Fault Diagnosis Of Large Wind Turbine Bearings Based On Information Entropy And PSO-ELM

In recent years,due to the huge power consumption in the world and the increase of environmental pollution,the development of wind power generation in the world is very rapid,but the maintenance cost of wind power equipment is still high,which seriously affects the safe operation of wind power and reduces wind power.Economic benefits.Aiming at the problem of fault diagnosis of key components of wind turbines,an algorithm based on information entropy and particle swarm optimization(LPO-ELM)algorithm was proposed to establish an optimization model,and the improved particle swarm optimization algorithm was used to solve the problem.Specifically completed the following work:(1)Through the empirical mode decomposition method and related improved model,the vibration signal of the mechanical bearing fault is analyzed,and the information eigen component is extracted by the information entropy,so that the fan bearing fault can be accurately diagnosed.The overall average empirical mode decomposition algorithm(EEMD)can be used to solve the problem of mode aliasing caused by intermittent signals,that is,to decompose different frequencies uniformly distributed in regular white noise.The adaptive noise complete set mode decomposition algorithm(CEEMDAN)method adds a limited number of adaptive white noises during the operation of the algorithm,which not only reduces the average number of operations,but also reduces the signal reconstruction error.(2)Using the improved CEEMDAN algorithm to decompose the original signal,and use the kurtosis index to select the IMF component for signal reconstruction and reconstruction signal selection,and finally realize the accurate extraction of the rolling bearing fault characteristics.The ELM and PSO-ELM algorithms are used to calculate and compare the data respectively.It is found that the accuracy of the PSO-ELM algorithm is higher than that of the extreme learning machine(ELM).The PSO-ELM algorithm is used for optimization calculation,and the results can reflect the real data better.Variety.(3)The online diagnostic test platform of wind turbine was built,the hardware design and software development of the system were completed,the experimental data was recorded,and the acceleration trend and waveform spectrum and envelope diagram of the vibration signal of the unit were analyzed.It is verified that the optimized control scheme proposed in this paper can better reflect the trend of fan status and effectively reduce the operation and maintenance cost of the fan.