According to the fatigue life prediction of wind turbine mainshafts and towers under complex alternating loads

it proposes a fatigue damage assessment method based on an improved three-point rainflow counting method.Firstly

by establishing a stress cycle recognition model with adaptive adjustment functionality

the method achieves precise identification of stress amplitudes and their cycle counts under non-stationary load conditions.Secondly

the prediction model is optimized using the SSA-BP neural network algorithm

enhancing the model's generalization capability.Finally

comparisons with traditional prediction models using MAPE and R2 values demonstrate that this method improves the accuracy of fatigue life predictions

providing valuable references for the life assessment and optimization of wind turbine mainshafts and towers.