行星齿轮箱耦合调制振动信号分离与故障诊断

针对故障状态下行星齿轮箱振动信号通常表现为稳态调制和冲击调制成分耦合问题,建立不同求解模型,提出耦合调制振动信号分离方法。应用稀疏全变分模型实现冲击调制成分分离,采用改进的非对称惩罚函数,根据峭度系数与方均根系数指标构建正则化参数优选策略,采用受控极小化(majorization-minimization, MM)方法进行模型求解；采用基追踪实现稳态调制成分分离。试验结果分析表明:该方法在噪声D为0.15的信号中能快速分离出微弱冲击调制成分,算法迭代次数少、收敛效果好。 

Coupled modulated vibration signal component separation andfault diagnosis of planetary gearbox

Aiming at the status that planetary gearbox vibration signal is usually appeared as coupling problem of steady state modulation component and shock modulation component under fault condition, different solution models have been established and a new method for the separation of coupled modulated vibration signals has been proposed. Shock modulated component separation was realized by the application of sparse total variation model, using an improved asymmetric penalty function, according to a strategy of coefficient of kurtosis and selection of regularization parameters based on mean square index, using majorization-minimization (MM) to solve the model. Steady state component separation was achieved by using base tracking. Simulation and experimental results verify the effectiveness of the proposed method for the separation and separation of vibration signals of planetary gearbox in a strong noise environment. The results of experimental show that this method can quickly isolate the weak shock modulation components under the noise D of 0.15. The algorithm has fewer iteration times and better convergence effect.