A knowledge based system to assist in the selection of appropriate geotechnical field tests

Abstract

The variety in geological conditions and range of geotechnical problems has led to the development of a considerable number of different in-situ test methods. The correct selection of the appropriate in-situ tests allows a safer and cost-efficient design to be achieved. A prototype Knowledge-Based System has been developed to assist in the selection of appropriate geotechnical in-situ tests. The system is model-based and has been implemented using PDC Prolog on a Personal Computer to perform two functions: i) general querying of the knowledge bases which it incorporates and ii) advise on selecting in-situ tests. The system consists of two knowledge bases (the Ground Knowledge Base and the Tests Knowledgebase), an Extended Inference Mechanism consisting of search rules developed to allow inheritance and transitivity inferences (as well as information retrieval facilities), an advisory rule developed for offering assistance in the selection of appropriate field tests, and a menu driven user interface to achieve ease of use. The Extended Inference Mechanism, and the user interface implemented for it, form a basic expert system shell. The knowledge required to be included in the system was obtained in two ways: i) from technical literature and ii) from a small knowledge elicitation exercise in the form of a questionnaire. The representation scheme adopted is the same for both knowledge bases and allows modifications (additions or deletions) of the existing knowledge to be easily made. Towards the completion of this research program, a comparative exercise was performed by re-implementing part of the system using the PROKAPPA software on a Sun Sparkstation 2 (both of which became available at that time). Throughout this exercise, the differences between the two implementation schemes were evaluated and the advantages and disadvantages of each of the schemes were identified.