向健平， 姜楠楠， Simon Jonathan Watson

(1． 长沙理工大学 能源与动力工程学院， 长沙 410114； 2． Faculty of Aerospace Engineering,Delft University of Technology, Kluyverweg 1, 2629 HS, Delft, Netherlands)

一种风力发电机自动故障诊断及预测方法

向健平1， 姜楠楠1， Simon Jonathan Watson2

(1． 长沙理工大学 能源与动力工程学院， 长沙 410114； 2． Faculty of Aerospace Engineering,Delft University of Technology, Kluyverweg 1, 2629 HS, Delft, Netherlands)

为了获得故障信号精确的时域和频域信息，提出了一种Morlet小波变换补偿方法，首先计算出故障信号的时间和频率信息，然后根据Morlet小波系数中心频率峰值对计算的故障信号的幅值进行补偿，得出故障信号的准确幅值.采用Simulink模型证明该方法的可行性，并将该方法应用到测试风力发电机和实际大型风力发电机的电功率信号分析中.结果表明：该方法可以自动获得故障信号按时序排列的振幅趋势图，显示了部件发生故障后的剩余使用寿命期限；风力发电机特别是海上风力发电机的维护维修计划可根据此时间信息进行制定，降低风电运维成本.

风力发电机； 故障诊断； Morlet小波变换； 补偿算法； 电功率信号

Abstract： A Morlet wavelet-based compensated algorithm was proposed to calculate the accurate amplitudes of faulty signals. The specific way is to compute the time range and frequency values of the faulty signals at first, and then to compensate the amplitudes calculated for above faulty signals according to the center frequency values of Morlet wavelet coefficients to further obtain the accurate amplitudes. A Simulink model was used to demonstrate the feasibility and generalization of the algorithm. At the same time, the algorithm was used to analyze the electric power signals of a test rig and large turbines. Results show that this algorithm can automatically find the amplitude trend of faulty components in a time sequence, and indicate the residual service life of wind turbines after faults are generated. Based on the information of the residual service life, the maintenance and repairing plan for wind turbines, especially offshore ones, can be developed to lower the cost of wind power in operation and maintenance.

Key words： wind turbine; fault diagnosis; Morlet wavelet transform; compensated calculation; electric power signal

自然环境中存在很多复杂信号，其中一些如心电图信号和地震信号都是非平稳的，但信号测量方式可以控制[1-2].通常运用小波变换来处理这些复杂的非平稳信号[3].由于这些非平稳信号的频率可通过控制测量方式来确保其信号频率不变，因此，与信号相关物理对象的状态可以通过小波图像比较来监测.然而，还有一些复杂信号，如不可控变工况下的变速异步风力发电机电功率信号中包含的一些元件故障信号，这些信号具有不连续性，其频率随时间发生变化，用……