The Development Of J2EE Based Integration Tool For Intelligent Prognostics

Reliability, efficiency and time to failure of machinery facilities are very crucial for a corporation to guarantee its operation safety, and improve its economic benefit and market competitiveness. In order to realize life-cycle intelligent management and near-zero-downtime of important facilities with key components, it is of great significant to evaluate their real-time performance and predict their remaining life dynamically, which can be achieved employing advanced mode of intelligent management, utilizing technologies of prognostics and network communication, and taking full advantage of information technology to facilitate the maintenance management of machinery facilities. On the other hand, as an emerging maintenance technology, WEB based life-cycle intelligent prognostics, which is characterized by real time, intelligence, systematization, prediction and networking, can realize real-time performance evaluation and diagnosis, intelligent system, and early stage prognosis, leading the trend of development in the field of facility maintenance. Therefore, aiming at the networking of intelligent prognosis, efforts are made to carry out research on the J2EE based intelligent prognosis tool in this thesis.A uniform four layers framework is proposed in the perspective of intelligent prognostics implementation. The function and design concept of each layer is presented clearly. Additionally, it is accentuated that standardization of data format is crucial for wide application of intelligent prognostics. Therefore, the intelligent prognostics framework based on J2EE was proposed.Research is carried out on the key techniques of J2EE based system for intelligent prognostics , including modules based on algorithms for signal processing and feature selection, design of condition monitoring modules based on Ajax, database and its interface, and the interactive technique for Matlab and Java, JfreeChart (Java drawing technology) and Joone (Java ANN technology).Finally, a J2EE based integration tool for intelligent prognostics was developed. The tool contains a condition monitoring module for machinery equipment based on Ajax, a searching module for machine history information based on SQL Server and JDBC, algorithms for feature extraction and feature selection developed utilizing Matlab Builder for Java, and ANN-based algorithm for intelligent prognosis developed utilizing Joone ANN framework. The algorithms for feature extraction include FFT algorithm and WPT algorithm, and the algorithms for feature selection include PCA algorithm and Fuzzy Clustering algorithm. To display the result of intelligent prognostics, full advantage of JfreeChart has been taken for more explicit understanding of the result. Database connection and graphic display is realized through secondary development of the configuration software in ICONICS so that the historical data can be displayed and the historical information of intelligent prognosis can be presented dynamically.