论文部分内容阅读
由于齿轮箱振动信号复杂及故障类型难以预知,因此提出了一种引入变异算子PSO的小波神经网络对齿轮箱故障进行诊断。仿真结果表明,该方法明显优于传统小波神经网络方法,不仅迭代次数大幅减少,而且收敛精度和收敛速度也有很大提高。将引入变异算子PSO的小波神经网络方法应用到高转速运转下齿轮箱故障诊断中,试验结果进一步验证了该方法的精确性,并能准确地识别齿轮的损坏程度。
Due to the complexity of gearbox vibration signal and the type of fault, it is difficult to predict the gearbox fault. Therefore, a wavelet neural network with mutation operator PSO is proposed to diagnose gearbox fault. The simulation results show that this method is obviously superior to the traditional wavelet neural network method, not only the number of iterations is greatly reduced, but also the convergence precision and convergence speed are greatly improved. The wavelet neural network with mutation operator PSO is applied to the gearbox fault diagnosis under high speed operation. The test results further verify the accuracy of the method and can accurately identify the gear damage.