A PHM architecture based on hybrid of model and data for electronic products

According to the current international PHM (Prognostics and Health Management) technology, this paper proposes a PHM architecture based on hybrid of data and model for electronic products. For prognostics, this paper presents a novel fault prediction method based on the hybrid of model and data—based on Bayesian Belief Networks(BBN) hybrid method. The proposed method considers both POF model of components and real-time monitoring data of systems, which makes the prediction result more accurate, more reliable, and more widely applicable. Then, a PHM platform according to the architecture and BBN techniques is implemented through the way of LabVIEW and MATLAB hybrid programming. By taking an elevator door circuit board as an object, the functions of PHM platform for electronic products are demonstrated and tested. The results indicate that the BBN hybrid fault prediction method proposed in this paper and the PHM platform can realize the function of fault prediction and remaining life prediction for electronic products, which lay an important foundation for the development of fault prediction technology for electronic products.