Making sense of non-destructive evaluation data with an artificial intelliaence approach

Abstract

This paper presents the feasibility of an artificial intelligence technique for processing and interpretation of non-destructive evaluation (NDE) data from assessments of engineering structures. The technique used is a learning and reasoning approach with belief network. With this technique, causal factors and consequent indicators in the data structure in relation with structure/material condition can be modelled, and their causal relationship can be established using the NDE data as the learning resource. Fundamentals of the technique are briefly presented, and then the potential applications of the technique to NDE data are demonstrated in two case studies.