Health estimation method of manufacturing systems based on multidimensional state prediction

Abstract

Systematic and accurate health estimation for the running manufacturing system is the prerequisite to implement production scheduling and predictive maintenance. This enables remedial actions to be taken in advance and reschedule of production if necessary. However, existing studies pay more attention to the failure diagnosis of equipment, while ignoring the output and input characteristics of the manufacturing system. Therefore, this paper presents a novel method for health estimation of manufacturing systems from three dimensions of equipment performance, product quality and task execution Firstly, the equipment performance state is represented based on the theory of polymorphism. Secondly, the quality state is defined to describe the qualified degree of the output products according to the response model. Thirdly, a task execution state modeling method is proposed, and the correlation between sub-task execution states is considered based on Copula function. Then, an integrated model is built to prognosis the change trend of manufacturing system health by integrating the above three states. Finally, a case study conducted to illustrate the effectiveness of the proposed method.