多传感器数据融合的风力机侧风状态评估OA

In order to explore the strain characteristics of wind turbine blades during continuous crosswind,a multi-sensor data fusion method for wind turbine crosswind state evaluation is proposed in this study.In this method,ensemble Empirical Mode Decomposition(EMD)and composite multi-scale permutation entropy-wavelet algorithm are used to jointly reduce the noise of wind turbine blade strain signals,and then Kernel Principal Component Analysis(KPCA)is used to integrate the multiple groups of strain signals after noise reduction.The Squared prediction error(SPE)statistic was used as the evaluation index to effectively divide the wind turbine crosswind process.The results show that the proposed method is effective in reducing the noise of non-stationary wind turbine blade strain signal,and can accurately reflect the strain variation law in continuous crosswind process.In addition,through the combination of KPCA and SPE statistics,the process of wind turbine crosswind is segmtioned,and the operating state of wind turbine crosswind under different influencing factors is analyzed.