Data acquisition and data mining in the manufacturing process of computer numerical control machine tools

Abstract

Networks of five-axis machine tools produce huge amounts of process data. These data directly reflect the running condition of the machine tool but are seldom used to examine the machine performance. This study proposes a new data acquisition method based on the Object linking and embedding for Process Control protocol without any additional monitoring equipment. The data collection principle is explained, and a client is developed based on the SIEMENS 840D system. Considering less influence on the manufacturing process, a communication architecture for the machine network is designed with a special computer transmitting the data to the server. A compression algorithm is applied to reduce the storage capacity of massive amounts of data. Finally, a method for predicting the future performance of the machine tool is proposed using similarity analysis of the time series. A Petri net model is also established to diagnose possible failure causes. These methods significantly improve the machine tool reliability and find potentially important information from the data in the manufacturing process.