Knowledge Discovery In Big Data For Complex Electromechanical Equipment's Maintenance And Its Application

Equipment maintenance is an important means to improve the reliability of equipment. The condition monitoring of equipment is an important way to determine the equipment failure and analyze the reliability of equipment. A large amount of data have been generated in the process of condition monitoring and equipment maintenance. While monitoring and maintenance is more concerned with some certain trend and threshold point of these data. And most of these historical data are collected during the normal operation period of equipment. Therefore, based on the large history data of equipment, this paper used the method of knowledge discovery in big data to discover new knowledge, and then realized the transformation of data to knowledge. Finally, these knowledge would be applied to equipment maintenance. The main research work of this paper are as follows:(1) The common maintenance standards are usually based on threshold maintenance. But in the actual industrial production, the threshold value is difficult to determine due to the difference of environment and use conditions. Based on this, this paper proposes the method of knowledge discovery based on big data. Through the AR fitting, the data retains the characteristics of the time, and then put them into the SOM network. Last, the knowledge discovery model is formed through training. This paper also defines the boundary of the original knowledge by clustering method to ensure the validity of knowledge. Finally, the validity of the method is verified by the data from experimental platform. And compared with the threshold method, the maintenance method based on knowledge is more sensitive.(2) The maintenance requirement of complex electromechanical equipment was analyzed from the complexity of equipment structure and manufacturing process. And on the basis of the analysis, a quantitative study of the maintenance difficulty of the equipment was carried out based on the method of expert scoring. Through quantitative analysis, the key system that requires special maintenance of electromechanical equipment can be found. The main characteristic of maintenance is difficult to be found by empirical knowledge. Based on this, the maintenance mode of knowledge discovery was used to maintain the key equipment in this paper, and thus a maintenance strategy based on data knowledge was put forward.(3) Finally, the method was verified by the data of HC-30 CNC lathe. Through the definition of key components, three indexes were selected, namely vibration acceleration, spindle acceleration and oil temperature of spindle box, to maintain the spindle system based on data knowledge. And then, by the discovery of maintenance knowledge, the spindle system can be judged to be running normally and it didn't need to be maintained. So the conclusion drawn by this method was consistent with the threshold method. In addition, the validity of the method is verified by the side.In this paper, a method of equipment maintenance was proposed, which was suitable for complex electromechanical equipment. The research results show that this method has a better application effect, which provides a new method and basis for the equipment maintenance decision.