Research And Application Of Fault Diagnosis And Remote Supervisory In Simulated Moving Bed Control System

Simulation Moving Bed (SMB) chromatogram separation technology is widely developed inthe 60's last century. A distributed control system was designed according to the technical flow ofit and has applied in practice. Based on the shortage of system in use, the following aspects arediscussed.The existing supervisor system has the ability to detect the overrun faults of some parameters,but it lack the function of diagnosing recessive fault of sensors, such as gain-changed fault. Toenhance the PH sensor reliability, the methods based on signal processing and neural networks arediscussed. Firstly, Radial Basic Function (RBF) neural network is used to built predictor modelwith a serial of value of sensor, and the residual between output of predictor model and actualvalue is computed to determine whether sensor is working normal or not. Then wavelet transformis used to extract the sensor feature from fault signal, and finally RBF diagnostic neural networkis used to map from fault feature to fault types. The application to the PH sensor test proveseffective.In order to change the shortcomings of department supervisor,according to the softwarestructure of SMB control system, the paper employs Web technology, database technology, ASPtechnology and ActiveX technology to realize data remote dynamic display, data real trend display,data overrun alarm. To act as support decision system of supervisor, the paper presentsimplementation of fault diagnosis based on Matlab web server and ActiveX module programmingmethod. From the result of usage, the remote supervisor system is effective.