Investigation on the Bi-direction Fusion of Data and Knowledge in Prognostic and Health Management

Data and knowledge fusion are hot topics of machine learning, and the important working direction of Prognostic and Health Management technology (PHM). In order to crack up the schismatically situation of data and knowledge in industrial community, we propose the bi-direction fusion of data and knowledge in PHM including knowledge mining and knowledge embedding. We present an overall framework and conduct exploratory studies through actual project cases. In terms of knowledge mining, a KMAHP (knowledge mining analytic hierarchy process) test system decision method for structural health monitoring is proposed. And for knowledge embedding, a physical-based health baseline construction method is proposed for critical telemetry health status monitoring of aircraft. The results of the two cases are in line with expectations, which show that the proposed fusion framework and the concrete implementation plan are both effective exploration research and engineering practice for PHM.