Study On The Application Of Fault Predictionin The Thermal Power Plant Combustion System Based On Grey Error Neural Network

With the widespread application of the automatic control system and the continuous expansion of the industrial scale,factories have higher demands on reliability and stability of the system.It is extremely meaningful to predict and diagnose all kinds of abnormal states or faults in time and accurately in order to reduce the loss to the lowest level.Fault prediction technology is a new technology developed with the transformation of maintenance concept and maintenance mode,which effectively realizes the active prevention of steering failure from passive responses to failure,and fault prediction based on intelligent system from traditional post steering system,which realizes the safety and economy operation of equipment to a certain extent.This paper studies the basic structure,process flow and the main safety performance index of the boiler in the thermal power plant.The T-S fuzzy neural network structure model and the multi parameter grey error neural network model are simulated by using the wall temperature value of 4# high temperature super heater as the research sample.The results show that the accuracy and speed of the multi parameter gray error neural network model are much higher.In order to study the model algorithm further,the fault prediction system is designed,and the simple and friendly interface and convenient operation mode have positive significance for the boiler related fault prediction.The boiler control system of the power plant adopts the TisNet-OnXDC800 series of DCS,which is integrated with the alarm technology,but a large amount of alarm information will be sent when the fault occurs.In order to realize the fault identification and prediction,it is necessary to analyze the alarm information by using expertise to extract the important information and realize the fault prediction.At the same time,this paper realizes the application of the multi parameter grey error neural network model in the XCU graphic configuration software,which has certain practical significance.