Research On Energy-saving High-speed Train Driving Strategy Prediction Method

Reducing the operating energy consumption of high-speed trains is an important way to improve the energy efficiency of rail transit and reduce pollution emissions.In order to promote the energy conservation and emission reduction of rail transit and deeply practice the green concept,in view of the characteristics that the controller handle transformation will have a significant impact on the energy consumption of train operation,this study proposes to transform the energy saving and consumption reduction problem of high-speed rail into the problem of classification of driver controller handles,and construct two neural network models to solve the energy-saving driving reference strategy of high-speed trains.Under the premise of ensuring the safety and punctuality of the train,the purpose of energy saving and emission reduction is achieved,and the comfort of passengers is improved to a certain extent.The main research contents and conclusions are as follows:(1)A GRU model based on time mode attention mechanism and simulated annealing algorithm is constructed to solve the energy-saving driving strategy of offline high-speed train.This model considers the influence of controller handle level on the driving of high-speed train,and transforms the problem of energy saving and consumption reduction into the problem of classification of controller handle level.In terms of specific implementation,data cleaning is first required to obtain data for training and testing.In general,data cleansing includes basic preprocessing and feature selection.In this study,the downsampling method was used to construct the balanced dataset,and the multiple feature selection method was used to remove redundant data to improve the accuracy and stability of the model prediction results.Finally,the model is validated and evaluated.In this study,the simulation road segment is used to verify that the energy consumption of the constructed model and other existing models is compared,and it is proved that the prediction strategy of this model has better energy saving effect than other common existing model prediction strategies.Driving with this strategy saves about 17.86%on short routes and 17.98%on long routes.This shows that an effective offline high-speed train energy-saving driving strategy can be obtained by the GRU model based on the time mode attention mechanism and the simulated anncaling algorithm.(2)A fusion network classification model is constructed to solve the energy-saving driving strategy of real-time high-speed trains.Considering the delay of high-speed trains,the fusion network classification model can select the corresponding dataset for training according to the current different operating conditions of the train,so as to achieve real-time prediction.Compared with the GRU model based on temporal mode attention mechanism and simulated annealing algorithm,the fusion network classification model has fewer iterations and faster raining efficiency.In the simulation results,the fusion network classification model can obtain the current driving reference strategy in a shorter time under the premise of satisfying passenger comfort,and the prediction strategy of the model has a good energy-saving effect,while meeting the real-time strategy prediction,the enargy saving is about 15.30%(short route)and 15.53%(long route),therefore,the model can provide efficient and reliable driving reference strategy for high-speed trains.(3)Based on the construction prediction model,the energy-saving driving assistance system of high-speed rail is designed to help drivers better understand the operation status of che train,and then formulate operation stratcgjes more accurately.