Research And System Design Of Electro-mechanical Equipment Health Condition Monitoring For Big Data

Electro-mechanical equipment is inevitable to become the integration of mechanical, electrical, hydraulic and mechanical technology. The integration of different knowledge systems, adding to complex conditions in industrial production leads to the result that the possibility of electro-mechanical equipment failure has greatly increased.Traditional methods for the research on health monitoring and diagnosis of electro-mechanical equipment mostly base on the collected state monitoring data during their good running. With the increasingly higher requirement for the electro-mechanical equipment health monitoring in industrial manufacturing domain, and the need of various parameters of the real-time monitoring of electro-mechanical equipment, it will be inevitable to result in the increase of the amount of conditions monitoring data, which exceeds the ability of the traditional data storage and data processing. Therefore, the research on the method of electro-mechanical equipment health monitoring for big data is of great significance. In this paper, some researches on the method of electro-mechanical equipment health monitoring for big data were conducted, the main contents are as follows:(1) Analyzing the monitoring data collected from electro-mechanical equipment, grouping the monitoring data and extracting the feature parameters. Using adaptive network-based fuzzy inference system(ANFIS) to establish the mapping relationship between the conditions monitoring data and electro-mechanical equipment health condition.(2) Setting up a Hadoop platform in the pseudo-distributed mode to storage the state monitoring data of electro-mechanical equipment with Hadoop distributed file system(HDFS) and to analysis the state monitoring data using Hadoop parallel calculation model of MapReduce.(3) Designing a health condition monitoring system for electro-mechanical equipment based on adaptive network-based fuzzy inference system(ANFIS) and Hadoop with C/S(Client/Server) model. The related function modules are realized.