Research on prediction of tower mechanical response in wind field based on multi-layer perceptron

There are often strong winds and typhoons in coastal areas, which brings great risks to the safe operation of transmission lines. The study of a fast calculation method for the response of transmission lines in wind farm is of great significance to ensure the safe operation of power grid. Most of the existing research methods are simulation modeling and statistical calculation, but there are still deficiencies in computing speed and universality. In this paper, the response of transmission tower in wind field is studied. Using the principle of machine learning, the tower response model under different weather and tower operation conditions is built, and the MLP multi-layer perceptron model is used to predict the tower stress value in wind field; Smote linear interpolation is used to increase the correctness of the number of unbalanced sample data. Finally, the prediction accuracy of minority samples is improved by modifying the loss function. The prediction accuracy of minority samples is significantly improved, and the accuracy is 96.5%.