop cluster membership analysis. A process operator or expert identifies the root cause based on this reduced data set. The examples showed that the analysis typically prunes 90% of the loops from

the analysis, thus reducing the effort required by the operators or experts. A GMM was chosen to represent normal states. Low membership in all GMM clusters indicates a deviation from normal operationâ&#x20AC;&#x201D;for example, changes in control operation mode, or reduced variation in signals that normally have large variation. The GMMs are updated regularly as a background process. The update frequency depends on the process dynamics. In papermaking, updating once per day with data over two weeks has been found appropriate to support problem periods of a few hours. In this research, the binned EM algorithm was used for GMM identification. It was demonstrated that for a given model accuracy, the binned algorithm is faster than unbinned ones when the feature data is two-dimensional. To implement the proposed method at full scale on-line is quite straightforward: modelling and analysis functions can be implemented as automated background processes in any modern data-base system. However, tailoring an efficient user interface for operators into an automation system may require a considerable amount of work and experimentation. The first version is expected therefore to be intended for an engineer whose

Fig. 14 - Set-point (SPA), measurement (ME), and control (CON) of the pressure in the starch dosing line control loop. Pressure control changes to manual operation around 02:45 when the cooking process was turned off.

J-FOR

task is to solve production problems and bottlenecks. The method can easily be adapted to other continuous production processes because conceptually it deals only with control loops and their signal triplets of set-point, measurement. and control action. The method is very easy to implement in current process data-base systems: automatic updating of the normal status models at regular intervals is a background process, and computing the memberships is fast. The models have a small number of parameters requiring little space in the data base: with a twodimensional feature vector, the model has 6K-1 parameters per loop, where K is the number of clusters, typically ranging from one to seven. Collective visualization of control-loop memberships by means of colour coding provides a quick overview of process state and a visual indication of abnormal loops. ACKNOWLEDGEMENTS

Financial support by TEKES as part of the Data to Intelligence program of DIGILE is gratefully acknowledged. Metso Process Automation Systems is acknowledged for providing the data.

Fig. 15 - Visualization of loop membership values. Colour indicates the dominant cluster and brightness the membership.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS

REFERENCES 1. 2.

9. 10.

11.

12.

Hodge, V.J. and Austin, J., “A Survey of Outlier Detection Methodologies”, Artif. Intell. Rev. 22:85-126 (2004). Isermann, R., Fault-Diagnosis Systems: An Introduction from Fault Detection to Fault Tolerance. Springer-Verlag, Berlin, Heidelberg, 2006. Patcha, A. and Park, J.-M., “An Overview of Anomaly Detection Techniques: Existing Solutions and Latest Technological Trends”, Computer Networks 51:3448-3470 (2007). Chandola, V., Banerjee, A., Kumar, V., “Anomaly Detection: A Survey”, ACM Computing Surveys 41(3): Article 15 (2009). Isermann, R., “Fault-Diagnosis Applications: Model-based Condition Monitoring: Actuators, Drives, Machinery, Plants, Sensors, and FaultTolerant Systems”, Springer-Verlag, Berlin, Heidelberg, 2011. Qin, J., “Survey on Data-Driven Industrial Process Monitoring and Diagnosis”, Annual Rev. Control 36:220234 (2012). Yin, S., Ding, S.X., Haghani, A., Hao, H., Zhang, P., “A Comparison Study of Basic Data-Driven Fault Diagnosis and Process Monitoring Methods on the Benchmark Tennessee Eastman Process”, J. Process Contr. 22:15671581 (2012). Yin, S., Ding, S.X., Xie, X., Luo, H., “A Review on Basic Data-Driven Approaches for Industrial Process Monitoring”, IEEE Trans. Ind. Electron. 61(11):6418-6428 (2014). Jolliffe, I.T., Principal Component Analysis, Second edition, SpringerVerlag, 2002. Wold, H., “Partial least squares”, in: Kotz, S., Johnson, N.L. (editors), Encyclopedia of Statistical Sciences, Vol 6. Wiley: New York, 1985, pp. 581591. Ku, W., Storer, R.H., Georgakis, C., “Disturbance Detection and Isolation by Dynamic Principal Component Analysis”, Chemometr. Intell. Lab. 30(1):179–196 (1995). Luo, R., Misra, M., Himmelblau, D.M., “Sensor Fault Detection via Multiscale Analysis and Dynamic PCA”,

J-FOR

13.

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

Ind. Eng. Chem. Res. 38(4):1489– 1495 (1999). Russell, E.L., Chiang, L.H., Braatz, R.D., “Fault Detection in Industrial Processes using Canonical Variate Analysis and Dynamic Principal Component Analysis”, Chemometr. Intell. Lab. Syst. 51(1):81–93 (2000). Choi, S.W. and Lee, I.B., “Multiblock PLS-Based Localized Process Diagnosis”, J. Process Contr. 15(3):295– 306 (2005). Mina, J. and Verde, C., “Fault Detection using Dynamic Principal Component Analysis by Average Estimation”, Proc. 2nd Intl. Conf. Elect. Electron. Eng., Mexico City, Mexico, 374–377 (2005). Marple, S.L., Digital Spectral Analysis with Applications. Prentice-Hall: Englewood Cliffs, NJ,1987. Rantala, S., Jokinen, H., Jokipii, M., Saarela, O., Suoranta, R., “Process Dynamics Study Based on Multivariate AR Modelling”, Proceedings, IECON’88 14th Annual Conference of the IEEE Industrial Society, Singapore, 1988, Vol. 2:320-325. Oguma, R., “Investigation of Power Oscillation Mechanisms Based on Noise Analysis at Forsmark-1 BWR”, Ann. Nucl. Energy 23(6):469-485 (1996). Saarela, O., Multivariate Autoregressive Analysis in Locating the Origin of Fluctuation in Continuous Industrial Processes. Thesis, Tampere University of Technology, Publications 366; 2002. Rosipal, R. and Trejo, L., “Kernel Partial Least Squares Regression in Reproducing Kernel Hilbert Space”, J. Mach. Learn. Res. 2:97–123 (2001). Cho, J.H., Lee, J.M., Choi, S.W., Lee, D., Lee, I.B., “Fault Identiﬁcation for Process Monitoring using Kernel Principal Component Analysis”, Chem. Eng. Sci. 60(1):279–288 (2005). Alcala, C.F. and Qin, S.J., “Reconstruction-Based Contribution for Process Monitoring with Kernel Principal Component Analysis”, Ind. Eng. Chem. Res. 49(17):7849–7857 (2010). Willis, A.J., “Condition Monitoring

24.

25.

26. 27. 28.

29.

30.

31.

32.

33.

34.

of Centrifuge Vibrations using Kernel PLS”, Comput. Chem. Eng. 34(3):349–353 (2010). Jack, L.B. and Nandi, A.K., “Support Vector Machines for Detection and Characterization of Rolling Element Bearing Faults”, Proc. Instn. Mech. Engrs. Conference: J. Mech. Eng. Sci. 215:1065–1074 (2001). Gao, C., Jian, L., Luo, S., “Modeling of the Thermal State Change of Blast Furnace Hearth with Support Vector Machines”, IEEE Trans. Ind. Electron. 59(2): 1134–1145 (2012). Duda, R.O., Hart, P.E., Stork, D.G., Pattern Classification, 2nd edition. Wiley: New York, 2001. Hyvärinen, A., Karhunen, J., Oja, E., Independent Component Analysis. Wiley: Hoboken, NJ, 2001. Chiang, L., Russell, E., Braatz, R., Fault Detection and Diagnosis in Industrial Systems. Springer-Verlag: London, 2001. Chiang, L.H., Kotanchek, M.E., Kordon, A.K., “Fault Diagnosis Based on Fisher Discriminant Analysis and Support Vector Machines”, Comp. Chem. Eng. 28(8):1389–1401 (2004). Chiang, L.H., Russell, E.L., Braatz, R.D., “Fault Diagnosis in Chemical Processes using Fisher Discriminant Analysis, Discriminant Partial Least Squares, and Principal Component Analysis”, Chemometr. Intell. Lab. Syst. 50(2):243–252 (2000). He, Q.P., Qin, S.J., Wang, J., “A New Fault Diagnosis Method using Fault Directions in Fisher Discriminant Analysis”, AICHE J. 51(2):555–571 (2005). Kano, M., Tanaka, S., Hasebe, S., Hashimoto, I., “Monitoring Independent Components for Fault Detection”, AICHE J. 49(4): 969–976 (2003). Lee, J.-M., Qin, S.J., Lee, I.-B., “Fault Detection and Diagnosis Based on Modiﬁed Independent Component Analysis”, AICHE J 52(10):3501– 3514 (2006). Lee, J.-M., Yoo, C., Lee, I.-B., “Statistical Process Monitoring With Independent Component Analysis”,

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

35.

36. 37.

38.

J. Process Contr. 14(5):467–485 (2004). Yu, J., “A New Fault Diagnosis Method Of Multimode Processes Using Bayesian Inference Based Gaussian Mixture Contribution Decomposition”, Eng. Appl. Artif. Intel. 26:456– 466 (2013). Bishop, C.M., Pattern Recognition and Machine Learning. Springer: New York, 2006. Chen, T. and Zhang, J., “On-Line Multivariate Statistical Monitoring of Batch Processes using Gaussian Mixture Model”, Comp. Chem. Eng. 34:500–507 (2010). Yu, J., “A Particle Filter Driven Dynamic Gaussian Mixture Model Approach for Complex Process Moni-

39.

40.

41.

toring and Fault Diagnosis”, J. Process Contr. 22:778–788 (2012). Dempster, A.P., Laird, N.M., Rubin, D.B., “Maximum Likelihood from Incomplete Data via the EM Algorithm (with Discussion)”, J. R. Statist. Soc. B 39:1-38 (1977). Nissinen, A. and Nuyan, S., “Performance Improvements and Sustainability Through a Remote Monitoring, Analysis and Warning System: Benchmarking Performance for Long-Term Stability”, Proc. Control Systems 2012, New Orleans, Louisiana, 684697 (2012). Jones, P.N. and McLachlan, G.J., “Maximum Likelihood Estimation from Grouped and Truncated Data with Finite Normal Mixture Models”,

42.

43.

44.

LIST OF ACRONYMS Expectation-Maximization (algorithm) EM FDA Fisher Discriminant Analysis Gaussian Mixture Model GMM ICA Independent Components Analysis KL Kullback-Leibler (divergence) Principal Components Analysis PCA PDF Probability Density Function Partial Least Squares PLS

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Appl. Stat. J. Roy. St. C 39(2):273–82 (1990). Cadez, I.V., Smyth, P., McLachlan, G.J., McLaren, C.E., “Maximum Likelihood Estimation of Mixture Densities for Binned and Truncated Multivariate Data”, Machine Learning 47:7–34 (2001). Bhattacharyya, A., “On a Measure of Divergence between Two Statistical Populations Defined by their Probability Distributions”, Bulletin of the Calcutta Mathematical Society 35:99– 109 (1943). Figueiredo, M.A.T. and Jain, A.K., “Unsupervised Learning of Finite Mixture Models”, IEEE Trans. Pattern Anal. 24(3):381–396 (2002).

Providing safe, reliable and efficient operations. Power and automation solutions for the pulp and paper industry.

Automation systems and industry environments are growing increasingly complex. Thousands of products and systems must work together seamlessly. Our solutions respond to the demand for the highest levels of quality, precision and accuracy in output, by delivering information that can be relied on to guarantee the best operational decisions for your plant. Read more on www.abb.com/pulpandpaper

CONDENSER COOLING-WATER CONTROL STRATEGY RECOVERS WASTE HEAT FROM POWER GENERATION ABSTRACT

DOUG BARBOUR, RYAN PRONTACK, JULIE BOISVERT*

An innovative cooling-water control strategy was used to recover waste heat from a condensing steam turbine generator at the Harmac Pacific pulp mill. Cold water is circulated through the turbine condenser to provide warmed water to the pulp mill, thereby displacing steam heating and saving energy. A recirculation loop around the condenser makes it possible to modulate warm-water flow to meet process requirements within turbine operating constraints. Harmac reduced their average process steam consumption by 15,000 kg/h, which is a substantial economic benefit to the mill. On average, 26% of the condenser waste heat was recovered. By recovering the waste heat from the turbine condenser, Harmac avoided the installation of a cooling tower, thereby saving capital costs, O&M costs, and space on a congested site.

INTRODUCTION

Harmac Pacific (Harmac) operates a northern bleached softwood Kraft (NBSK) pulp mill on Vancouver Island near Nanaimo, British Columbia. Nanaimo Forest Products Ltd., the owner of Harmac Pacific, re-started the mill with a single production line on October 3, 2008, producing NBSK at the rate of 226,000 tonnes per year. In 2009, a second production line was re-started, and capital upgrades in 2010 and 2011 brought the mill’s production capacity up to 365,000 tonnes of pulp per year. In 2013, Harmac completed the Biomass Electrical Generation Project. As part of this project, Harmac installed a 25 MW straight condensing steam turbine generator to take advantage of excess biomass boiler capacity. The mill operates two chemical recovery boilers and one biomass-fired power boiler. Before implementation of this project, all high-pressure steam (4,033 kPag, 397°C / 585 psig, 750°F) was directed to a 30 MW extraction back-pressure turbine that controlled the medium- and low-pressure steam systems. High-pressure steam was controlled by varying the biomass boiler firing rate. There was a unique opportunity to take advantage of unused biomass boiler

J-FOR

capacity to generate green power. BC Hydro was interested in acquiring green power from independent power producers to help limit their greenhouse gas emissions and stimulate the renewable energy industry in British Columbia. Under the Canadian government’s Green Transformation Program, Harmac implemented thirty energy reduction projects to improve steam usage efficiency in the mill. This increased the availability of biomass steam for power generation, and in 2011, Harmac signed a power sales agreement with BC Hydro and began the design of the Biomass Electrical Generation Project. During the design phase, Harmac realized that they had a unique opportunity

DOUG BARBOUR

RYAN PRONTACK

Production Engineer and Energy Manager, Harmac Pacific, Canada

Engineering Superintendent, Harmac Pacific Canada

to recover the waste heat rejected in the turbine condenser. This paper describes a turbine condenser cooling-water system that is integrated with the main mill water supply and provides warmed process water to the pulp mill. ENERGY-SAVING OPPORTUNITY

Some years ago, Harmac implemented several pinch point projects and process improvements. These changes eventually resulted in an adequate amount of hot water for the process and a significant increase in warm water temperature for part of the year. There remained a cyclical steam demand that matched the

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

JULIE BOISVERT

Senior Process Engineer, Temec Engineering Group Ltd. Canada *Contact: jboisvert@temeceng.com

TRADITIONAL AREA CONTRIBUTIONS cold river-water temperature entering the mill. Because steam was being used to heat cold water, there was an opportunity to use condenser waste heat to displace steam heating. A condensing steam turbine produces low-temperature waste heat. Normally, it is not economical to recover this low-grade heat, and it is discharged to the environment using a cooling tower, an aircooled condenser, or once-through cooling. Given adequate hot and warm water supplies, it made sense to produce warm water from the new turbine condenser to displace cold-water supply and reduce or eliminate seasonal steam demand. To maximize heat recovery potential, modifications were made to the water supply system. Mill processes that would benefit from receiving warm water were supplied preferentially with turbine condenser cooling water. By recovering the waste heat, energy savings were realized, and the projectâ&#x20AC;&#x2122;s economic return was enhanced. PROCESS DESIGN

A photograph of the Harmac turbine and condenser is shown in Fig. 1. The turbine was supplied by GE Thermodyn and the condenser by Maarky Thermal Systems. The turbine is an axial exhaust turbine with the generator located on the east side and the condenser on the west side on the same axis as the turbine. The cooling-water supply and return lines are visible on the south side of the condenser (in the upper left corner of the photograph). A condensing turbine exhausts wet steam at sub-atmospheric pressure. Turbine exhaust steam is a mixture of saturated steam and saturated water with a typical quality of 88% to 92%. The steam portion must be condensed to be reused as boiler feed water. The condenser is a shell-and-tube heat exchanger. Steam is condensed on the shell side, whereas cooling water circulates on the tube side. The condenser is responsible for setting the temperature

J-FOR

Fig. 1 - Turbine and condenser.

of the steam exiting the turbine. Because pressure and temperature are linked at saturation, the condenser effectively controls the turbine exhaust pressure. The cooling-water configuration for the Harmac G4 turbine condenser is once-through cooling with a recirculation loop around the condenser. A simplified process flow schematic is shown in Fig. 2. Cold raw water from a blended source

of river and well water is circulated in the turbine condenser to condense the turbine exhaust steam. The cooling water is pumped by two 50% capacity booster pumps through the condenser and onwards to two warm-water storage tanks. At the tank outlets, two pumps provide warm water to mill process users. Excess water not used by the process overflows the warm-water tanks to sewer.

Fig. 2 - Simplified process flow schematic.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

On the condenser outlet, there is a recirculation line (with a temperaturecontrolled recirculation valve) back to the condenser inlet. WHY RECIRCULATION AROUND THE CONDENSER?

The purpose of the recirculation loop is to enable the amount of warm water generated by the condenser heat-rejection system to vary. Varying the amount of recirculation makes it possible to modulate the net flow of water exiting the cooling loop, thereby generating sufficient warm water to match (as closely as possible) the mill warm-water requirements. Inside the cooling loop, the flow is maintained high enough to promote good heat transfer and prevent fouling of the condenser tubes. The flow in and out of the cooling loop is smaller than the flow through the condenser. WHY WARM-WATER STORAGE?

One of the design challenges was to manage water supply/demand imbalances that might occur between the turbine condenser and mill process users. It was therefore important to decouple the two systems, which was accomplished by storing warm water in two large-capacity tanks. These existing tanks were reassigned as water storage buffers with a combined capacity of 500 m3 (132,000 USG). If process demand is lower than turbine condenser water needs, water will overflow to sewer. If process demand exceeds the amount of cooling water available from the cooling loop, the tanks may be drawn down, and if high demand persists, cold water makeup is supplied to the tanks to maintain a minimum level. MILL STEAM PRESSURE CONTROL

Before the new G4 condensing turbine was added, all the high-pressure steam generated by the three boilers was directed to the existing G3 extraction back-pressure steam turbine. G3 controlled the pressures of the medium- and low-pressure steam

J-FOR

systems supplying process steam users. The pressure of the 4,033 kPag (585 psig) steam was controlled by adjusting the biomass boiler firing rate. Biomass combustion has a relatively slow response time to a change in steam demand; therefore, variations occurred in 4,033 kPag steam pressure due to process steam variations. Supplemental natural gas firing was used in the biomass boiler to maintain steam temperature when wet biomass was fed to the boiler. To maximize biomass boiler efficiency, Harmac installed a new natural gas-fired package boiler. By transferring load swings to the package boiler, biomass firing can be stabilized and maximized in the biomass boiler, improving its biomassburning capacity. The new gas-fired boiler is normally operated in header-pressure control mode. The new G4 condensing turbine is normally operated on load (MW) control to match the amount of steam available from the biomass boiler and meet power delivery targets. Operation of G4 in inlet pressure control mode is possible, but is not the normal mode of operation. Because two generators are involved in power delivery, the load set-point on G4 is typically cascaded to offset swings in the extraction back-pressure turbine to maintain constant power delivery to the grid. CONTROL PHILOSOPHY

The condenser cooling-water control system would be required to do the following: • Provide sufficient cooling water to the turbine condenser at all times; • Keep the cooling water to the condenser within a given temperature range to avoid an under- or over-pressure condition at the turbine exhaust; • Handle transients in turbine operation and process mill water demand. The control system attempts to match steam turbine cooling-water requirements to mill process water demand. In case of an imbalance between the water required by the process and by the

condenser, the control system must enable the turbine to stay on-line and operating within the target power-generation range and must satisfy the mill process users. The amount of heat rejected in the condenser depends on turbine power output. For example, at 20 MW, 172 GJ/h (163 MMBtu/h) is transferred to the cooling water. The amount of heat rejected is equal to the flow of cooling water through the condenser multiplied by the temperature rise across the condenser. For a given turbine power output, the only way to vary warm-water flow to process is to allow the temperature exiting the cooling loop to vary. This can be illustrated as follows: For 20 MW output, the heat rejected to the cooling loop is: Heat rejected = 172 GJ/h = flow of warm water exiting cooling loop x ( T warm water - T blended water ) The temperature of the blended water (entering the cooling loop) varies seasonally, and therefore on a short-term basis, it is assumed constant. By inspection of the above equation, it follows that to vary warm-water flow to the process, the temperature of the warm-water leaving the cooling loop must vary. This is achieved by acting on the recirculation valve. The temperature exiting the condenser will normally be between 32°C and 46°C (90°F and 115°F). CONDENSER FLOW-CONTROL LOOP

The condenser flow-control loop consists of the following: • 2 x 50% capacity condenser booster pumps with variable frequency drives; • A flow transmitter measuring flow through the condenser. A simplified schematic of the condenser flow-control loop is presented

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS

Fig. 3 - Condenser flow control loop.

in Fig. 3. Note that inside the cooling loop, there are multiple combinations of pump flows and recirculation-valve positions that will provide the same warm-water flow exiting the cooling loop. To avoid an overly complex control strategy, it was decided that pump flow would be a function of power output only and that the recirculation valve would act solely to control warm-water flow exiting the cooling loop. The controller varies the frequency of the booster-pump motors to maintain a target flow through the turbine condenser. The flow set-point is between 536 and 757 l/s (8,500 and 12,000 USGPM) and is a function of turbine power output. Varying the flow through the condenser with power output saves energy at low power outputs and ensures sufficient cooling water flow through the condenser when power output is increased. This proposed strategy is simple and has proved stable in operation. There are other possible control strategies that may improve turbine efficiency, such as minimizing recirculation flow at all times. The design team chose a simple strategy to commission and to gain experience with the control system. Harmac may review and adjust the control strategy in the future.

WARM-WATER FLOW CONTROL LOOP

The warm-water flow control loop consists of the following: • a level indicator controller measuring the level in the warm-water storage tanks; • a temperature indicator controller measuring the temperature of the cooling water exiting the condenser; • a temperature control valve acting on the recirculation flow from the condenser outlet to the condenser inlet. A simplified schematic of the warm-

water flow control loop is presented in Fig. 4. This loop controls the amount of water leaving the condenser cooling recirculation loop and available for the process. The level controller at the warmwater tanks will request a raising or lowering of the temperature set-point for the recirculation valve. Lowering the temperature set-point means a request for more process water. The recirculation valve will close, thereby increasing the net flow exiting the cooling loop. Similarly, raising the temperature set-point means a request for less process water, which will cause the recirculation valve to open. The temperature set-point is not allowed to increase above 41°C (105°F). If the process requires less water and requests to raise the temperature set-point above 41°C, the request will be ignored. Excess warm water will overflow the warm-water tanks to sewer. The temperature set-point is not allowed to drop below the minimum dictated by the turbine low exhaust pressure limiter. When the exhaust pressure nears the low limit, the turbine controller will override the process water request and prevent the recirculation valve from closing any further. In this case, there will be less warm water available than the process requires, and the level in the tanks will drop. If high demand persists, cold water make-up is supplied to the tanks to maintain a minimum level.

Fig. 4 - Warm-water flow control loop.

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

OPERATING CONSTRAINTS

The performance of the condenser is characterized by the steam-side pressure it can maintain while condensing its design duty at the given cooling water flow and temperature. This condenser has the following design rating: Design pressure: 10 kPaa (2.95 in HgA) Steam flow: 104,356 kg/h (230,000 lb/h) Cooling water flow: 757 l/s (12,000 USGPM) CW supply temperature: 21.1°C (70°F) CW range (delta T): 19.4°C (35°F) At design conditions, when the turbine is operating at maximum power generation (25 MW), the condenser will generate 757 l/s (12,000 USGPM) of warm water at 41°C (105°F). If the conditions are different than the above, such as a lower power output or a different cooling water flow or temperature, the condenser will produce an exhaust pressure different from the design pressure. For example, if the cooling-water source is colder than 21.1°C, all other parameters being equal, the turbine exhaust pressure will drop below 10 kPaa, and the turbine power output will increase slightly above 25 MW. The reverse is true as well; a warmer supply of cooling water (above 21.1°C) will result in reduced power output. If the turbine happens to be in power control mode, where power set-point is maintained, then a change in exhaust pressure will affect the steam flow required by the turbine. It is apparent from this discussion that variations in cooling-water flow or temperature to the condenser affect turbine performance. Some variation can be tolerated, but too much variation can lead to problems. There are limits that cannot be exceeded on the turbine exhaust pressure (see Table 1). If the pressure is allowed to drift too high, the turbine will trip. If the exhaust pressure drops too low, then the turbine control system will intervene to prevent operation below the low exhaust pressure limit (no trip). From a steam turbine point of view,

J-FOR

TABLE 1

Important pressure / temperature parameters for the Harmac G4 steam turbine.

Max. allowable exhaust pressure (turbine trip) Rated exhaust pressure Min. allowable exhaust pressure

it is most advantageous to operate the cooling-water system to maximize power output. This means supplying the condenser with as much cooling water flow at as low a temperature as required to maintain the exhaust pressure above the minimum allowable value. However, coupling the condenser with the mill water requirements means that the condenser will see varying cooling-water flows and temperatures. This sets up a potential conflict between the turbine needs and the process needs. The first priority is to maintain the steam turbine on-line and producing power to meet the target power production schedule. It was therefore very important that the control strategy ensure a reliable and steady supply of cooling water and that the temperature of the cooling water to the condenser be in line with the desired turbine exhaust pressure.

Shell-side pressure (Psat) kPaa (in Hg A) 25.0 (7.38) 10.0 (2.95) 6.2 (1)

Shell-side temp. (Tsat) °C (°F) 65 (149) 46 (114) 37 (98)

TURBINE EFFICIENCY CONSIDERATIONS

Adding a recirculation loop warms the cooling water entering the condenser and therefore impacts the power generation efficiency of the steam turbine. A cooling tower would have a similar effect, especially in summer when the temperature to the condenser would be dictated by the ambient wet bulb temperature and the tower approach temperature. Harmac is operating the cooling-water system to maximize power generation at times and to minimize water consumption at times depending on power generation schedule, boiler availability, and other operating factors. Overall, the effect on power generation efficiency is similar to what it would have been had Harmac installed a cooling tower.

Fig. 5 - Annual energy profile for G4 power generation.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS

OPERATING EXPERIENCE

The turbine has been operating reliably since November 2013. It has had excellent availability with no unplanned outages for the first few years of operation. The turbine is operated to deliver power according to a power generation schedule, as shown in Fig. 5. This schedule takes into account the availability of biomass fuel, the amount of steam required by the process, and the power generation of the existing G3 steam turbine. Because Harmac must meet power generation targets, it sometimes has to operate at maximum power output to catch up following a mill-related process upset or outage. The condenser cooling-water system has also been operating in a stable and reliable manner. The only issue identified since startup was the need for a second water strainer on the river water supply to the blended cold mill water system. During run-off season, river water debris is high, and dual strainers (one on-line at all times) proved necessary to avoid fouling of the turbine condenser. ENERGY SAVINGS

Harmac has been able to maximize recovery of turbine condenser waste heat to the limit of current process water usage by the mill. Figure 6 shows Harmac’s specific steam consumption before and after implementation of the G4 project. Specific steam consumption is the total amount of steam consumed by the mill after subtracting the steam sent to G4 for power generation. It is a measure of the energy efficiency of pulp production and is used to track ongoing mill performance. Annual average mill steam consumption has been reduced by 181 kg/ADt (0.4 klb/ADt) at the average pulp production of 1,050 ADt/d. The energy saved due to the heat recovery project is actually higher than reported in the above graph. Had there been no condenser heat

J-FOR

Fig. 6 - Specific steam consumption before and after G4.

recovery, the specific steam consumption would have increased due to two factors: - Increased de-aeration load due to the need to reheat the G4 turbine condensate; - Increased steam consumption due to re-blading of G3 to increase its efficiency. By means of a heat balance we calculated the increase in specific steam consumption that should have occurred had there been no condenser heat recovery. The results are presented in Table 2. The actual steam savings due to implementation of the condenser heat recovery project were 363 kg/ADt (0.8 klb/ADt). This represents an annual average steam generation reduction of 15,430 kg/h (34,000 lb/h), which is a significant energy saving for the mill. The amount of waste heat recovered is 26% of the amount of heat rejected in the condenser at an average annual power TABLE 2

output of 17 MW. This amount could have been higher had there been more process water demand by the mill. OTHER SAVINGS

The condenser heat recovery project enabled Harmac to avoid the capital and operating expenses associated with installing a cooling tower to reject heat to the environment. The estimated capital cost for a 222 GJ/h cooling tower and all accessories is $7.3 M CAN in 2012 dollars. This estimate was partially based on cost experience from another project. Offsetting costs were incurred for piping modifications and new pumps required for the heat recovery project. Caution must be exercised in comparing the costs of engineered and installed equipment and of equipment avoided and not engineered. Therefore, the precise capital cost savings are difficult to estimate. The Harmac site is congested, and allocating space for a cooling tower

Determination of speciﬁc steam savings due to condenser heat recovery. Impact on specific steam consumption

Increased de-aeration load due to addition of G4 Increased efficiency of G3 due to re-blading Expected increase in specific steam consumption Measured change in 2014 compared to 2012–2013 Actual reduction due to heat recovery project

+ 136 kg/ADt + 46 kg/ADt + 182 kg/ADt - 181 kg/ADt 363 kg/ADt

(0.3 klb/ADt) (0.1 klb/ADt) (0.4 klb/ADt) (-0.4 klb/ADt) (0.8 klb/ADt)

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

would have been a challenge. CONCLUSIONS

Harmac has designed and built a turbine condenser cooling-water system that is integrated with the main mill water supply, providing warmed process water to the pulp mill. The project was a success, with excellent reliability since operation started in the third quarter of 2013. The heat recovery project enabled Harmac to reduce their specific steam consumption by approximately 363 kg/

ADt (0.8 klb/ADt). This represents an annual average steam savings of 15,430 kg/h (34,000 lb/h), which is a substantial economic benefit to the mill. On average, 26% of the heat that would have been rejected to the environment was recovered by warming process water. The rest of the heat was sent to secondary effluent treatment for disposal into Georgia Strait. Had there been more process water usage, the energy savings would have been greater. Harmac continues to look for new opportunities to make use of the excess warm water generated by

the heat recovery project. By recovering the waste heat from the turbine condenser, Harmac avoided the installation of a cooling tower thereby saving capital costs, O&M costs, and space on a congested site. ACKNOWLEDGEMENTS

The authors would like to acknowledge the contribution of Doug McKenzie of Allnorth Consultants and his role in the conceptual design of this cooling water system.

WHY JOIN PAPTAC? Access to a wide and well-established network, including an on-line directory for members; Participation priviledge on PAPTAC's technical communities through meetings, e-mail discussion groups, conference calls, on-line forums and blogs; Priviledged member access to J-FOR on-line; A one-year subscription to Pulp and Paper Canada magazine, and the PaperAdvance and Le MaĂŽtre papetier e-newsletters; A free complete set of Engineering Data Sheets and Standard Testing Methods; Reduced rates on books, events, courses and more.

PAPTAC is a wide and well-organized network of individuals and represents a world of information exchange and experience for the benefit of the professional development of its members and the pulp and paper industry. We are inviting members to get actively involved in our industry by joining our technical communities, by attending conferences and presenting papers, and by participating in our TECH courses and Webinars. The benefits of involvement and participation will provide value back the individual member, the company and to the industry.

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

PROSYGMA.CA

Phone: 450 442-5100 | Toll-free: 1 800 736-6160 370, Chambly Rd, # 300 Longueuil (Qc) J4H 3Z6

ASB DREDGING AND ODOUR CONTROL ABSTRACT

JENNIFER FOWLER, DAN DAVIES*, WEB TEETZEL

Normal operation of the Hinton mill’s ASB (aerated sludge basin) results in slow accumulation of solids in this pond. As the level of these solids increases, they can cause various problems. These include decreasing the effective volume of the pond (and hence the residence time) and generation of TRS (total reduced sulphur) odours. The TRS odours are the result of anaerobic digestion of the sludge. Under normal circumstances, these odours are a nuisance, causing complaints from the surrounding community. However, the pond is dredged every two years to remove sludge, and agitation of the sludge and its exposure to the atmosphere causes much higher TRS levels, particularly near the pond and the dredging operation. Because these higher TRS levels are a potential health hazard for personnel at or near the pond and the dredging operation, the TRS is eliminated before it can be released by application of H2O2 (hydrogen peroxide). This paper will discuss the dredging operation and the steps taken to control TRS odours. Before implementation of the hydrogen peroxide addition, there had been a notable increase in community odour complaints while dredging was in progress.

INTRODUCTION

Common methods of treating liquid effluent from a pulp mill are clarification, aeration, and settling. The Hinton pulp mill’s aeration stabilization basin (ASB) is one step in this process, aerating the effluent and then allowing solids to settle. Despite surface aeration, the settled solids can become anaerobic, allowing generation of TRS species. Although TRS can be slowly released through normal operation, any agitation causes a much greater release. Approximately every two years, the mill dredges the solids from the ASB, which both agitates the solids and exposes them to air, causing much greater release of TRS. This TRS release may be a nuisance to the mill’s neighbours, but can be a health hazard for personnel working on the dredging operation or in the vicinity. This dredging operation is a necessary part of ASB maintenance, ensuring that solids accumulation does not adversely affect the pond’s retention time and lessening TRS release during routine pond operation. The TRS consists mainly of H2S, and hence H2O2 was deemed suitable to control the odour. H2O2 is also effective in oxidizing the other odorous compounds of concern in mill effluent, including methyl mercaptans, dimethyl sulphide, and dimethyl disulphide. Past experience

J-FOR

has shown that during dredging, most of the H2S was released from the de-watering equipment. This equipment consisted of a grizzly screen, a large mix tank with agitation, a centrifuge and a belt press, and finally a short conveyor to transport the sludge into dump trucks. DREDGING OPERATION

Hinton Pulp dredges the north pond of the ASB every two years to remove settled biological sludge. On average, 10,000 BDt (bone-dry tonnes) of sludge are dredged over a period of three months. The goals of the dredging program are to increase the ASB residence time and to minimize the amount of settled sludge in the basin undergoing anaerobic digestion. This digestion is a source of TRS odours if left

unchecked. Sludge is dredged from the bottom of the pond using a cutter head dredge. The material is pumped to a mixing tank, where polymer is added to promote coagulation of biological matter. The sludge is then sent through a centrifuge for water removal; typically, sludge enters the centrifuge at 8% solids and leaves at 30%–35% solids. The de-watered sludge is then transported to landfill for disposal. The potential to release TRS to the atmosphere exists particularly during centrifuging. It is this TRS release that can have an impact on the Town of Hinton, as well as on personnel working as part of the dredging operation or nearby. According to the Hinton Pulp approval issued by AESRD (Alberta

JENNIFER FOWLER

DAN DAVIES

WEB TEETZEL

Technical & Environmental Superintendent, Hinton Pulp, A Division of West Fraser Ltd. Canada

Applications Manager Evonik Canada Inc. Canada *Contact: dan.davies@evonik.com

Sales Manager, Western Canada Evonik Canada Inc. Canada

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS Environment and Sustainable Resource Development), dredging can occur only from May to September; this period overlaps with the highest frequency of odour complaints. Since 2007, Hinton Pulp has used hydrogen peroxide addition to the sludge settling tank to oxidize the TRS compounds before centrifuging. Community testing with the Jerome monitor in 2011 and 2013 has shown this oxidation process to be effective in minimizing odours. This effectiveness is critical in controlling community odour inquiries because residents are more susceptible to odours due to increased time spent outdoors in spring and summer. TRS ABATEMENT

A portable H2O2 tank and dosing skid were set up between the pond and the dewatering equipment. This made it possible to use the de-watering equipmentâ&#x20AC;&#x2122;s power and water supply for the H2O2 equipment. Although the de-watering equipment is a relatively large and complex installation, for temporary equipment, the H2O2 equipment was quite simple. Besides the portable storage tank, the dosing skid consisted of a gear pump with a variablespeed motor. The dosing rate was manually controlled by the variable-speed pump and an in-line rotameter. Other preparations included H2O2 safety training for all contractor personnel and supervision of the dosing line installation. The safety training ensured that the contractors were prepared to handle the H2O2 in the exposed, temporary operation; they were comfortable with the equipment and confident with the peroxide. A particular concern with the location in which this operation took place was the normally windy conditions. In this situation, the end of the dosing line could not be submerged, but had to be above the highest level in the settling tank. This was necessary to avoid any possibility of submerging the end of the dosing line, possibly allowing backflow of pond fluids into the dosing line. If any pond fluids

J-FOR

Fig. 1 - Pond before treatment.

were to backflow into the dosing line or H2O2 storage tank, they would contaminate the tank, possibly causing the H2O2 to decompose. This would release large amounts of heat and steam, wasting

the H2O2. However, this configuration meant that there would be a free drop of H2O2 from the end of the dosing line into the settling tank. Although the winds were

Fig. 2 - Positioning of hydrogen peroxide equipment.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

not excessive, they were sufficiently strong to present the risk of splashes of windborne H2O2 from the free-falling H2O2. To avoid this possibility, a splash / wind shield was installed on the grating above the dosing point. This simple sheet-metal shield protected the open H2O2 stream from being affected by the wind as it left the end of the dosing line. In addition, it ensured that any splashes that did occur would not escape the settling tank. CONCLUSIONS

H2S, mercaptans, and related compounds are nuisances not only because of their odour, but very offensive even at low concentrations in air (detectable as low as 1 ppb), causing problems for employees, locals, and even distant neighbours. In addition, concentrations of these gases can reach hazardous levels during dredging operations if control measures are not in place. Special care must be taken with H2S because it dulls the sense of smell, which can prevent a warning of high concentrations. H2O2 addition is a simple method of treating a potentially serious, even hazardous, odour problem. Under these conditions, H2O2 is a robust treatment that is very forgiving of temperature, pH, residence time, contaminants, or concentration conditions. It reacts very quickly,

Fig. 4 - Hydrogen peroxide dosing pump.

J-FOR

Fig. 3 - Dewatering and treatment equipment.

enabling its addition in pipe, tank, or centrifuge. Hinton Pulp operates their dredging program biannually to remove sludge, reduce odour, and increase freeboard in the pond. Use of H2O2 for TRS control during the dredging operation has become a permanent part of this program.

REFERENCES Davies, D., Christy, T., O’Connor, B., 1. “Combatting Odour Problems in a Kraft Mill Using Hydrogen Peroxide”, Proceedings, PAPTAC Annual Meeting, Montreal QC, February 2000. Fowler, J., Start, G., O’Connor, B., El2. lis, S., “Identification of Key Odour Sources and Abatement Strategies Implemented at Hinton Pulp”, Proceedings, PACWEST Conference, Jasper AB, May 2014.

Fig. 5 - Splash shield for hydrogen peroxide addition point.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

POST-D0 TOWER BRIGHTNESS – USEFUL OR NOT? ABSTRACT

JAMES GOLDMAN*, MICHAEL SUMMERFORD

This paper reports a study of data from the post-D0 brightness measurement and shows how it correlates with pre-D0 and post-EOP measurements. It is well established that D0 brightness is not very useful in a conventional bleach plant. Instead, this study examines data from two oxygen-delignified bleach lines to evaluate whether this measurement is useful for control. These data correlations will be used to determine whether this measurement should be used for feedback control of ClO2 on the D0 stage. Whether to use this measurement for feedforward control of oxygen and peroxide usage in the EOP stage will also be evaluated. This initial data study has revealed a reasonable potential to use post-D0 brightness of 60 or above to improve control of the D0 and EOP stages.

INTRODUCTION

The use or non-use of absolute brightness after the D0 stage has long been debated with regard to bleach plant evaluation and control. The common understanding is that the delignification reaction is not complete at this point and that the pulp first needs to undergo extraction before the stock’s brightness value becomes important. The D0 stage acts to oxidize the lignin bonds in the pulp fibre, but the lignin remains in the pulp until it is solubilized and washed out in the subsequent EOP stage. However, in today’s modern bleach plants, the oxygen delignification stages are acting to reduce pulp lignin content significantly and to brighten it before the stock ever enters the D0 stage. Due to these increased brightness levels entering and exiting the D0 stage, the potential usefulness of the D0 after-tower brightness measurement should be re-evaluated. For purposes of this evaluation, two bleach plants were studied. Data from both bleach plants were snapshots at one-hour intervals. For Line A, one year’s data were available. For Line B, only three months of data were available for evaluation. MEASUREMENT EVALUATION

Before oxygen delignification, in conven-

J-FOR

tional bleach plants, unsuccessful efforts were made to measure post-D0 brightness for use in either feedback or feedforward control of front-end bleaching. Depending on incoming kappa, particularly on softwood, exit D0 brightness values between the mid-30s and the low to mid-40s were common even with significant chemical addition. At these low values (30s and 40s), the pulp is still so dark that an absolute brightness value does not provide any “bleaching degree” indication whatsoever. Bleaching cannot take place until enough of the lignin is gone. Up to this point, the pulp is truly undergoing only extended and perhaps even some remaining bulk delignification. As is well known, at this stage in the process, the oxidized lignin (a substantial amount) has not yet been dissolved so that most of it can be removed. Therefore, kappa readings in particular are of little to no use until the oxidized lignin in the fibre has been removed in an alkaline medium. Fast forward to modern pulping sequences with pressurized oxygen delignification, particularly two-stage processes, and one will make a different observation. Kappas entering the D0 stage are now 50%–65% lower than previously, with corresponding absolute brightness values entering the stage from the mid-30s

to the low to mid-40s instead of that same range leaving the stage as in a conventional bleach plant. Figure 1 illustrates the relationship of inlet brightness to kappa for a modern softwood bleach plant. Note the

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

JAMES GOLDMAN

Product Manager – Pulp Solutions Valmet USA 7281 Northlake Rd Sterlington, LA 71290 U.S.A *Contact: james.goldman@valmet.com

MICHAEL SUMMERFORD

Field Service Engineer Valmet USA 1697 Bill Futch Rd Ellabell, GA 31308 U.S.A

TRADITIONAL AREA CONTRIBUTIONS high correlation of inlet brightness with kappa, which is indicative of good brownstock washing on this line. For this trend, the data were sorted by inlet kappa number, and data sets of 50 were averaged to obtain the overall trend. At these higher inlet brightness levels, bulk delignification is largely exhausted, and much extended delignification has also taken place. The D0 stage, albeit still performing the primary role of extended delignification, is now also providing some small amount of “bleaching”. Modern D0 stages are smaller than conventionally, with shorter residence times. This lack of excess retention (causing extended durations of residual loss) minimizes the significant former issue of reversion that made meaningful D0 brightness measurement a challenge. Absolute brightness values of well-washed, non-extracted pulps exiting D0 now range from the upper 50s to the mid- to upper 60s. At these levels (absolute brightness values above 55–60), the pulp is now bright enough to provide a meaningful (repeatable) measurement of what has taken place in the D0 tower. Figure 2 shows the relationship between D0 inlet kappa and D0 outlet brightness on two softwood bleaching lines. The data compare the time-delayed results from inlet kappa to the resulting outlet brightness. Once again, the data were sorted, and data sets of 50 were averaged for trending. This same method was also used for most of the remaining trends. For ClO2 dosage determination, both bleach lines use some form of modified kappa-factor control. The biggest difference between the two is that Line B has cleaner pulp (lower COD) than Line A. The conclusion that can be drawn from these data is that the D0 outlet brightness is predictable and linear on Line B compared to inlet kappa and could potentially be of value for feedback control. This is not as evident for Line A, even though the inlet kappas are in the same range. Pre-bleach carryover and D0 mixing may be contributing factors, but the differences are not fully understood at this time. It does appear, even on this line, that outlet D0 brightness values above 60

J-FOR

Fig. 1 - Inlet brightness vs. inlet kappa.

Fig. 2 - D0 kappa vs. D0 outlet brightness.

could be useful for feedback. To use this feedback for D0 chemical control, this same measurement, or that of delayed inlet kappa number, must be capable of being used for feedforward control of EOP oxidation or bleaching. In this manner, the long feedback from post-EOP measurements would be eliminated, and multiple front-end corrections would be avoided. Effectively, conventional long feedback would be lessened to short feedback after the D0 stage. From the above argument, it was proposed to determine whether the D0

brightness measurement could be useful as a feedforward tool for EOP control. Normally, the delayed D0 inlet kappa is used as a feedforward signal to the EOP stage. Figure 3 shows how the post-D0 brightness compares to the time-shifted EOP exit brightness. Figure 3 shows that D0 exit brightness correlates somewhat with EOP discharge brightness, although quite differently in the two bleaching-line examples. Figure 4 shows how this observation compares to the more common delayed D0 kappa versus EOP brightness relationship.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Fig. 3 - D0 brightness vs. EOP brightness.

Fig. 4 - D0 inlet kappa vs. EOP brightness.

Fig. 5 - D0 brightness vs. EOP kappa.

Fig. 6 - Delayed D0 inlet kappa vs. EOP kappa.

Surprisingly, the relationship shown in Fig. 4 is not stronger than the regression relationship of D0 [brightness] to EOP brightness, although approximately the same. Given this, it was then proposed to compare the correlations between delayed inlet kappa and EOP kappa against those between D0 brightness and EOP kappa. Figure 5 shows the relationship between D0 outlet brightness and the resulting EOP kappa for both bleaching lines. These correlations seem to be very repeatable. Figure 6 shows the more commonly used measurement of delayed D0 inlet kappa versus the resulting EOP kappa for both lines. The result shown in Fig. 5 suggests strongly that EOP kappa can be reasonably predicted from D0 exit brightness on both bleach lines. Even more surprisingly, this relationship appears stronger than that of delayed inlet kappa to EOP outlet kappa. To evaluate this finding further, it

J-FOR

was decided to plot peroxide charge per unit of EOP inlet brightness and per unit of delayed inlet kappa versus brightness gain across EOP as well as delayed kappa drop (delignification). These comparisons are shown in Figs. 7 and 8. Chemical usage data were available only from Line A, and hence these charts show results from Line A only. In Fig. 7, the correlation of a dropoff in brightness development with increased “brightness factor” peroxide charge seems fairly odd. Likewise, Fig. 8 appears to show a strong linear correlation with decreased delignification at higher peroxide “kappa factors” or ratios. This phenomenon can be explained in both charts by an over-charge of baseloaded peroxide at low lignin demands. One informative detail in Fig. 7, which partially confirms the previous finding (D0 exit brightness in the higher range as a good delignification indicator), is that

the kappa drop increases linearly with peroxide charge. However, this correlation needs to be studied further. Lastly, the D0-EOP stage combination was reviewed with respect to conventional kappa-factor associations with both delignification and brightness gain. In addition, using a ClO2 brightness factor based on D0 inlet brightness was also investigated. The results are shown in Figs. 9 and 10. Note that because bleach line A did not measure absolute pre-D0 or inlet brightness, these values were extracted for purposes of analysis using the inlet kappa to brightness equation provided in Fig. 1 from a different softwood bleach line (Line B). Although this type of analysis should not be used for accurate performance assessment or exact prediction, it does highlight two important points. First, conventional front-end delignification control using kappa-factor bleaching is highly

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TRADITIONAL AREA CONTRIBUTIONS

Fig. 7 - EOP stage benefits vs. D0 brightness.

Fig. 8 - Benefits versus delayed inlet kappa.

Fig. 9 - Delignification as a function of inlet brightness and kappafactor ClO2 charge.

Fig. 10 - Brightness gain as a function of inlet brightness and kappa factor ClO2 charge.

repeatable and a good form of control, as has historically been documented. Second, it appears that for relatively high inlet brightness values, post-D0 absolute brightness may have worthwhile potential as a â&#x20AC;&#x153;middle-manâ&#x20AC;? controller for the bleach-plant front end. CONCLUSIONS

The use or non-use of D0 brightness for front-end dosage control will continue to be debated long after this paper. However, it appears that in a certain brightness

J-FOR

range, it does have potential for use in D0 feedback and EOP stage feedforward control. As O2 stages continue to delignify pulp further and bleach-plant inlet brightness continues to increase, the potential for using D0 brightness for meaningful control will continue to increase as well. To obtain a further understanding of this potential, a partnership with an appropriately configured mill or mills should be established to study this topic in greater depth. Meanwhile, EOP kappa remains as a true indicator of front-end performance and should continue to be used for

front-end control. ACKNOWLEDGEMENTS

We would like to thank the engineers that provided the data used in this study, although they have chosen to remain anonymous. We would also like to thank the PAPTAC bleaching committee for posing this question and providing many responses to the question of the usefulness of D0 brightness earlier this year. Many of the responses were either considered or used in preparing this paper.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

3D PRINTING – A REVIEW OF TECHNOLOGIES, MARKETS, AND OPPORTUNITIES FOR THE FOREST INDUSTRY ABSTRACT

TINGJIE LI*, JOSEPH ASPLER, ARLENE KINGSLAND, LYNE M. CORMIER, XUEJUN ZOU

3D printing has been used to produce prototypes and molds in industry for many years. It is now gaining increasing interest, due to its expansion into consumer products markets as well as the exploration of potential applications in manufacturing large architectural structures. Recent intellectual property changes (especially the expiration of early patents) have led to the availability of low-cost 3D printers and novel applications of 3D printing technology. This has turned 3D printing from a small-scale prototyping tool into a potential game changer for product development and manufacturing. This report summarizes the development of 3D printing, its markets, and its applications in the short and medium term. 3D printing is ideal for mass customization, but is not yet capable of mass production because of limitations on cost, speed, and materials. The near future of 3D printing will focus on production tooling, on-demand parts at low volumes, design and educational tools, products made at home, and even large architectural structures. The potential applications for forest products have been examined. Forest industry products based on wood fibres and wood-based biomaterials are already being prototyped using 3D printing. There is also increasing interest in using wood-based materials as feedstocks for 3D printing to impart a unique appearance to printed products and meet the demand for sustainability from consumers. This development could become part of transforming the forest industry from production of commodities to production of high-value-added specialty products. Some opportunities for the forest industry have been identified.

INTRODUCTION

Only five years ago, 3D (three-dimensional) printing technology was rarely seen outside of trade shows and development centres, producing prototypes and structures that could not be produced by conventional manufacturing. Since then, 3D printing has started to expand very rapidly. Home hobbyists can now buy 3D printers at low cost. The aerospace, automotive, and military industries are investing heavily in the field with the stated goal of fabricating high-quality objects in

TINGJIE LI

low to medium volumes using 3D printing [1]. The news media—both technical and popular—now carry articles and features on 3D printing on a regular basis.

Material Extrusion STATUS OF TECHNOLOGY DEVELOPMENT Types, Technologies, and Materials that can be Handled

Many processes and materials are defined as 3D printing. Based on ISO/ASTM

JOSEPH ASPLER FPInnovations FPInnovations 570 boul. St-Jean, 570 boul. St-Jean, Pointe-Claire, Quebec, Pointe-Claire, Quebec, Canada Canada *Contact: tingjie.li@fpinnovations.ca

J-FOR

52921:2013 E, Terminology for Additive Manufacturing, 3D printing technologies can be categorized as in Table 1. Fused deposition modelling (FDM) (Fig. 1) systems represent the largest installed base of 3D printers [11]. In FDM, the printing nozzles (extrusion heads) melt and fuse thermoplastic material. The liquefied material is deposited layer by layer, directed by a digital file, to form the final

ARLENE KINGSLAND

LYNE M. CORMIER

XUEJUN ZOU

FPInnovations 570 boul. St-Jean, Pointe-Claire, Quebec, Canada

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

EMERGING AREA CONTRIBUTIONS

TABLE 1 Type

Technologies Fused deposition modelling (FDM) [2]

Summary of key technologies.

Materials (feedstocks) Thermoplastics (e.g., PLA, ABS)*, liquid metals, edible materials

Pros

Cons

Makers

Ideal for conceptual and engineering models and functional prototypes

Low resolution and high surface roughness

3DSystems (USA) Stratasys (Israel)

Aqueous slurries and dispersions

Ideal for biomaterials

Photopolymer

Excellent for highLow speed, resolution limited choice of applications with photo-polymers complex geometries

Material extrusion Bio-printing [3]

Vat photopolymerization

Stereolithography (SLA) [4]

Material jetting

PolyJetTM [5] Powder-bed fusion (e.g., SLS [6], SHS [7]) **

Granular material binding

Sheet lamination

Thermoplastics, metal powders, ceramic powders

Binder jetting (3DPTM) [8]

Thermoplastics, plaster

Direct energy deposition [9]

Metal powders

Laminated object manufacturing (LOM) [10]

Paper, metal foil, plastic film

Maintaining EnvisionTEC structural integrity during drying 3DSystems Stratasys EnvisionTEC (Germany) - 3D Systems - Voxeljet (Germany) - SLM Solutions (Germany) - ExOne (USA) - EOS (Germany) - Arcam (Sweden)

Ideal for durable, functional parts with a variety of applications

Low speed

Easy access to printing materials, wood-like 3D

Mcor (Ireland) Low strength, narrow application

*polylactic acid (PLA), acrylonitrile butadiene styrene copolymer (ABS) **selective laser sintering (SLS), selective heat sintering (SHS)

solid object. Since 2009, the number of low-cost 3D printers from both major and start-up companies has greatly increased, with many using the now off-patent FDM technology. Bio-Printing

Bio-printing is another form of 3D printing that is particularly suited to biological materials. Bio-printers [3,12] may be constructed in a configuration similar to FDM printers, but the print heads deal with lower-viscosity feedstocks rather than hot-melt plastics. Because the feedstocks for bio-printing are aqueous slurries, a large volume of water may have to be removed. Bio-printing allows constructing a 3D â&#x20AC;&#x153;scaffoldingâ&#x20AC;?, onto which the biomaterial is then printed. The application of this and similar

J-FOR

devices to printing of biomaterials, including cellulosics, will be discussed below.

object. The PolyJet 3D printing technology possesses the advantage of higher printing resolution (layer height

Vat Photopolymerization

In stereolithography (SLA) [4], the first patented and commercialized 3D printing process, a platform which serves as the base for the object is submerged into a vat of polymer. A UV laser cures and hardens these polymers with each pass over the object. Once a pass is finished, the platform lowers slightly into the vat, allowing more uncured polymer to cover the object. Material Jetting

PolyJet 3D printers jet layers of liquid photopolymer onto a building tray and cure them with UV radiation [5]. The layers build up one at a time to create a 3D

Fig. 1 - Fused deposition modelling (FDM) [2].

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

of 0.089–0.12 mm for PolyJet, compared to 0.1–0.3 mm for FDM) [5]. Granular Material Binding

The powder-bed fusion, binder jetting, and directed energy deposition technologies share the similarity of binding granules into solid objects. Granular material binding technologies have fewer feedstock constraints. One layer of a powdered raw material is added by an applicator that passes over the building plate. The materials are dried or hardened by laser, electron beam, or heat. The applicator makes another pass, and the next layer is added. As will be discussed below, this technology may be of critical interest to the forest products industry if the hardening agent is a printed adhesive. One form of granular material binding is shown in Fig. 2. Similarly to fused deposition printing, as the early patents for this process expire, a new series of low-cost printers may proliferate. Current Market and Applications for 3D Printing

3D printing can facilitate understanding between designers and engineers as they bring design into reality. With a physical prototype, engineers can see the concept better in real terms (Fig. 3a). 3D printing also benefits the fashion and art design industries. 3D printing is starting to provide major benefits to medical device manufacturers to provide customization to patients at low cost, to satisfy the need to introduce innovative products, and to advance bioengineering (cell culturing). The feedstocks are mainly polymers, ceramics, metals, and biological cells. The value of

Fig. 2 - Powder-bed fusion [6].

J-FOR

3D-printed medical products is predicted to grow from $11 million in 2012 to $1.9 billion in 2025 [13]. With increased printing speed and material choices, 3D printing can open up more applications for direct part production. Hewlett-Packard, a leader in laser and ink-jet printing, has announced that they are developing their own proprietary “Multi-Jet Fusion” 3D printing technology. According to their most recent press releases, this technology will be officially introduced in 2016 [14]. They claim that this technology will be both faster and cheaper than those currently used. Metal powder printing (followed by near-instantaneous high-temperature sintering and fusion) is growing rapidly. General Electric’s Additive Development Center (part of its aviation division) produced a working model (approximately 30 cm in length) of a jet engine fabricated by 3D printing. Of greater importance, GE obtained approval for the commercial introduction of a jet engine nozzle manufactured by 3D printing. According to a news release, they expect more than 100,000 3D printed engine parts to be on the market by 2020 [15]. In a related application, a proposal has been made to 3D-print a metal bridge across a canal in Amsterdam [16]. 3D printing technology also shows potential for fabricating large architectural structures. Contour Crafting [17] is a fabrication process by which large-scale parts can be fabricated quickly in a layerby-layer fashion. Attempts are being made to use fibres to develop printable composite building materials. A private firm in China recently used a 10 m × 6.6 m printer to spray a mixture of cement and

construction waste to fabricate sections of buildings layer by layer (Fig. 3b). Ten full-sized, detached single-story houses were made in one day [18]. A group at the Université de Nantes in France has proposed 3D printing of emergency shelters. The prototype—see the video link in the attached reference—was built out of a polyurethane material, but clearly there is room for other liquid/gel systems that are capable of being extruded [19]. 3D printing may even be used in space. NASA is supporting a revolutionary suite of technologies called “SpiderFab” [20] to enable efficient fabrication in orbit of spacecraft and space-station components that are too large to be transported in current launch vehicles, such as antennas, solar panels, trusses, and other multifunctional structures (Fig. 3c). Any material that can be passed through an extruding nozzle can be printed, including food products and pastes. NASA awarded $125,000 to Systems and Materials Research Consultancy (SMRC) to study how to “print” food during long space missions. The project made headlines largely because of the first item on the menu: a 3D-printed space pizza. Cellulose-based food additives could be included in the recipe for printing ingredients [21]. Main Feedstocks and Requirements

Feedstocks for 3D printing can be categorized into thermoplastics, photopolymers, metal powders (stainless steel, sterling silver), and other powders (glass, ceramics, resin, sandstone, rubber, etc.; Fig. 4). The global materials market in 2013 in total was about 2,000 tons, which is equivalent to $450 million. The materials value

Fig. 3 - Representative objects produced by 3D printers: (a) printed architectural design model; (b) printed house; (c) concept for printed spacecraft component [22].

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

EMERGING AREA CONTRIBUTIONS should exceed $600 million (about 9,700 tons) by 2025 in a development combining increased demand with reduced prices [23,24]. Thermoplastic filaments (Table 2) for FDM 3D printing accounted for 40% of the total feedstock market in 2013. Filament materials for FDM printers that meet customers’ needs for speed, strength, accuracy, surface resolution, chemical and heat resistance, colour, and mechanical properties must continue to be developed. Achieving the desired mechanical, thermal, and chemical resistance properties in a 3D-printed object involves a complex interplay between feedstock material properties and process parameters. Aside from the material properties, filament geometry also plays an important role in 3D printability and quality of the printed objects. Inconsistent filament diameter can lead to extruder failure and inconsistent extruded volume. Challenges in Moving 3D Printing to Manufacturing

Significant technological and business hurdles must be overcome before 3D printing can live up to its most ambitious promises. 3D printing in its current state is very good at recreating geometric and organic complexity. However, 3D printed objects are usually not as durable as traditionally manufactured products. Although a printed wrench is functional, it will not last as long as one produced through drop

TABLE 2 Produced from Properties Extruder temp. Pros

Petroleum-based, deteriorates in sunlight

Cons

PLA Plant starch Tough, strong 160°C–220°C Bioplastic and non-toxic Slow cooling, low heat resistance, easier to break than ABS

PVA (PVOH) * Petroleum Water-soluble, excellent film formation, good barrier properties 190°C–210°C Biodegradable, recyclable, non-toxic Expensive, deteriorates with moisture, special storage necessary

*Polyvinyl alcohol

forging of high-strength steel. Therefore, to reach the point where the average consumer can print “ready-to-go” objects from a home printer, durability must be improved. Another technical challenge for printed objects is the “as printed” look and feel. Almost all 3D printing technologies to date require some level of post-processing (e.g., deburring, sanding, priming, and, airbrushing), although many expect a 3D printed part to look completely smooth and finished when it is done printing. Besides the properties of the printed objects, printing speed is also a hurdle for moving 3D printing to mass manufacturing. Although ideal for mass customization, the technology is not yet capable of mass production. The practical usefulness of 3D printing for a manufacturer is to fabricate 100 airplane parts rather than one million smartphone cases. A new product manufacturing concept, community-driven manufacturing, is

Fig. 4 - Breakdown of the printing materials market in 2013 (total ~2000 tons) [23].

J-FOR

Thermoplastic feedstocks for FDM 3D printers.

ABS Petroleum Durable, strong, slightly ﬂexible, heat-resistant 210°C–250°C Great plastic properties, solidifies quickly, durable and difficult to break

currently emerging. For example, Shapeways is an on-line service platform that enables users to design and upload 3D printable files. The objects are then printed and shipped by Shapeways [25]. 3D Hubs is another on-line 3D printing service that operates a network of 3D printers with over 15,000 locations in 140 countries [26]. FOREST PRODUCTS 3D PRINTING AND BIOPRINTING

Considerable research is ongoing to use wood and forest products as 3D printing feedstocks. These efforts span the range from large-scale building manufacturing to small-scale, highly specialized value-added bio-printing. Through market analysis and technology assessment, FPInnovations is creating a vision of how the forest industry can use 3D printing technologies (Fig. 5).

Fig. 5 - FPInnovations’ vision of 3D printing for the forest industry.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Niche Markets for Speciality Wood Products Use of lignocellulosic materials as fillers in powder and fused-deposition printing - 3D printing of thermoplastic

filaments has grown considerably in recent years, and there is already a demand for “green” substitutes for petroleum-based thermoplastics. Polylactic acid (PLA) is currently the principal biomaterial for fused-deposition 3D printing, but has issues of low strength, low flexibility, poor thermal stability, and poor 3D printability. Biomaterials extracted from forest products can be tuned to address these issues (Table 3). However, their market potential and value remain to be established. Significant efforts world-wide in recent years have led to the production of large quantities of biomaterials in pilot and demonstration plants. However, technical challenges exist in developing feedstock formulations that will match wood fibres, cellulosic biomaterials, or lignin with the most appropriate 3D technology. Work at FPInnovations and elsewhere has shown that these problems can be solved, at least at the scale of proof-of-principle. FPInnovations and Emily Carr University of Art and Design (Vancouver) collaborated to explore 3D printing of powdered lignin using powder-binding technology. The objects in Fig. 6(a) are composed of lignin plus binder. This work is being continued by FPInnovations and includes the development of woodbased filaments for 3D FDM technology. The objects shown in Fig. 6(b) are made from filaments composed of thermoplastic polymers containing lignin filler and

TABLE 3

J-FOR

Potential solution CNC(1), CF(2), NFC(3), and MFC(4) as reinforcing agents

Premature oxidation

Moisture-sensitive

Many synthetic polymers are not compostable

What we know CNC reinforces PLA films [27,28] CNC is compatible with PVA [29,30]

Use lignin fillers in 3D printing

Lignin is a known anti-oxidant (free radical scavenger) [31,32] Lignin is a fire retardant [33]

Functional derivatives of CNC, CF, NFC, and MFC to give water resistance Biopolymers are compostable

The surfaces of CNC, CF, NFC, and MFC can be chemically modified [34,35] N/A

Flammable

(1) Cellulose nanocrystals (2) Cellulose filaments (3) Nanofibrillated cellulose (4) Microfibrillated cellulose

fibre materials. Wood-containing thermoplastic filaments supplied by non-systems manufacturers for fused-deposition printing are already on the market. In 2012, CCProducts created Laywoo-D3 [36], which contains up to 40% recycled wood fibre combined with a thermoplastic polymer binder (Fig. 7a). ColorFabb, produced by Helian Polymers of The Netherlands, offers WoodfillTM, which contains 25%–30% milled wood fibre in a thermoplastic resin (Fig. 7b) [37]. Use of cellulosic materials in bioproducts - As already noted, 3D printing is be-

ing actively developed by medical device manufacturers. In one recent life-saving case, a bronchial “splint” was 3D-printed to help expand the bronchial tubes of an infant. The device was made from a biopolymer (polycaprolactone) and is expected to be gradually absorbed into the child’s body, allowing its own pulmonary system to take over [38]. The ease with which cellulosic derivatives form stable dispersions and gels in aqueous media provides an

Fig. 6 - Developing lignocellulosic-based feedstock for 3D printing at FPInnovations: (a) prototypes made using the powder-binding technology; (b) prototypes made using FDM technology.

Potential improvement for filaments with biomaterials.

Problem with existing filaments Strength, hardness, ﬂexibility, heat resistance

excellent opportunity for new bioproducts. Although removing large amounts of water presents an obvious drawback, this remains an active area of world-wide research. Using a bio-printer, Swansea University in Wales has prepared 3D-printed wound dressings (Fig. 8a) using nanofibrillated cellulose (NFC) supplied by the Norwegian Paper and Fibre Research Institute (PFI) [39]. They described their resulting material as strong, able to be kept under moist conditions, and possessing inherent anti-microbial activity—this last perhaps the result of surface modification during the oxidative process used to prepare the nanofibrils. The Swansea group also used more conventional materials to build a 3D scaffolding that effectively formed a collagen-like structure [40], as shown in Fig. 8b. A different approach was taken by a group in Sweden and involved a dissolving pulp (in an ionic solvent) as the cellulose source [41]. The printed cellulosic gel was then immediately coagulated by overprinting a layer of water. Successive layers could then be built up. Although the solvent used in that particular work is rather exotic, there exists a very extensive

Fig. 7 - Application of wood fibre in 3D printing: (a) Laywoo-D3TM [36]; (b) WoodFillTM developed by ColorFabb [37].

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

EMERGING AREA CONTRIBUTIONS

Fig. 8 - Bioplotted 3D structures: (a) wound dressing with inherent anti-microbial properties, printed from nanofibrillated cellulose (NFC); (b): printed model human ear made from a collagen-like material.

chemistry for dissolving and then regenerating cellulose, which could be applied. Large-scale uses for commodity wood products - 3D printing of building com-

ponents and even entire buildings has gained considerable attention around the world. So far, workers in the field tend to use concrete or related materials in extrusion-type printers [e.g., [17]]. At the same time, forest industry researchers are examining the potential for using wood products in large-scale 3D printed construction. Several difficulties exist with any potential deposition system for slurries, especially with high water content or if fine detail is desired. A slurry: 1. Must have the appropriate rheology to flow out of the nozzle. This may become an issue with higher solids content. 2. Must be solid enough to support the next layer. In other words, it must dry quickly and not collapse. For some structures, forms may help. 3. Must bind well with the next layers printed. 4. Cannot solidify within the chamber or the nozzle. Researchers at the University of Sydney (Australia) recently received funding to develop “micro timber” [42]. The process, which was not fully described, claims to use wood products as one element in

J-FOR

Fig. 9 - Truncated cone generated by 3D printing with spruce chips and gypsum, cut open to show the inner texture [39].

3D panel printing. Potential ingredients include hardwood and softwood fibres, wood flour, and even wood waste. Other details (including binders) have not yet been disclosed. This work can be compared to concrete extrusion printing [e.g., [17]], in which either the concrete could contain wood fibres as a reinforcing component, or the main solid portion of the structure could be wood-based. Concrete as an aqueous slurry is also subject to the requirement that the structure rapidly become self-supporting—certainly before the next layer is printed. The Oxman group of the MIT Media Lab is dedicated to both the art and science of 3D printing. In particular, they have a special interest in both permanent and biodegradable structures. A very recent publication on “water-based engineering” [43] described printing composites made from chitosan and cellulose, among many other materials. Another variation on this theme is to use wood products (e.g., fibres, wood flour, or even wood chips) in a form of powder-bed printing. Just as in powderbed printing with sand or any other material, the wood product would form the “powder bed”, and complex structures would be built up by printing a suitable adhesive layer by layer. Researchers at the Technische Universität München demonstrated a truncated cone (Fig. 9) generated by 3D printing with wooden chips as the bulk material and gypsum, methyl

cellulose, sodium silicate, and cement as the binder [44]. 3D printing has also been proposed for manufacturing packaging and absorbent products by both Golden Hongye Paper and a research group at Shaanxi University of Science and Technology in China. Their concept is to print and form customized paper products with special functional requirements by alternatively depositing fibre layers and adhesive layers onto a build plate according to a digital model. This process is claimed to offer the potential benefits of reduced energy consumption, material waste, and pollution compared to traditional paper production and of a shorter production cycle [45]. SUMMARY AND IMPLICATIONS

In spite of the ongoing debate on whether 3D printing is a game-changer [46], applications of 3D printing are expanding beyond simple prototyping. Niche commercial applications can already be found in the medical, customized fashion, and high-tech industries. 3D printing is ideal for mass customization, but is not yet capable of mass production because of limitations on cost, speed, and materials. The near future of 3D printing will be focussed on production tooling, on-demand parts at low volume, design and educational tools, products made at home, and large, high-value architectural structures. A growing interest in “green” feedstocks has created potential opportunities

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

for wood fibres, other cellulosic materials, and biomaterials. Appearance, toughness, temperature and water resistance, durability, biocompatibility, and castability are among the key attributes of this feedstock. Forest materials with the desired characteristics (flexibility, strength, ease of chemical modification) could provide or enhance these desired feedstock functionalities. In addition to taking advantage of the increasing demand for feedstocks, the forest industry could also create new, high-value products for 3D manufacturing technology, starting with newly developed biomaterials such as CNC, CF, NFC, MFC, and lignin. In the longer term, 3D printing may have potential in producing sophisticated, high-performance absorbent and advanced packaging products. REFERENCES 1.

3. 4.

PriceWaterhouseCoopers LLP, “3D Printing: A Potential Game Changer for Aerospace and Defense”, Available at: http://www.pwc.com/us/ en/industrial-products/publications/ gaining-altitude-with-pwc/issue7-3d-printing.jhtml. 2013; retrieved May 2, 2014. Tsouknidas, A., “Friction-Induced Wear of Rapid Prototyping-Generated Materials: A Review”, Advances in Tribology, 2011. EnvisionTEC, http://envisiontec. com/3d-printers/3d-bioplotter/. Retrieved August 12, 2015. De Boer, J. “Shapeways and the Democratization of Production”. Available at: http://fr.slideshare.net/ jochem/shapeways-the-democratization-of-production-presentation. October 15, 2008; retrieved January 05, 2014. http://www.engr.usask.ca/services/ engineering-shops/facilities/rapidprototyper.php#1, Retrieved January 25, 2014. http://www.sculpteo.com/en/3dprinting/3d-printing-and-traditionalmanufacturing-processes/, Retrieved November 23, 2015. Beaman, J.J. and Deckard, C.R., “Selective Laser Sintering with Assisted Powder Handling”, U.S. Patent

J-FOR

9. 10. 11. 12.

13.

14.

15.

16.

17. 18. 19.

20.

4938816, assigned to the Board of Regents of the University of Texas, July 3, 1990. https://thre3d.com/how-it-works/ powder-bed-fusion/selective-heatsintering-shs, Retrieved January 25, 2014. Sachs, E.M., Haggerty, J.S., Cima, M.J., and Williams, P.A., “Three-Dimensional Printing Techniques”, U.S. Patent 5204055, assigned to Massachusetts Institute of Technology, 1993. http://www.arcam.com/technology/ electron-beam-melting/. Retrieved January 25, 2014. http://www.mcortechnologies.com. Retrieved January 22, 2014. Wohlers Associates, “3D Printing and Additive Manufacturing: State of the Industry”, Wohlers Report, 2014. EnvisionTEC, http://envisiontec. com/envisiontec/wp-content/uploads/MK-MPC-BioplotterPublishedPapers-V01-FN-EN.pdf. Retrieved August 12, 2015. Lux Research, Inc., http://www2. luxresearchinc.com/component/ content/article/164.html. Retrieved January 22, 2014. Hewlett-Packard, http://www8. hp.com/us/en/commercial-printers/ floater/3Dprinting.html, accessed August 5, 2015. GE Reports, http://www.gereports. com/post/118394013625/these-engineers-3d-printed-a-mini-jet-enginethen; http://www.geglobalresearch. com/innovation/3d-printing-createsnew-parts-aircraft-engines. Retrieved August 5, 2015. The Guardian, http://www.theguardian.com/technology/2015/jun/17/ dutch-startup-plans-first-3d-printedsteel-bridge-to-span-amsterdam-canal, retrieved July 23, 2015. http://craft.usc.edu/CC/modem. html, retrieved January 05, 2014. http://www.bbc.com/news/blogsnews-from-elsewhere-27156775, retrieved April 23, 2014. h t t p : / / w w w. u n i v - n a n t e s . fr/1435591545031/1/ficheactualite/ &RH=INSTITUTIONNEL_FR, retrieved, August 14, 2015. http://www.tethers.com/SpiderFab.

21.

22.

23. 24. 25. 26. 27.

28.

29.

30.

31.

32.

33.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

html, retrieved January 05, 2014. http://www.space.com/21308-3dprinting-nasa-space-food. html#sthash.lrROmBJw.dpuf, retrieved January 05, 2014. h t t p : / / w w w. s t r a t a s y s. c o m / r e sources/case-studies/architecture/ rietveld-architects; http://www.bbc. com/news/blogs-news-from-elsewhere-27156775; http://www.tethers. com/SpiderFab.html, retrieved August 12, 2015. h t t p : / / w w w. w i r e d . c o m / d e sign/2012/11/staples-goes-3-d, retrieved January 06, 2014. Kneissl, W., 3D Printing Materials 2014-2025: Status, Opportunities, Market Forecasts. IDTechEX, 2013. www.shapeways.com/. https://www.3dhubs.com/. Lin, N., Huang, J., Chang, P.R., Feng, J., Yu, J., “Surface Acetylation of Cellulose Nanocrystal and its Reinforcing Function in Poly (Lactic Acid)”, Carbohydrate Polymers 83:1834 (2011). Ramirez, M.A., Cellulose Nanocrystals-Reinforced Electrospun Poly (Lactic Acid) Fibers as Potential Scaffold for Bone Tissue Engineering, MSc thesis, North Carolina State University, 2010. Peresin, M.S., Habibi, Y., Zoppe, J.O., Pawlak, J.J., Rojas, O.J., “Nanofiber Composites of Polyvinyl Alcohol and Cellulose Nanocrystals: Manufacture and Characterization”, Biomacromolecules 11:674 (2010). Zhang, W., He, X., Li, C., Zhang, X., Lu, C., Zhang, X., Deng, Y., “HighPerformance Poly (Vinyl Alcohol)/ Cellulose Nanocrystal Nanocomposites Manufactured by Injection Molding”, Cellulose 21:485–494 (2014). Pouteau, C., Dole, P., Cathala, B., Averous, L., Boquillon, N., “Antioxidant Properties of Lignin in Polypropylene”, Polymer Degradation and Stability 81:9 (2003). Dizhbite, T., Telysheva, G., Jurkjane, V., Viesturs, U., “Characterization of the Radical Scavenging Activity of Lignins—Natural Antioxidants”, Bioresource Technology 95:309 (2004). Chiricoa, A., Armaninia, M., Chinib, P., Cioccolob, G., Provasolia, F., and

EMERGING AREA CONTRIBUTIONS

34.

35.

36. 37.

38. 39.

Audisio, G., “Flame Retardants for Polypropylene Based on Lignin”, Polymer Degradation and Stability 79(1): 139–145 (2003). Stenstad, P., Andresen, M., Tanem, B.S., Stenius, P., “Chemical Surface Modifications of Microfibrillated Cellulose”, Cellulose 15:35 (2008). Habibi, Y., Lucia, L.A., Rojas, O.J., “Cellulose Nanocrystals: Chemistry, Self-Assembly, and Applications”, Chemical Reviews 110:3479 (2010). https://www.lulzbot.com/products/ laywoo-d3-laywood-wood-filament3mm, retrieved August 13, 2015. h t t p : / / w w w. 3 d e r s . o r g / articles/20131204-3ders-at-euromold-2013-digifabb-colorfabb.html, retrieved August 13, 2015. http://www.scientificamerican.com/ article/3-d-printed-windpipe/, retrieved August 14, 2015. Gethin, D., Rees, A., Powell, L., Chinga-Carassco, G., Claypole, T., Deganello, D., Hill, K., Thomas, D., and Syverud, K., Studies on the 3D

40. 41.

42. 43.

44.

Printing of Nanocellulose Structures, Adv. Printing Sci. Technol. 41:91 (2014). http://3dprint.com/62259/celleronbioprinting/, retrieved August 14, 2015. Markstedt, K., Sundberg, J., and Gatenholm, P., “3D Bioprinting of Cellulose Structures from an Ionic Liquid”, 3D Printing and Additive Manufacturing.1:115 (2014). http://sydney.edu.au/news/84. html?newsstoryid=15063. Retrieved August 14, 2015. Soldevila, L.M. and Oxman, N., “Water-Based Engineering and Fabrication: Large-Scale Additive Manufacturing of Biomaterials”, Materials Research Society (MRS 2015). Symposium NN Adaptive Architecture and Programmable Matter: Next Generation Building Skins and Systems from Nano to Macro (2015). Henke, K. and Treml, S., “WoodBased Bulk Material in 3D Printing Processes for Applications in

45.

46.

Construction”, Eur. J. Wood Prod. 71: 139–141 (2013). Wang, Y. and Liang, J., CN 103061194, Paper Producing Method and Paper, 2012; Zhu, J. and Liang, J., CN 102989032, Absorbing Core Body and Preparation Method Thereof as well as Absorbent Product Using Same, 2012; Zhao, C., CN 104354297, Method for Producing Paper Package Box by 3D Printing Technology, 2014. Uglow, B., Coe, N., Vig, R., Wood, A., and Davies, R., Capital Goods: 3D Printing, Mogan Stanley Blue Paper, September 2013; Manyika, J., Chui, M., Bughin, J., Dobbs, R., Bisson, P., and Marrs, A., Disruptive Technologies: Advances that will Transform Life, Business, and the Global Economy, McKinsey Global Institute, May 2013; Berman, B., “3-D Printing: The New Industrial Revolution”, Business Horizons, 55

WHY JOIN PAPTAC’s TECHNICAL COMMUNITIES Sharing information on specific topics & challenges facing the Canadian pulp and paper industry. Accessing an exclusive Canadian technical pulp and paper network. Continuing to learn from your peers, identifying and developping new problem-solving solutions. Being aware of the latest technological advancements and innovations. Greater value derived from participating in PAPTAC events (PaperWeek, PACWEST, conferences, webinars, etc.). The technical communities have been formed among PAPTAC members who are interested in a particular aspect of the industry. Members discuss practical and scientific developments in their area of specialty. They work on projects to learn more about the industry, and then disseminate this information to the industry through technical presentations and reports.

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

ENVIRONMENTAL EVALUATION OF DURAPULP BIO-COMPOSITE USING LCA: COMPARISON OF TWO APPLICATIONS ABSTRACT

FRIDA HERMANSSON, MATTY JANSSEN*, FREDRIK GELLERSTEDT Bio-composites are materials made up of a bio-fibre and a polymer, the latter being either bio-based or fossil-based, and are developed for various reasons, for example to decrease the use of non-renewable resources such as crude oil. DuraPulp, a bio-composite of wood pulp and poly-lactic acid (PLA) developed by Södra in Sweden is such a material. The goals of this study were to evaluate the main environmental impacts arising from the different life cycle phases of DuraPulp using life cycle assessment (LCA) to compare its impacts with those of several other composites and to evaluate two applications of the material from an environmental point of view. These two applications are a car door panel (a product with a long lifespan) and a packaging unit (a product with a short lifespan). The results show that bio-composites have a smaller environmental impact than composite materials made of 100% fossil-based plastics. The results also indicate that the packaging unit is the preferred application from an environmental point of view. Furthermore, the main contributors to the impact depend on the application. Lastly, the results indicate that replacing PLA with bio-polyethylene lowers the environmental impact of DuraPulp regardless of its application.

INTRODUCTION

Due to structural changes in the pulp and paper market, with digital media replacing paper in many segments, the forest industry is looking for other applications for the pulp it produces. One of these applications is bio-composites, in which the properties of wood fibres contribute to their desirable properties. A bio-composite is commonly defined as a material consisting of a bio-fibre and a polymer matrix that can be either fossil-based or bio-based. The main advantages of using renewable bio-based matrices rather than fossil-based ones are the decrease of dependency on non-renewable, fossil-based polymers with similar properties, the minimization of non-degradable plastic waste (note that not all biopolymers are biodegradable), and the reduction of carbon dioxide emissions [1]. The rewards of choosing natural fibres over fossil-based fibres are the CO2 neutrality of natural fibres and the absence from the processing environment of agents causing skin and respiratory problems for the employees [2]. It is nevertheless important to investigate the environmental impacts

J-FOR

of bio-composites in their various applications, as well as to explore the advantages and drawbacks of using a renewable polymer rather than a fossil-based one. Södra Skogsägarna Ekonomisk Förening is a forest company based in the south of Sweden. Sawn timber and market pulp are the two most important products in the Södra product portfolio. Due to the structural changes in the pulp and paper market, Södra has developed a unique product called DuraPulp, a bio-composite

FRIDA HERMANSSON

Environmental Systems Analysis, Department of Energy and Environment, Chalmers University of Technology, SE-41296 Göteborg, Sweden

based on pulp and polylactic acid (PLA), which is a renewable and biodegradable biopolymer produced mostly from corn. DuraPulp can be used in various product applications with different lifespans. An example of a short-lifespan product is packaging material for high-end goods, where DuraPulp is substituted for a pure polymer material such as polystyrene or polypropylene. A long-lifespan product could be a car part such as a door panel (the component protecting the electronics

MATTY JANSSEN

FREDRIK GELLERSTEDT Environmental Systems Södra Skogsägarna Analysis, Department of Ekonomisk Förening Energy and Environment, (Södra Innovation), Chalmers University of SE-43286 Technology, SE-41296 Väröbacka, Göteborg, Sweden Sweden *Contact: mathias.janssen@chalmers.se

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

EMERGING AREA CONTRIBUTIONS on the inside of the car door), where DuraPulp could replace pure plastic components and conventional composites with fossil-based fibre. Currently, packaging products manufactured from bio-based material (excluding traditional fibre-based packaging such as cardboard) make up 2% of the market [3]. Life cycle assessment (LCA) can be used to explore the environmental advantages and drawbacks of using biocomposites. LCA is an analytical tool for assessing the environmental aspects and potential impacts associated with a product throughout its life cycle [4]. The first step of an LCA study is the goal and scope definition, in which the purpose of the study and its modelling aspects are described (Fig. 1). This is followed by the life cycle inventory analysis, in which the inputs (resources used) and outputs (emissions) of the processes included are identified. These inputs and outputs are then classified and characterized to calculate the life cycle environmental impact of the product or service. Finally, the results are interpreted. An LCA study of coffee packaging alternatives, which included a composite material, was performed by De Monte et al. [5]. The results of the study showed that choosing plastic polylaminate packaging over metallic cans decreases the environmental burden. Another study focussing on packaging material was performed by Xie et al. [6]; it compared a composite material with 100% polyethylene. The results of the study showed that the composite packaging has a slightly higher environmental impact. The main reasons for this higher impact are the lack of reuse of the composite material and of proper recycling processes. Alves et al. [7] conducted a study that compared the environmental performance of car parts manufactured from a conventional composite using glass fibre with those made from a bio-composite using jute fibre. The conclusion was that the bio-composite enhances the environmental performance of the car by reducing its weight, which leads to lower fuel use during operation. Luz et al. [8] compared the

J-FOR

environmental benefits of replacing talc with sugarcane bagasse in polypropylene (PP) composites for use in the automotive industry. Among the conclusions were that: 1) the growing of sugarcane in the cultivation phase offsets the global warming potential due to uptake of CO2, 2) the production process is cleaner than before, and 3) the bagasse/PP bio-composite weighs less than the talc/PP composite at equivalent performance, which reduces the use of gasoline. The literature indicates that replacing pure plastic with a bio-composite in a product generally decreases the environmental impacts of the product. However, a good enough recycling scheme for bio-composites has still not been developed, which may be detrimental for their application from an environmental point of view [6]. Such a recycling scheme would further lower the environmental impact of bio-composites. Recycling of bio-composites differs from recycling of pure fibre or 100% plastic products: the different properties of the bio-composite components must be taken into account, as well as how these are changed when the bio-composite is recycled. Soroudi and Jakubowicz [9] reviewed the possibilities for recycling biopolymers and bio-composites and concluded that mechanical recycling is preferable, although research into recycling bio-based materials is still at a preliminary stage. Xie et al. [10] attempted to determine the best waste management strategy for composite packaging waste in China. Their results showed that this depends on policy goals, e.g., energy savings or emissions reduction. Piemonte [11] performed a study investigating different end-of-life treatments for two different starch-based biopolymers. The results of this study showed that mechanical recycling always leads to a decrease in environmental impacts. Applications of bio-composites can currently be found in the automotive and packaging industries. However, the makeup of the bio-composites differs between these applications. A desirable outcome of LCAs that compare different products made from the same bio-composite, but

with different lifespans, would be to evaluate for which product this bio-composite would be better used, e.g., a product with a long lifespan but without recycling possibilities, or a product with a short lifespan but with recycling possibilities. Therefore, methodological options for performing LCAs to compare different applications of the same material are needed, for example choosing an appropriate functional unit and a reference flow that enable such comparisons. Furthermore, information on recycling of bio-composites is scarce. In addition to more fundamental research on bio-composite recycling, more LCA studies need to be done on possible future scenarios for recycling these materials to make informed decisions about recycling of bio-composites. The purpose of this study was to determine the environmental sustainability of different products with different life spans, but all made from DuraPulp, using life cycle assessment. Specifically, the goals of this study were to: 1. Compare DuraPulp to a range of alternative composite materials. 2. Investigate the main contributors to the environmental impacts of DuraPulp. 3. Identify the preferred use of DuraPulp from an environmental point

Fig. 1 - Framework for life cycle assessment (LCA).

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

of view by comparing a short- and a long-lifespan product, represented by a packaging unit and a car door panel respectively. 4. Examine the environmental impacts associated with recycling the shortlifespan product.

material was placed in an open, heated cavity. This cavity was closed, and heat and pressure were applied until the material had cured and solidified [12]. The electricity consumption in this phase was 4.5 kWh/kg of moulded material. Set-up of the LCA

The results of the study are intended to provide the stakeholders at Södra with the environmental highlights of the intended short- and long-lifespan products and to help them decide where to focus further efforts towards development of DuraPulp products. METHODS Description of the Systems Analyzed

The LCAs for the car door panel and the packaging unit were carried out from cradle to grave, including raw material extraction, composite manufacturing, product use, and end-of-life treatment (Fig. 2). The attributional life cycle assessment methodology was chosen due to the comparative nature of this study. In the first manufacturing step, the bio-composites, consisting of polymer in the form of fibre and pulp (northern bleached softwood Kraft, total chlorinefree bleaching), were mixed together with chemical addition by dispersing the polymer fibre in a tank of pulp fibres. All the pulp fibres were assumed to be produced from Swedish wood extracted in the vicinity of Södra Cell Värö (in Väröbacka, approximately 60 km south of Göteborg, Sweden), where the pulp is manufactured. The pulp manufacturing process was assumed to be the same for all types of bio-composites (using either biopolymers or fossil-based polymers). Next, the mixture was dried in a counter-current flash dryer and shaped into bales. The make-up of the bio-composite was 70% pulp fibre and 30% polymer. Next, the bio-composite was converted into sheet form before entering the compression moulding process, where the material was shaped into the product (car door panel or packaging unit). During compression moulding, the

J-FOR

DuraPulp is an already commercialized bio-composite, and a range of other composite materials were identified as alternatives for producing the car door panel and the packaging unit. Pure virgin polypropylene was considered as the reference material for the car door panel. For the packaging unit, the reference material was expandable polystyrene. The 70% pulp fibre and 30% polymer alternatives included PLA and other polymers (bio-polyethylene and fossil-based polypropylene). Wood-flour plastic composite (WPC) and fibre plastic composite (FPC) were identified by Södra as the main competing materials for the car door panel with a fibrebased bio-composite and were therefore included in the comparison. The composite materials considered when analyzing the packaging unit were: a) DuraPulp: PLA (D:PLA) (70% pulp fibre and 30% polylactic acid); b)

DuraPulp: PLAR (D:PLAR) (70% pulp fibre and 30% polylactic acid with renewable energy credits included [13]); c) pulp mix: bio-PE (PM:PE) (70% pulp fibre and 30% bio-polyethylene); d) pulp mix: PP (PM:PP) (70% pulp fibre and 30% fossil-based polypropylene); e) pure virgin polypropylene (PP:P) (100% fossil-based polypropylene); and f) pure virgin expandable polystyrene (PS) (100% fossil-based polystyrene). The composite materials considered when analyzing the car door panel were: a) DuraPulp: PLA (D:PLA) (70% pulp fibre and 30% polylactic acid); b) DuraPulp: PLAR (D:PLAR) (70% pulp fibre and 30% polylactic acid with renewable energy credits included [13]); c) pulp mix: bio-PE (PM:PE) (70% pulp fibre and 30% biopolyethylene); d) pulp mix: PP (PM:PP) (70% pulp fibre and 30% fossil-based polypropylene); e) pure virgin polypropylene (PP:P) (100% fossil-based polypropylene); f) WPC (40% wood flour and 60% fossil-based polypropylene); and g) FPC (50% pulp fibre and 50% fossil-based polypropylene). Two functional units were used in this study, one for the product with a long lifespan (car door panel), and one for the

Fig. 2 - Simplified flowchart for the life cycle of: a) the car door panel and b) the packaging unit made from bio-composite. The red dashed line indicates that the use phase is excluded in the case of the packaging unit. Note that in the case of 100% fossil-based materials, tree cultivation and harvesting are not included.

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

EMERGING AREA CONTRIBUTIONS product with a short lifespan (packaging unit). For the long-lifespan product, the functional unit was chosen to be one car door panel, and for the short-lifespan case, it was set to one unit of packaging (a high-end food packaging tray). The reference flow for the car door panel was 1 dm3 of composite material, corresponding to the assumed car door panel dimensions of 1m by 0.5m by 0.002m, with a lifespan of 17 years (the average lifespan of a car in Sweden) [14]. The reference flow for the packaging unit was 0.026 dm3 of composite material, corresponding to the volume of a packaging unit manufactured using DuraPulp or the other composites. The lifespan of the packaging unit was set to one month. Next, the reference flow was normalized to one year for both products. Normalization was performed to take into account the lower input of raw material for the long-lifespan product and the increased input of material for the short-lifespan product due to a higher product turnover rate. This illustrates the different requirements for raw material extraction and inputs depending on the bio-composite application and the product lifespan and enables quantitative comparison of two applications of the same material. The life cycle impact assessment (LCIA) was carried out using the CML characterization method [15]. For this study, the following impact categories were considered: global warming potential (GWP), eutrophication potential (EP), acidification potential (AP), photochemical ozone creation potential (POCP), and depletion of abiotic resources potential (DAP). Global warming potential is strongly connected to manufacturing processes and fuel use (in the case of the car door panel), the eutrophication potential of fertilizing raw materials for biopolymer production (in this study, corn for production of PLA and sugar cane for production of bio-PE), as well as acidification potential, photochemical ozone creation potential, and depletion of abiotic resources due to the energy used during manufacturing as well as in the use phase. Although DuraPulp is a material made from entirely

J-FOR

renewable components, depletion of biotic resources, impacts due to land use and land-use change, and impact of biogenic CO2 emissions were not considered in this study. Additional Modelling Choices

Manufacturing of the pulp and bio-composite, the use phase, and end-of-life treatment of the bio-composite were assumed to take place in Sweden. The electricity used by these processes was assumed to be generated in Sweden. The manufacturing and raw material extraction of the polymers were assumed to take place in various locations around the world, e.g., PLA is manufactured in the United States and bio-polyethylene (bio-PE) in Brazil. Whenever site-specific data were not available, information was gathered from the literature and from inventory data collected in the ecoinvent database [16]. When manufacturing WPC and FPC, 10% of the output is returned to the process input. The pure polypropylene, WPC, and FPC composites are moulded using the injection moulding technique, which consumes 3.4 kWh per kg moulded material [17]. The car door panel was assumed to be incinerated. This assumption was based on current car dismantling practices [18]. The input of gasoline in the use phase was allocated based on the mass of the car door panel (i.e., the amount of gasoline needed to transport one car door panel) (Fig. 2a). At its end of life, the packaging unit was assumed to be sent to incineration, and when possible to anaerobic digestion. To identify clearly the differences arising from various waste management systems for the packaging unit, it was assumed that either all waste was incinerated or all was sent to anaerobic digestion. The use phase and the transportation of the packaging unit were not included based on the assumption that the environmental impacts are allocated to the goods being packaged rather than the packaging (Fig. 2b). Recycling of the Packaging Unit

A future scenario in which the packaging unit is recycled was also considered. When recycling DuraPulp and pulp mixes,

Fig. 3 - Recycling process for the packaging unit (short-lifespan product).

polymers are added (Fig. 3). DuraPulp is wet-laid, but this cannot be done when recycling the DuraPulp because the thermoplastic fibres have melted and turned into a matrix. Therefore, recycling must be done by extrusion. Because it is impossible to process a material with 70% fibre content using an extruder, more matrix material must be added to the biocomposites to reach approximately 40% fibre content. The polymer added in the recycling process is of the same kind as in the original matrix. There is no recycling of expanded polystyrene due to the lack of proper recycling schemes [19], and therefore it is assumed that polystyrene is not re-used, but incinerated. RESULTS AND DISCUSSION

One of the main differences among the various composite materials is the amount of polymer used in the material, ranging from 30% in DuraPulp and the pulp mix materials (PM:PE and PM:PP) to 100% in the pure virgin plastic alternatives. Another difference is the input of electricity, which varies according to the choice of moulding technique and where the polymer is manufactured. Furthermore, the product mass changes with the material used due to differences in material density (Tables 1 and 2).

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

TABLE 1

Density of the materials used to manufacture the packaging unit.

Functional Unit: Packaging unit DuraPulp: PLA/PLAR (D:PLA/PLAR) Pulp mix: PP (PM:PP) Pulp mix: bio-PE (PM:PE) Compression moulding pure PP (PP:P) Polystyrene (PS)

Results for the Packaging Unit

The pure polypropylene and expandable polystyrene packaging units (the latter is the reference material for this application) have the highest environmental impact in terms of global warming potential (GWP), photochemical ozone creation potential (POCP), and depletion of abiotic resources (DAP) because they are based on fossil resources only (Fig. 4). The local electricity mix contributes to the GWP and DAP impact categories (Figs. 4a and 4e). The difference between the impacts for these categories for DuraPulp: PLA (D:PLA) and DuraPulp: PLAR (D:PLAR) is that manufacturing D:PLAR results in lower greenhouse-gas emissions and decreased use of non-renewables because renewable energy is bought in quantities equivalent to the electricity used in the manufacturing process [13]. The packaging units made of bio-

Density [g/dm3] 755 630 630 909 755 (assumed)

TABLE 2

Density of the materials used to manufacture the car door panel.

Functional Unit: Car door panel DuraPulp: PLA/PLAR (D:PLA/PLAR) Pulp mix: PP (PM:PP) Pulp mix: bio-PE (PM:PE) Compression moulding pure PP (PP:P) WPC FPC

composites with bio-based polymers (DuraPulp and the pulp mix with bio-PE) have a higher eutrophication potential because of the fertilizers used in cultivation of corn (in the case of PLA) and sugar cane (in the case of bio-PE) (Fig. 4b). When comparing the composites that use bio-based polymers, the bio-PE composite has an equal or lower eutrophication potential than the PLA composite (D:PLA and D:PLAR). Bio-PE is produced in Brazil from ethanol through its dehydration to ethylene. To produce 1 kg of bio-PE, 21.5 kg of sugar cane is needed [20,21]. This amount of biomass is significantly greater than the 2.5 kg of corn needed to manufacture 1 kg of PLA [22]. However, this difference in renewable resource use is compensated for by the difference in density between DuraPulp and the pulp mix:bio-PE composite and by the fact that sugar cane cultivation is much less

intensive in terms of fertilizer use [23]. The choice of end-of-life treatment is important for the DuraPulp alternatives. On the one hand, DuraPulp that is sent to anaerobic digestion has a lower global warming potential than DuraPulp that is incinerated (Fig. 4a). On the other hand, DuraPulp that is sent to anaerobic digestion has a slightly higher acidification potential than DuraPulp that is incinerated because acids are formed by bacteria in the anaerobic process (Fig. 4c) [24]. Due to its attributional nature, this study did not consider the environmental benefits of replacing combustion of conventional (fossil) fuel with incineration of DuraPulp. Moreover, this study did not account for replacing fossil methane with biogas. The higher acidification potential for the pulp mix:bio-PE (PM:PE) packaging unit compared to the pulp mix: PP (PM:PP) packaging unit is due to the

Fig. 4 - Environmental impacts for packaging units made from the various material options.

J-FOR

Density [g/dm3] 755 630 630 909 715 753

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

EMERGING AREA CONTRIBUTIONS increased acidic emissions associated with ethanol fermentation and transportation across the Atlantic Ocean (Fig. 4c). D:PLAR has a higher photochemical ozone creation potential than D:PLA because of the natural gas used in the D:PLAR manufacturing phase (Fig. 4d). The packaging unit made of expanded polystyrene has the highest impact in this category due to the use of benzene and other chemicals in the polystyrene manufacturing phase. Note that polypropylene and polystyrene were produced using the best available technologies, whereas the biopolymer (bio-polyethylene and polylactic acid) production processes are not yet fully mature. Further development and optimization of these processes may therefore lead to lower environmental impacts. Results for the Car Door Panel

The results for the car door panel follow the same trend for all impact categories, with the life cycle impacts dominated by the use phase due to the fuel used when driving the car (Fig. 5). The mass of the car door panel has a significant influence on the fuel input in the use phase. The greater the mass of the car door panel, the

Relative impact of the long-lifespan product (car door panel) and the short-lifespan product (packaging unit) made of DuraPulp: PLA TABLE 3 (incinerated) and pure polypropylene. If the relative impact > 1, then the DuraPulp product has a lower impact. Impact category Global warming potential (GWP) Eutrophication potential (EP) Acidification potential (AP) Photochemical ozone creation potential (POCP) Depletion of abiotic resources potential (DAP)

more fuel is needed to drive the car. The use of a pure virgin polypropylene (PP:P, the reference material for this analysis) door panel has the highest impact in all categories due to its high density and its fossil-based nature. The car door panels made of the PM:PE and PM:PP materials had the lowest environmental impact due to their low densities. Furthermore, the environmental impacts of using car door panels made of DuraPulp and those made of WPC and FPC materials were similar, which can mostly be explained by their similar densities. The manufacturing phase and the end-of-life treatment contribute only slightly to the environmental impact. Even if the door panel were to be used in an average European electric car, which would

Car door panel 1.31 1.20 1.23 1.25 1.35

Packaging unit 2.84 0.91 1.44 2.84 4.56

significantly lower the global warming potential, the impact due to the use phase would still overwhelm the impacts due to the other life cycle phases [25]. Comparison of LCIA results

To establish the preferable bio-composite application from an environmental point of view, the LCIA results for the packaging unit (the short-lifespan product) and the car door panel (the long-lifespan product) were compared. This was done by calculating the ratios of the environmental impacts of the two types of products made from the bio-composites (DuraPulp: PLA and the pulp mix: bio-PE composites) and from the 100% fossil-based polypropylene composite:

Fig. 5 - Environmental impacts of car door panels made from the various material options.

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Relative impact of the long-lifespan product (car door panel) and the short-lifespan product (packaging unit) made of pulp mix: bio-PE TABLE 4 and pure polypropylene. If the relative impact > 1, then the pulp mix: bio-PE product has a lower impact. Impact category Global warming Potential (GWP) Eutrophication potential (EP) Acidification potential (AP) Photochemical ozone creation potential (POCP) Depletion of abiotic resources potential (DAP)

The overall results show that changing the material used in the packaging unit from 100% polypropylene to a biocomposite leads to the greatest reduction in environmental impacts (Tables 3 and 4). However, the eutrophication potential increases for the packaging unit for both bio-composites because the fertilizers used in the cultivation phase of the raw materials (corn and sugarcane) have a large impact. This is not the case for the car door panel because its impact is determined mainly by the use phase, which is determined by the mass of the car door panel. This mass is lower when the panel is made of bio-composite materials. The results (Tables 3 and 4, Figs. 4 and 5) also show that using bio-PE instead of PLA as the polymer matrix in the bio-composite material leads to an equal (EP and POCP of the packaging unit) or lower environ-

Car door panel 1.34 1.41 1.45 1.48 1.63

Packaging unit 3.90 0.91 1.56 2.84 11

mental impact. From an environmental point of view, SĂśdraâ&#x20AC;&#x2122;s DuraPulp product may therefore benefit from changing the polymer matrix from PLA to bio-PE. However, the availability of bio-PE on the market is still low and may not be as sustainable economically. Recycling of the Packaging Unit

The end-of-life phase contributes significantly to the global warming potential of the packaging unit (Fig. 4a). Therefore, its recycling may decrease its environmental impact. The global warming potential of 100% recycling (all waste is collected and sent to recycling), 50% recycling, and 0% recycling of the packaging unit was assessed (Fig. 6). The impacts were calculated by dividing the cumulative global warming potential for each collection rate with the resulting number of uses of the

material at the given collection rate [26]. When increasing the collection rate, the cumulative impact (the impact of the total number of uses) increases. This is due to the addition of virgin polymers to the recycling process after the first use (see Fig. 3): the higher the collection rate, the more new polymer must be added to achieve the required material composition. This suggests a trade-off between collection rate and environmental impact. However, this trade-off does not exist when considering the impact per use of the packaging unit (Fig. 6): the higher the collection rate, the lower is the environmental impact. The global warming potential is higher for recycled fossil-based polymers, which is related to the input of virgin polymers into the recycling process (this argument is also valid for AP, POCP, and DAP, but the results are not shown here). In the case of eutrophication potential (not shown), the biopolymers have a higher eutrophication potential than the composites based on fossil-based polymer and 100% polypropylene. This higher value is due to the fertilizers used in the raw material cultivation phase.

Fig. 6 - Global warming potentials for 0%, 100%, and 50% recycling. The total impacts for each case were divided by the total number of uses for each collection rate. The number of uses was calculated according to ISO/TR 14049 [26].

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

EMERGING AREA CONTRIBUTIONS

CONCLUSIONS

This study was intended to investigate the environmental impacts of the bio-composite DuraPulp and of alternative composite materials for two applications, a car door panel and a packaging unit. This was done using life cycle assessment. DuraPulp and the alternative biocomposites have a lower environmental impact than composites that consist 100% of fossil-based material in both applications. DuraPulp’s environmental impact can be further decreased by changing the polymer matrix from polylactic acid (PLA) to bio-polyethylene (Bio-PE). In the case of the car door panel, this decrease is mostly due to the resulting lower density of the material, which consequently leads to lower fuel consumption during the use phase. In the case of the packaging unit, this impact is mostly due to lower fertilizer use during sugar-cane cultivation. The main contributors to the environmental impact of using DuraPulp (and all alternative composite materials) depend on the application. In the case of the car door panel (the long-lifespan product), the use phase mainly determines its environmental impact, which depends on the mass of the door panel. In the case of the packaging unit, the manufacturing phase and end-of-life treatment are the main contributors. This implies that efforts to reduce DuraPulp’s environmental impact depend on its application and may not or cannot be directly influenced by its producer. The packaging unit is the preferred application of DuraPulp (and of the pulp mix: bio-PE composite material) from an environmental point of view according to this study. Furthermore, a recycling scheme for DuraPulp packaging, which is currently non-existent, and a recycling process that adds a minimum amount of virgin polymer will further decrease the environmental impact of this DuraPulp application. The results also suggest that if the packaging unit has been recycled once, it is preferable for it to be recycled again to compensate for the impacts of the virgin polymers added in the first recycling.

J-FOR

Techno-economic assessment and assessment of the physical and mechanical properties of the composite materials under study may be included in future work to help stakeholders make better and more sustainable decisions regarding composite material design and applications. REFERENCES 1.

2. 3.

Pilla, S. Handbook of Bioplastics and Biocomposites Engineering Applications, 1st ed., Scrivener, Salem, MA, USA (2011). John, M.J. and Thomas, S., “Biofibres and Biocomposites”, Carbohydrate Polymers 71(3):343-364 (2008). Johansson, C., Bras, J., Mondragon, I., Nechita, P., Plackett, D., Šimon, P., Svetec, D.G., Virtanen, S., Baschetti, M.G., Breen, C., Clegg, F., and Aucejo, S., “Renewable Fibers and Bio-Based Materials for Packaging Applications—A Review of Recent Developments”, Bioresources 7(2):2506-2552 (2012). International Organization of Standardization, Environmental Management—Life Cycle Assessment—Principles and Framework (ISO 14040) (2006). De Monte, M., Padoano, E., and Pozzetto, D., “Alternative Coffee Packaging: An Analysis from a Life Cycle Point of View”, Journal of Food Engineering 66(4):405-411 (2005). Xie, M., Li, L., Qiao, Q., Sun, Q., and Sun, T., “A Comparative Study on Milk Packaging Using Life Cycle Assessment from PA-PE-Al Laminate and Polyethylene in China”, Journal of Cleaner Production 19(17):21002106 (2011). Alves, C., Ferrão, P.M.C., Silva, A.J., Reis, L.G., Freitas, M., Rodrigues, L.B., and Alves, D.E., “Ecodesign of Automotive Components Making Use of Natural Jute Fiber Composites”, Journal of Cleaner Production 18(4):313-327 (2010). Luz, S.M., Caldeira-Pires, A., and Ferrão, P., “Environmental Benefits of Substituting Talc by Sugar Cane Bagasse Fibers as Reinforcement in Polypropylene Composites: Ecode-

10.

11.

12.

13.

14. 15.

16. 17.

18.

sign and LCA as Strategy for Automotive Components”, Resources, Conservation, and Recycling 54(12):11351144 (2010). Soroudi, A. and Jakubowicz, I., “Recycling of Bioplastics, Their Blends and Biocomposites: A Review”, European Polymer Journal 49:2839-2858 (2013). Xie, M., Qiao, Q., Sun, Q., and Zhang, L., “Life Cycle Assessment of Composite Packaging Waste Management—A Chinese Case Study on Aseptic Packaging”, The International Journal of Life Cycle Assessment 18:626-635 (2013). Piemonte, V., “Bioplastic Wastes: The Best Final Disposition for Energy Saving”, Journal of Polymers and the Environment 19(4):988-994 (2011). Harper, C.A. and Petrier, E.M., Plastics Material and Processes: a Concise Encyclopedia. Wiley: Hoboken, NJ (2003). Vink, E.T., Glassner, D.A., Kolstad, J.J., Wooley, R.J., and O’Connor, R.P., “The Eco-Profiles for Current and Near-Future NatureWorks® Polylactide (PLA) Production”, Industrial Biotechnology 3(1):58-81 (2007). Ekerot, S., “Stålets kretslopp”, Report no. D792, Jernkontoret, Stockholm, Sweden (2003). Guinée, J.B., Gorrée, M., Heijungs, R., Huppes, G., Kleijn, R., Koning, A. de, Oers, L. van, Wegener Sleeswijk, A., Suh, S., Udo de Haes, H.A., Bruijn, H. de, Duin, R. van, Huijbregts, M.A.J., Handbook on Life Cycle Assessment: Operational Guide to the ISO Standards, Kluwer Academic: Dordrecht, The Netherlands (2002). Ecoinvent, http://www.ecoinvent.org (2013). Thiriez, A. and Gutowski, T., “An Environmental Analysis of Injection Moulding”, Proceedings of the 2006 IEEE International Symposium on Electronics and the Environment, 195-200 (2006). Jensen, C., Felix, J., Ljunggren-Söderman, M., Alongi, S., and Skenhall, T.R., “Utvärdering av förändrad demontering och återvinning av uttjänta fordon i Sverige”, Report no. B2068, IVL, Göteborg, Sweden (2012).

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

19.

20.

21.

Tan, R.B. and Khoo, H.H., “Life Cycle Assessment of EPS and CPB Inserts: Design Considerations and End of Life Scenarios”, Journal of Environmental Management 74(3):195205 (2005). Monteiro, J.M., Veiga, L.B.E., and Coutinho, H.L., “Sugarcane Crop for Biofuel Production, Demand on Soil Resource and Food Security in Brazil”, Proceedings of the 19th World Congress of Soil Science, Brisbane, Australia (2010). Liptow, C. and Tillman, A.-M., “A

22. 23.

24.

Comparative Life Cycle Assessment Study of Polyethylene Based on Sugarcane and Crude Oil”, Journal of Industrial Ecology 16:420–435 (2012). Kingsland, C., PLA: A Critical Analysis, Institute of Packaging Professionals (2010). Luo, L., van der Voet, E., and Huppes, G., “Life Cycle Assessment and Life Cycle Costing of Bioethanol from Sugarcane in Brazil”. Renewable and Sustainable Energy Reviews 13:16131619 (2009). van Haandel, A. and van der Lubbe,

25.

26.

J., Sludge Treatment I: Handbook of Biological Wastewater Treatment. Quist: Leidschendam, The Netherlands (2007). Hermansson, F., Environmental Evaluation of Bio-Composites using LCA, Master’s thesis, Chalmers University of Technology (2013). ISO, Environmental Management— Life Cycle Assessment—Examples of Applications of ISO 14041 to Goal and Scope Definition and Inventory Analysis, 1st ed., BSI, London, UK (2000).

PAPTAC's JOB ANNOUNCEMENT SERVICE! The “Job Announcement Service” section of PAPTAC’s Website provides members priority access to local and national job offers in the pulp and paper industry. This section is updated as soon as a new job announcement is submitted to us. Simply click on the job posting you are interested in to read the details, and directly send your resume to the contact person. You wish to post a job offer on PAPTAC’s Website? This feature is another benefit for PAPTAC Sustaining Members and Member mills, which have the opportunity to post as many job offers as they want for the time they need. Send us all the information at tech@paptac.ca in order to update the section with the position you need to fill.

If you are not a PAPTAC Sustaining member but you wish to advertise a job offer, contact us to know more on terms and fees.

J-FOR

Journal of Science & Technology for Forest Products and Processes: VOL. 5, NO. 2

Building Buildingfor forthe the and Paper NewPulp Pulp and Paper Community Community

FOR THE ADVANCEMENT OF THE FOREST INDUSTRY

J-FOR Journal of Science & Technology for Forest Products and Processes contains papers which are the property of the Pulp and Paper Technical Association of Canada (PAPTAC). Papers may not be reprinted or reproduced without permission. These papers have been peer reviewed and approved for publication by the Editorial staff. J-FOR is published bi-monthly by PAPTAC, 740 Notre Dame West, Suite 1070, Montreal, QC, Canada H3C 3X6, to which all general correspondence should be addressed. J-FOR Journal of Science & Technology for Forest Products and Processes is indexed and/ or abstracted with CAS (Chemical Abstracts Service, a division of the American Chemical Society), Thomson Reuters, and ProQuest. Vol. 5 subscriptions for a printed copy and on-line access are: Canada (includes all applicable taxes) – $850.00 Cdn; USA – $850.00 Cdn; Overseas – $875.00 Cdn. No part of this Journal may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, without the prior written permission of PAPTAC or, in case of photocopying or other reprographic copying, a licence from Access Copyright (the Canadian Copyright Licensing Agency), 1 Yonge Street, Ste. 1900, Toronto, ON, Canada M5E 1E5. ISSN 1927-6311 (Print) ISSN 1927-632X (Online) Printed in Canada

Some of the photos in the Journal provided courtesy of FPAC.

J-FOR A PAPTAC JOURNAL

JOURNAL OF SCIENCE & TECHNOLOGY FOR FOREST PRODUCTS AND PROCESSES

Unpack the power of Maximyze for packaging. ®

Let Buckman help you improve sheet strength and increase productivity. Buckman announces new Maximyze enzymatic technology for recycled packaging. It can significantly improve sheet strength and drainage, so you can increase machine speeds. With a customized Maximyze program you can reduce fiber costs, steam consumption, transportation costs and your environmental footprint too. No wonder it’s an EPA Presidential Green Chemistry Challenge Award winner! Find out more. Contact your Buckman representative or visit buckman.com. Better drainage Production on a recycled linerboard machine was limited due to drainage. Buckman’s Maximyze application improved drainage, so machine speeds could be increased by as much as 100 mpm. Steam use was reduced 8%, and CO2 emissions were reduced by 1806 metric tons per year. Reduced energy A core and tube producer wanted to increase production, have greater flexibility in its fiber selection and reduce energy use. Buckman applied Maximize to the pulper, which conditioned the fiber faster with less refining energy, preserving fiber strength. Speed increased 10%. Refining energy decreased 30%. And tensile strength increased from 20 to 26 kgf/15mm.