Research On Gas Turbine Health Status Evaluation And Fault Diagnosis

With the increasing emphasis on the living environment, more and more gas steam combined cycle power plants has been built in recent years. At present, gas turbine equipment abnormality is just monitoring by alarm value and jump value of the preset parameters. There are the following shortcomings:1) Setting the threshold simply cannot have a good consideration of the impact of the parameters on the working condition,2) in general, the alarm or jump machine only happened after serious failure, and it cannot achieve the early warning of failure; 3) when the fault occurs, it’s difficult to have a good fault diagnosis.To begin with, the mechanism and criterion of typical gas turbine failure are analyzed on the basis of full investigation of the structure characteristics, working principle and common failure mode of the three heavy duty gas turbine units from GE, MIT and SIEMENS. According to the basic principle of gas turbine fault diagnosis, the relationship between thermal performance parameters, mechanical performance parameters and the operating state of the unit is analyzed.Then, taking the operating condition parameters of the gas turbine as input, the residual of original characteristic parameters as the output, the ANFIS method is used to predict the abnormal deviation of the original characteristic parameters in different operating conditions. With gas turbine running data as input, after feature extraction by OLPP method, the SOM is used to evaluate the overall health status of gas turbine, and this method can be used to send out an alarm signal before machine failure for the purpose of failure warning. And SGT5-2000E (V94.2) single shaft gas turbine is used as the reference for verifying the case.Finally, it quantify the fault symptom, and then the diagnosis rule of expert knowledge base is compared, and the fault pattern is recognized according to the principle of minimum stick schedule. Based on the fault analysis, the fault cause, influence and treatment measure are found out, and the diagnosis report is generated as an important reference for the maintenance decision of gas turbine.