Development of a diagnostic expert system for a whipped toppings process

Abstract

Abstract A diagnostic expert system for the whipped toppings process was developed in an effort to: improve product quality/consistency; improve process efficiency by reducing downtime; improve product yield by minimizing production of marginal product; and standardize diagnostic performance by further educating the plant operators. The expert system is PC-based and has been implemented using VP-Expert, a rule-based development tool, and VP-Info, a database management system. The main principles of operation are as follows. First, the expert system runs off-line and is triggered when the plant operator recognizes, during normal inspection, that a problem exists in the process or with the product. Then, the diagnostic expert system, acting as a consultant to the plant operator, seeks to locate the physical or process malfunctions that are the root causes of the observed process abnormalities. During the course of diagnostic consultations, the expert system serves to focus the attention of the plant operators and aids in coordinating the diagnostic activity of the plant operators and plant maintenance personnel. The expert system proposes corrective actions upon discovery of a malfunction in the plant. The system architecture is of a modular construction so that it may act as a template for future diagnostic expert systems. This paper also includes a discussion of lessons learned from experience in building the expert system.