BP Neural Network-based Prediction of Bolt Tightening Force for Wind Turbine Tower Flange
The present work aimed to addressing high cost and low efficiency in full-coverage real-time monitoring of wind turbine bolts,and proposed a back propagation(BP)neural network model to predict the bolt tightening force of wind tur-bine tower flange.The network consisted of a four-layer network structure,and took the bolt tightening force value of each monitoring point on the wind turbine flange as the input of the network,iterated in turn through forward propagation and reverse propagation,and thereby completed the model training process.The proposed BP neural network-based pre-dictive model was verified by simulation data to achieve an increase in correlation,an increase in absolute coefficient,and a reduction in root mean square error by 6.73%,15.53%,and59.95%,respectively,compared with the conventional inter-polation algorithm when predicting bolt tightening force on wind tower flange.
wind power generatortower flange bolttightening force predictionBP neural network
NETL
NSTL
万方数据
