Generic Design Standards Processing in an Expert System Environment

Abstract

Checking design conformance with governing standards, codes, specifications, regulations, and manuals is a mandatory step in designing engineering systems. Automating this design checking process, therefore, is an important goal that has far‐reaching design productivity implications. Current conformance checking automation generally involves the use of application programs into which the standards, or an interpretation of their intent, have been directly coded. This approach is extremely inflexible and often error‐prone. This paper investigates the feasibility of casting design standards in an alternative form, one suitable for flexible processing in a knowledge‐based expert system environment. The proposed form not only enables a designer to readily check the conformance of designs, but it also allows him to evaluate the effects of changing design variables as well as evaluate the impact of down‐stream activities and events, such as change orders, on the initial design. Two prototype standards processing systems utilizing the production system approach has been constructed and are described. Although the obvious direct translation for a design requirement in a standard to a rule in a rule‐based expert system has its advantages, a more generic and flexible computer representation for design standards is desirable. The approach advocated in this paper is to represent standards as databases of facts which are used on an as‐needed basis by a rule‐based design conformance checking program to monitor design evolution. The representation is derived from a unified view of standards obtained by using the standards modeling tools proposed by previous researchers in this field during the past decade. A generic knowledge‐based standards processing architecture is proposed and described. An implementation of this architecture is also discussed.