Research On Temperature Early Warning Of Hydro-generator Stator Winding

As the core component of the hydro-generator,the stator is one of the main heat sources of the generator due to its complex structure and the accumulation of heat due to the flow of current.Compared with other components,the stator is more prone to failure,and statistics show that the probability of winding failure is far greater.However,the current real-time monitoring of the stator temperature of hydropower stations generally still uses the alarm when the temperature exceeds the set upper limit,and the emergency shutdown is exceeded when the temperature exceeds the set upper limit.When the temperature exceeds the alarm is monitored by this method,the stator temperature Internal faults are generally extremely serious.In view of the above problems,this paper selects the stator winding as the research object,uses its temperature data to carry out fault warning research,reduces the situation that the generator continues to work at excessively high temperature,and protects the safety of the generator.This paper uses the equipment abnormality monitoring and early warning method in the data mining early warning method to build the generator stator winding temperature early warning system.Therefore,the early warning algorithm in this paper is mainly divided into the use of the stator winding normal temperature rise model to obtain model prediction data and compare model prediction data with real-time There are two parts of monitoring data for early warning.The main research contents are as follows:This article takes the No.2 unit of Yunqiao Power Station in Shizhu County as the research object,gives a detailed introduction to its structure,the installation of stator winding sensors and temperature limits,and then models the normal temperature rise of the stator windings of the generator set and establishes the temperature rise curve model of the stator winding in normal operation is adopted.The key parameters in the model are identified by the nonlinear least square method to obtain the thermal time constant of the unit and the identification results are analyzed.A new method is proposed to solve the different generators.The load temperature rises relationship model of the steady-state temperature rise value under load,and the model prediction data obtained by calculating the normal temperature rise model can be used for subsequent early warning.First,perform real-time online monitoring of temperature data.In order to reduce the influence of null or bad value data in the data chain on subsequent algorithms,the real-time temperature data is preprocessed.Then,in order to solve the situation of missing or wrong judgment of the temperature data trend in a single time window,a variable window Kendall-f algorithm is adopted,and the two-time windows of fast sliding and slow sliding are used to judge the temperature rise trend.The judgment of the temperature rise trend is more accurate.Finally,by using the discrete Fréchet distance algorithm to compare the similarity between the predicted temperature data and the real-time temperature data,the abnormal temperature rise warning can be carried out.Use SQL Server database and MATLAB software as the platform to build the entire generator stator winding temperature early warning system.The hydropower station uses GUI user interface to interact and display the entire system,and use web pages to monitor real-time temperature data,stator winding status and early warning in the cloud the diagnosis results are displayed.The early warning system designed in this paper can provide early warning for abnormal temperature rises that have not reached the upper limit alarm value,issue warnings to attract the attention of operating personnel,and protect the safe and stable operation of the unit.