AUTHOR=Ye Yunfei , Xiong Xiong , Cui Yang , Yang Fan 

TITLE=Quality control algorithm of wind speed monitoring data along high-speed railway

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 11 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2023.1160302

DOI=10.3389/fenrg.2023.1160302

ISSN=2296-598X

ABSTRACT=In order to scientifically carry out the research on high-speed railway traffic safety, high-quality second level wind speed data is required as the research basis. Since the wind speed data is easily disturbed during the collection and storage process, the quality of second level wind speed data is greatly reduced. Therefore, it is a crucial link to control the quality of second level wind speed data along the high-speed railway. Based on the strong instability and non-linear characteristics of wind speed data along the high-speed railway, this study combines Convolutional Neural Network, Long Short-Term Memory and Isolated Forest from the time dimension to form a quality control algorithm for wind speed monitoring data based on CNN-LSTM-IF. Firstly, Convolution Neural Network is used to extract the features of original data; Then, the extracted original data features are transferred to the Long Short-Term Memory Network for one-step prediction, and the prediction residual of the model is obtained; Finally, the prediction residual is sent to the Isolated Forest, and the abnormal value position in the original wind speed data is calibrated by detecting the abnormal value position in the prediction residual. The comparative experiments of three different quality control methods show that the error detection rate of CNN-LSTM-IF in this research method is about 0.95, and it has certain robustness and generalization in different terrain and seasons.