An Objective Weight Determination Method Based on Model Healthy status Recognition Rate

Health measurement model is used to objectively and accurately express the current healthy status of the product. In this paper, three common equipment health measurement models with different description forms are analysed, and the health measurement model based on Mahalanobis distance and the health measurement model based on overlapping degree are optimized on the basis of practical application. Then according to the application of these health measurement models, a set of rule reasoning process for model selection is established and automatic model selection is realized. In the process of analysing and optimizing the models, this paper also finds several subjective factors of different degrees in these models, which are parameter values determined by the experts’ experience. In order to eliminate such subjective factors, a method for determining subjective model factors based on healthy status recognition rate is proposed based on the objective operation data of different healthy status and the identical healthy status of equipment. Finally, taking the weight determination method in the health measurement model based on Euclidean distance as an example, a mathematical model and its optimization objective are explained in detail.