Scalable Analytics for Extracting Control Insights from Historical Process Data: with Applications in the Pulp and Paper Industry

几十年来，林产品工业一直是加拿大经济的推动力。该行业的活动帮助创造了数千个就业机会、新的社区、优秀的基础设施和世界上任何地方都无法超越的生活方式。该行业占加拿大制造业总出货量和国内生产总值的最大份额。近年来，该行业面临着许多新的挑战，包括来自国外生产的低成本材料的激烈竞争，以及旨在降低温室气体排放和提高环境可持续性的政府法规的增加。因此，加拿大的木制品行业目前正在经历其历史上最严重的经济衰退。

The forest products industry has been a driving force in Canada's economy for decades. Activity in this sector has helped create thousands of jobs, new communities, excellent infrastructure and a lifestyle that is unsurpassed anywhere in the world. This industry accounts for the largest portion of Canada's total manufacturing shipments and gross domestic product. In recent years, this industry has faced a number of new challenges, including stiff competition from low cost materials produced abroad and increasing government regulations aimed at lowering green house gas emissions and improving environmental sustainability. As a result, Canada's wood products industry is currently undergoing the most severe economic downturn in its history. Among the industry's many products, pulp and paper make up a significant fraction. The pulp and paper industry contributes billions of dollars to the economy each year, and directly employs tens of thousands of people. Thousands of people are also employed by secondary businesses that support the pulp and paper industry. This sector includes companies that manufacture equipment and systems that are exported to all parts of the world. Given the importance of the pulp and paper industry to Canada's economy and its current economic troubles, it is imperative that the various processes in this industry are operated at levels that are highly efficient, autonomous, and environmentally sustainable. With this goal in mind, this research project addresses important challenges in model identification and fault detection and diagnosis. Three established UBC researchers, in process control, computer science, and mathematics, will lead a team including 2 Ph.D. students to refine current models of the kraft process and develop data analytics solutions to transform historical process data into knowledge and decisions. The academic and industrial partners will work together closely to keep the project focussed and relevant.