Application of parallel processing techniques to eddy-current NDT instrumentation

A model-based fault diagnostic procedure is proposed for an eddy-current test instrument which utilises the multiple model estimation algorithm (MMEA). This algorithm essentially consists of a set of Kalman filters and is an ideal candidate for realisation by parallel-processing techniques. A practical instrument has been developed, consisting of a commercially available eddy current test instrument linked to a parallel processor consisting of an array of four transputers. A set of tubular metal samples, with known inner and outer wall defects, has been fabricated to enable the performance of the experimental instrument to be evaluated, and some practical results are given.