SBIR-STTR Award

A feasibility study is proposed to examine the application of artificial intelligence programming to existing problems in the control of processes in the pulp and paper industry. In particular, the study will assess the use of the "expert systems" program technique employing empirically based production rules to simulate an expert human operator of complex processes which cannot be modeled mathematically. Process control candidate problems exist in the Tomlinson black liquor recovery boiler, the chemical pulp digester, and the paper machine. A candidate problem or problems will be selected as most suited to an expert systems approach toward automatic process control, and a preliminary engineering/economic analysis will be performed.The potential applications as described by the company: It is expected that a Phase II development effort, targeted by the feasibility study, will result in software leading to a commercial automatic process control package. Incremental improvement in the control of these processes would create large energy and production cost savings.

In this project, a software research and development effort will create a knowledge-based expert system to assist paper machine operators in deciding when to replace felts in the wet press section. Knowledge engineering techniques will be used to represent human categorical reasoning in this diagnostic problem domain. The performance of the prototype consultation program will be compared to that of human experts as the program is refined iteratively. The performance of the end-result program will be documented in terms of economic benefit and cost.Anticipated Results/Potential Commercial Applications as described by the awardee: If the Phase II performance goals are reached, it is expected that commercial expert consulting systems will be packaged, sold, and serviced in all paper mill product segments, creating large energy and production cost savings. Successful commercialization would lead to many other applications.