压缩感知框架下的共振解调故障诊断方法

Fault Diagnosis Method of Resonance Demodulation under Compressive Sensing Frameworks

风力机滚动轴承早期故障诊断中,压缩感知算法能够利用信号的稀疏性对信号去噪,但稀疏度的选取对去噪结果影响较大。由于信号故障成分在傅里叶域的稀疏度已知,故可通过傅里叶变换基和压缩感知子空间追踪(CS_SP)算法对风力机信号的包络特征进行不完全重构,以降低噪声和其他无关信息的影响,获取直接反映故障特征的信号成分,从而提取故障特征频率。研究结果表明,压缩感知框架下的的共振解调技术能有效获取风力机滚动轴承的故障特征信息,验证了所提方法的有效性。

In the early fault diagnosis of wind turbine rolling bearings,the compressive sensing algorithm might denoise the noisy signals utilizing the signal sparsity,but the selection of sparseness had a great influence on the denoising effectiveness.Since the sparseness of signal fault components was known in the Fourier domain,the envelope feature of wind turbine signals might be reconstructed by Fourier transform basis and compressive sensing subspace pursuit(CS_SP)algorithm.As a result,the effects of noises and other irrelevant informations would be reduced,and the fault characteristic frequency might be extracted from the signal components directly reflecting fault characteristics.The results demonstrate that the resonance demodulation technique under compressive sensing frameworks may effectively acquire the fault characteristic informations of wind turbine rolling bearings.The effectiveness of the proposed method was verified.