Research On Condition Monitoring And Fault Diagnosis Method Of Medium Speed Coal Pulverizer

With the progress of China’s electric power technology,electric power system is developing towards the direction of high parameters and high-capacity, which puts forward higher requirements to device of stability and safety. By the end of 2012, coal-fired generating units account 71.5% of the country’s total installed generating capacity, and the coal-fired units will continue to be dominant in country’s energy structure quite a long period of time in the future. Coal pulverizer is a significant auxiliary equipment of boiler.Any faults happened to it may affect the stability operation of the unit, even the safety of the grid. Monitoring the condition of the power equipment and making quick fault diagnosis become more and more important.In this paper, the coal-pulverizer of a power plant in western Henan is chosen for studying.All the faults occurred from 2007 till now has been counted, all the faults has been broken down into three categories according to a large number of historical data analysis and fault mechanism analysis:1)Fault of single-point or multi-point parameters can be measured,such as loading oil pressure low, oil temperature of grinding roller high; 2) Coupling fault of multi-point parameters can be measured, such as coal breaking, spontaneous combustion of coal, full of coal; 3)Fault of no parameter settings, such as coal pulverizer vibration.For the first one, proposed fault diagnosis method based on threshold method combined with the example of abnormal loading pressure fault diagnosis,example showed that the threshold method was a simple and effective method for fault of single-point or multi-point parameters can be measured;For the last one, proposed a method that the current residual sequence of coal pulverizer to sign the coal pulverizer vibration, and the current residual sequence of coal pulverizer referred to the difference between the predict current and the measured current, then used wavelet packet analysis to extract sub-band energy ratio as characteristic quantity and LS-SVM method to diagnostic faults,example showed that this method can construct the coal pulverizer vibration and distinguish the two reason of it accurately.