Research On Fault Diagnosis Of Regenerative System Based On RBF Neural Networks

The operation conditions of the regenerative system have an important impact on the safety and economy of the units, research regenerative system fault diagnosis strategy for improving the unit operation level has important theoretical and application value. In this paper, regenerative system's performance is discussed, and common faults are analyzed. In order to diagnose the fault in different load, this paper calculates the parameter target values. The fault sample character knowledge base has been established combined with the expert experiences and correlative literature. A method was brought forward that double-deck fault mode recognition method of system-level and device-level by adopted fault diagnose of regenerative system and the knowledge bases have been training commendably by RBF network. It was exploited what is the fault diagnosis system through the method of calling Matlab by means of Delphi. And this system has the characteristics such as quick diagnosis, friendly interface and easy operation etc. With matlab, making use of RBF network learning rules, the knowledge construction is completed; With Delphi calling Matlab; fault diagnosis application software for the regenerative cycle system was exploited.