Operational State Evaluation and Maintenance Decision-making Method for Multi-state CNC Machine Tools based on Partially Observable Markov Decision Process

Abstract

With the development of science and technology, the manufacturing technology of computer numerical control (CNC) has been one of the irreplaceable important technologies in the manufacturing industry. However, due to the lack of sufficient data, the existing evaluation methods cannot accurately evaluate the operational state of CNC machine tools, and an effective maintenance scheduling strategy cannot be proposed. Aiming at the lack of monitoring data and the unclear understanding of the degradation mechanism of CNC machine tools, an operational state evaluation and maintenance decision-making method of multi-state CNC machine tools based on partially observable Markov decision process (POMDP) was proposed. On the one hand, considering the imperfect operational information of CNC machine tools and the influence of different maintenance activities on the degradation process of NC machine tools, a dynamic evaluation model for the operational state was constructed. On the other hand, a maintenance decision model for CNC machine tools was established with the goal of minimizing the total cost of losses (including maintenance costs and potential failures costs). The proposed method is applied to CNC machine tools of cylinder to verify the effectiveness and advancement of the method.