Reconstructing the problem of galloping monitoring of traditional complex analytical mechanism into a prediction method for machine learning algorithm modeling
Yongfeng Cheng, Jingshan Han, Jing Zhang, Wei Hao · 2018
To avoid a series of shortcomings of the traditional complex analytical mechanism, this paper selects the algorithm suitable for the transmission line galloping prediction model according to the characteristics of the traverse movement and the characteristics of the related galloping history data. After mining effective information from the selected and preprocessed data, the one class SVM algorithm is used to carry out unsupervised learning of the galloping history data. The Bagging algorithm is used in the ensemble learning algorithm to learn the classifier, avoiding the over fitting, and improving the anti-noise ability of the machine learning algorithm and the noise ability of the machine learning algorithm. The generalization ability, using the K folding cross validation method to verify the data of the galloping monitoring platform of the China Electric Power Research Institute, and using Fl-score to describe the performance of the traverse galloping early warning model, and verify the effectiveness of the machine learning method used in this paper in the aspect of galloping prediction.