An expert system for field inspection of concrete dams: Part 1, engineering knowledge

Abstract

This paper covers in more detail the material that was introduced in Ref. [1] and describes the final engineering reasoning and knowledge used for an expert system to assist during the field inspection of existing concrete dams. The knowledge base of this expert system was created by combining the traditional expert system approach with the reasoning procedures of qualitative physics. The traditional expert system approach was used to identify the shallow knowledge pertaining to the problem, and qualitative physics was used to identify the deep knowledge. The deep knowledge structured and organized the identification and representation of the shallow knowledge. The expert system uses shallow knowledge to combine a qualitative expert reasoning with a quantitative procedural evaluation of the stability of a dam. The present paper concentrates on the description of the engineering knowledge and Ref. [2] will concentrate on the artificial intelligence issues associated with the knowledge acquisition and representation associated with this expert system. The validation process compared results obtained by the expert system with those obtained by independent field inspection experts. Some of the results are presented here.