Research On Short-term Wind Speed Prediction Method Along High-speed Railway Based On Machine Learning

With the continuous improvement and development of high-speed railway network,the operating environment of high-speed railway becomes more and more complex and strict.The safety of high-speed railway operation has gradually become the theme of railway development,and the biggest impact is wind disaster.At home and abroad,a strong wind monitoring and warning system has been developed for the sections prone to strong winds,but it can only realize the monitoring function of wind speed and cannot predict the change trend of wind speed in advance.Moreover,there are very few studies on short-term wind speed prediction along the high-speed railway.Therefore,to build a short-term wind speed prediction model suitable for high-speed railway is to improve the accuracy of wind speed prediction.It is an important measure to reserve more time for dispatching personnel to make early warning and slow down the train in advance so as to ensure the safety of operation.Based on this,this paper established a variety of short-term wind speed combination prediction models for high-speed railway on the basis of machine learning method,and then takes the data fragments collected along the Beijing-Zhangjiakou high-speed railway to reach the 1min interval of wind speed warning value as the object,improves the accuracy and stability of wind speed prediction,and provides theoretical support for short-term wind speed prediction research along the high-speed railway,the research content of this paper is as follows:(1)Analysis of wind speed data along the high-speed railway.Firstly,the paper introduces the relevant data of the study area and the subsequent prediction;Secondly,data is preprocessed by data cleaning method to realize outlier detection and noise filtering,which provides good data quality for subsequent prediction.Finally,the characteristics of the collected wind speed data are analyzed,and the average wind speed,wind speed characteristics,wind direction distribution,annual and diurnal variations of wind speed along the high-speed railway are displayed with the visualization method,so as to explore the spatial and temporal distribution characteristics and correlation rules of wind speed along the high-speed railway.(2)Construction of short-term wind speed prediction model based on CEEMDAN-GWO-LSSVM.Firstly,the original high-speed railway wind velocity sequence was decomposed into multiple modal subcomponents by using Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN).Then,the Least Square Support Vector Machine(LSSVM)model was used to carry out single-step prediction,and the Grey Wolf Optimization(GWO)was used to optimize the parameters.Finally,the results of each component are reconstructed to obtain the final predicted value.The results show that the CEEMDAN-GWO-LSSVM model can significantly improve the accuracy and accuracy of wind speed prediction,and the Mean Square Error(MSE),Mean Absolute Error(MAE)and Root Mean Square Error(RMSE)are the three evaluation indexes with the highest improvement degree.(3)Construction of multi-step prediction model of short-term wind speed along the high-speed railway based on CEEMDAN-VMD-EHPO-LSTM.Firstly,the CEEMDAN method was used to decompose the original high-speed iron wind velocity sequence for the first time,calculate the sample entropy of each modal component,and carry out Variational Mode Decomposition(VMD)for complex components with high sample entropy to reduce the volatility of the original wind velocity sequence.Secondly,the initial search population diversity and search range of the Hunter Prey Optimizer(HPO)are improved by combining Tent chaotic mapping and Adaptive Inertia Weight(AIW).Finally,the Long Short Term Memory(LSTM)is used to predict the short term wind speed along the high-speed railway in one,three and five steps.The results show that the CEEMDAN-VMD-EHPO-LSTM model has a good overall prediction effect,solves the problem of prediction delay,and realizes the advanced multi-step prediction of short-term wind speed along the high-speed railway.(4)Construction of short-term wind speed prediction model along the high-speed railway based on multi-objective optimization.Following the single-step prediction model in(2),two optimization objective functions of MSE and standard deviation were set in combination with the Multi Objective Ant Lion Optimization(MOALO),and weight assignment was carried out.The results show that the solution performance of the algorithm based on multi-objective ant lion optimization is better than other algorithms and has stronger search performance.The proposed multi-objective optimization prediction model can effectively solve the multi-objective optimization problem in the short-term wind speed prediction of high-speed railway,and improve the accuracy and stability of wind speed prediction.In this paper,single-step,multi-step and multi-objective models for short-term wind speed prediction along high-speed railways are constructed respectively.The decomposition,optimization and prediction methods used in the models all show good performance,improve the accuracy and stability of wind speed prediction in advance,and provide feasible ideas for short-term wind speed prediction along high-speed railways.