Manufacturing Data Privacy Protection System for Secure Predictive Maintenance

In Industry 4.0 technology of industrial equipment is becoming more and more complex as physical-digital systems are combined to form a new industrial phase. The complexity of the equipment affects the frequency of equipment failures, and to avoid failures we need to predict the health of the equipment. We found that PdM is the most effective health prediction, but it has the challenges of data security and data availability. To address this aspect, we design and develop a manufacturing data privacy protection system (MDPPS). Data are collected by smart sensors and then encrypted. After that, machine learning models are used to extract features and predict the remaining useful life of the industrial equipment. Finally, the predicted results will help workers to carry out maintenance activities. The performances of the MDPPS achieve an average of 87.40% Accuracy, 95.57% Precision, 77.74% Recall, and 83.86% F1-Score.

I. Introduction

Industry 4.0 has been regarded as a new industrial stage where physical and digital systems are converging to provide novel solutions for the manufacturing process. It is related to a systemic adoption of Internet-of-Things, big data analytics, and artificial intelligence [1]. However, industrial equipment is becoming technologically more advanced, more complex, and more difficult to control [2]. According to [3], equipment idling, minor stoppages, and sudden shutdowns account for the largest number of production problems. Because of these breakdowns, manufacturing systems are made vulnerable to economic losses. Thus, manufacturing enterprises are realizing that there is a critical need for proper maintenance strategy of industrial equipment [4].