基于自回归求和滑动平均模型的风力发电机轴承寿命预测

摘要

提出了基于自回归求和滑动平均(ARIMA)模型的风力发电机轴承寿命预测方法.以经小波包去噪、平稳化处理后轴承振动信号的均方根值为特征量,利用贝叶斯信息准则确定ARIMA模型的自相关阶数和滑动平均阶数,并用矩估计方法求出ARIMA模型的参数.利用求得的ARIMA模型得出振动信号均方根值的变化趋势,进而预测出风力发电机轴承的寿命.仿真结果表明,基于ARIMA模型的风力发电机轴承寿命预测方法能够有效地利用实时数据对轴承寿命进行预测,为风力发电机组的设计和维护提供理论依据.%Aiming at the prediction of bearing life for wind generator,a bearing life prediction method based on autoregressive integrated moving average (ARIMA) model for wind generator was proposed.The root mean square (RMS) value of bearing vibration signal was extracted by wavelet packet denoising and smoothing as characteristic quantity.The autocorrelation order and the moving average order of the ARIMA model were determined with the Bayesian information criterion,and the parameters of the ARIMA model were calculated by moment estimation.With the ARIMA model,the RMS value trend of the vibration signal was obtained,and then the bearing life of wind generator was predicted.The results show that the bearing life prediction method of wind generator based on the ARIMA model can predict bearing life by using real-time monitoring data,which provides a theoretical basis for the design and maintenance of wind turbines.