The Great Plains/Rocky Mountain Hazardous Substance Research Center
Click here to skip the navigation menuHome Contact Search Publications Conferences Programs Research Phytoremediation Links

Intelligent Process Design and Control for the Minimization of Waste Production and Treatment of Hazardous Waste

Principal Investigators
L.T. Fan, Kansas State University

Abstract

Goal: The goal of this research is to develop intelligent or computer-aided systems tools for synthesizing and designing environment-friendly processes and controlling such processes with minimum waste generation.

Rationale: Waste minimization can be realized through source reduction and recycling. Efficient process synthesis and design, robust control, reliable diagnosis, and flexible production scheduling are important techniques to effectuate source control of waste.

Approach: The approach being employed is comprehensive and unique. Comprehensiveness of the proposed approach arises from the fact that all three levels of synthesis and design of processes and control systems for these processes, namely, macroscopic, mesoscopic, and microscopic levels, are incorporated into it. Uniqueness arises from the fact that the approach resorts to the most modern graph-theoretic methods, which are mathematically and theoretically rigorous, and to the techniques of artificial intelligence and neural networks, which are logically sound and computationally efficient.

Status: At the macroscopic level, the investigator has approached process synthesis by first identifying operating units represented by special directed bipartite graphs, specifically process or P-graphs, which uniquely specify each process structure in the synthesis. Materials are denoted by a classification scheme. For a given set of operating units, a large number of combinatorial structures are possible, but only an inordinately small fraction is actually feasible, which is determined on the basis of axioms stating self-evident facts. This results in a set of mixed integer nonlinear programming problems of relatively small size that can be solved effectively and speedily by means of the novel accelerated branch-and-bound algorithm. At the mesoscopic level, the structure of a process has been refined to improve its efficiency. This has been accomplished through heuristics rules much less computationally complex than traditional algorithmic techniques. It is envisioned that integration of these heuristic rules and the algorithmic method based on the P-graph can eventually give rise to a highly efficient technique at this level. At the microscopic level, detailed interconnections among process units have been identified, and the control system has been incorporated into the process structure. Improved systematic techniques based on AI methodology allow highly controllable mass and heat exchanger networks to be synthesized. The research to date has demonstrated that this comprehensive approach to process design or synthesis is highly effective and capable of synthesizing efficient systems for production, in-plant waste treatment, and integrated production and treatment. The third year of this project has been completed but some work is continuing.

Clients/Users: Results are of interest to design engineers who wish to incorporate waste minimization into process synthesis and control system design.

Key words: waste minimization, process synthesis, robust control, pollution prevention, graph theory

Return to Research

 


Click here to skip the navigation menu| Home | Contact | Search | Publications | Conferences | Programs | Research | Phytoremediation | Links |