Process control, real-time expert

A Real-Time Expert System for Process Control... 

The requirements for expert systems for process control have inspired new designs based on real-time knowledge base inferencing. Object oriented representation of plant equipment, knowledge representation of the interactions of processes and models of process behavior - heuristic as well as analytical --are incorporated into a real-time expert system for process control. The application of inference in real-time requires using metaknowledge to focus the inferencing resources of the expert system. Finally truth maintenance requires a temporal model of the time dependence of the truth of data and inferred results. 

Moore, R. Lindenfelzer, P. Hawkinson, L. Matthews, B. "Process Control with the G2 Real-Time Expert System," Proc. the First International Conference on Industrial Engineering Applications of Artificial Intelligence Expert Systems. Univ. of Tennessee Space Institute, June 1988. 

Expert systems are computer programs that simulate the decision-making process of human experts. The hallmark of expert systems is that decisions are based on heuristics (rules of thumb) when data is incomplete or there is not enough time to consider all possibilities [19]. An expert system can be a set of IF-THEN rules in FORTRAN, or it can be a written in one of the languages designed for expert systems, such as LISP. Expert systems can be used off-line to aid in cure cycle selection, or they can be used as real-time advisers or controllers. 

Expert Systems for Control of Plant Operations in Real Time Using Expert Systems to Monitor Safety Margins for High Hazard Processes... 

Moore, R.L. Hawkinson, L.B. Levin, M. Hofmann, A.G. Matthews, L.B. David, M.H. "Expert System Methodology for Real-Time Process Control," Proc. 10th World Congress on Automatic Control. IFAC. Munich, July 1987, Vol. 6, pp. 274-281. 

Each of the modular expert systems will be limited to a relatively narrow domain of expertise. However, together, these modules will encompass the many varied aspects of decision-making that would be normally carried out by the human expert from method selection, to real-time process control and then to reporting and documentation. 

This idea was embodied in the IUPAC definition of automatic systems "The use of combinations of mechanical and instrumental devices to replace, refine, extend, or supplement human effort and faculties in the performance of a given process, in which at least one major operation is controlled, without human intervention, by a feedback system"[367]. The key point for an automatic system is therefore the incorporation of a feedback system, i.e. the "combination of a sensing and a commanding device that can modify the performance of a given act" [367[. In other words, the response of a sensing device is used for the real-time modification of system operation. These definitions also hold for flow analysis automatic analysers are expert flow systems (Fig. 8.26), often referred to as smart or intelligent flow systems. 

Process-control systems wiU become more integrated with condition-based monitoring systems that in turn link to CMMS/EAM systems or even back to the condition-based equipment suppliers for real-time troubleshooting. Condition-based monitoring systems wUl eventuaUy link data coUection directly back to suppliers such as CSI, ENTER, SKF, and others via the Internet or RF. Reliability data analysis by off-site reliability experts providing contract services wUl be as close as e-mail. 

Three approaches dominate the real-time intelligent control field (1) expert systems, (2) neural net controllers (neurocontrollers), and (3) fuzzy logic controllers [50]. These intelligent control systems are based on two types of information processing symbolic and subsymbolic processing [15,51]. 

SEM s simulate human performance when controlling dynamic systems (Rouse 1980). They match overall human performance very well, or work even better by employing mathematical formulations of systems dynamics. For application purposes computational formulations are often not feasible. SEM s are useful for special purposes such as monitoring, adjusting and predicting outputs in real time. An evaluation of SEM s and expert systems for decision support systems in process control is provided by Zimolong et al. (1987). 

---