基于加权ANSMD和参数优化RLD的风电机组液压型变桨轴承损伤检测

Damage detection for hydraulic pitch bearing ofwind turbine based on weighted ANSMD and parameter optimized RLD

围绕风电机组液压型变桨轴承损伤检测这一工程难题,分别提出了加权自适应非线性稀疏模态分解(ANSMD)方法和参数优化Lucy-Richardson反褶积(LRD)方法,并将二者相互融合,用于实现液压型变桨轴承局部损伤检测。首先,通过ANSMD方法将原始振动信号分解为一系列稀疏分量,计算各分量的香农熵值及权重系数,并通过稀疏分量加权叠加获得信噪比更好的重构信号。其次,通过自相关能量比优化策略对LRD方法的形态控制参数进行搜寻,并利用参数优化LRD方法进一步处理重构信号,实现连续性周期脉冲强化放大。最终,通过反褶积信号的包络解调分析提取变桨轴承的损伤特征频率成分。液压型变桨轴承测试信号验证结果表明,所提方法可有效检测变桨轴承局部损伤,为解决风电机组运维问题提供新的思路。

For the difficult problem of damage detection of hydraulic pitch bearing of wind turbine,the weighted adaptive nonlinear sparse mode decomposition method and the parameter optimized Lucy-Richardson deconvolution method are respectively proposed,and they are combined to achieve damage detection of hydraulic pitch bearing.Firstly,the original vibration signal is decomposed into several sparse components by adaptive nonlinear sparse mode decomposition algorithm,the Shannon entropy and the weight coefficient of every component is calculated,and then the reconstructed signal with better signal noise ratio can be obtained by sparse component weighted superposition.Furthermore,the morphological control parameter of Lucy-Richardson deconvolution algorithm is searched by autocorrelation energy ratio optimization strategy,the reconstructed signal is processed using parameter optimized Lucy-Richardson deconvolution algorithm,and then the continuous periodic pulses are enhanced and amplified.Finally,the damage feature frequency component of pitch bearing can be extracted by envelope demodulation analysis of deconvolution signal.The testing signal verification result of hydraulic pitch bearing shows that the proposed method can effectively detect the local damage of pitch bearing,and provide a new thought for solving the operation and maintenance problem of wind turbine.