The Fault Diagnosis And Prediction Of The Turbine

In this paper,Firstly,the Research Status and Development Trend of the fault diagnosis were analyzed.Through collation and summing up the relationships between the advance vibration signs and faults of the Turbine,research methods and ideas of the problem were analyzed.Secondly,artificial neural nets theory was used in fault diagnosis of turbine electric generating set.Combined with the Sohre's advance vibration signs form and specific conditions of turbine electric generating set,the characteristic value form was designed which applied to vibration reason of turbine electric generating set.Then put the advance vibration signs into artificial neural network to do fault diagnosis.The result can preliminarily meet the needs of online monitoring of turbine electric generating set.In the another hand,for the vibration characteristics of rotating machinery,the paper discussed in depth the use of time-series models to predict the trend of the vibration method. A method was proposed to deal with field data of non-stationary,non-normality,model type and singular value,in order to construct a suitable model to the trend of implementation of the vibration of the multi-step prediction accuracy.