Extraction of Strain Characteristic Signals from Wind Turbine Blades Based on EEMD-WT

Analyzing the strain signal of wind turbine blade is key to studying load blade, so as ensure safe and stable operation in natural environment. The under continuous crosswind state has typical non-stationary unsteady characteristics. contains a lot noise, which makes analysis error. Therefore, it very important denoise extract features measured signals before analysis. In this paper, joint algorithm ensemble empirical mode decomposition (EEMD) wavelet transform (WT) used for first time achieve sufficient noise reduction effectively feature signals. application process EEMD-WT optimized. This optimization can avoid repeated selection basis function number layers due different conditions. EEMD adaptively decomposes into intrinsic functions, judge frequency IMFs, remove high-frequency components, retain useful components. components are denoised twice by transform, residual terms after secondary denoising reconstructed obtain characteristic signal. was applied simulating results were compared with EEMD. showed that method better performance, characteristics signals, lays solid foundation accurate