Research On Gnss-based Train Integrity Inspection Simulation Method

With the rapid development of railway construction in China,the expansion of railway capacity calls for heavy-haul and fast goods transport network.The acceleration and deceleration of the wagons make the coupler device connected to the wagon subject to frequent tension and extrusion.There is a risk of breaking the coupler device,leaving the wagons on the track,endangering the traffic safety.Train integrity inspection refers to the inspection of the train during operation to ensure that the train does not decouple and ensure the overall integrity of the train.Train integrity inspection system is an important subsystem in train operation control system.CTCS-N(Chinese Train Control System-New)requires the GNSS(Global Navigation Satellite System)and other sensors to deliver sensor fusion train location,thus achieving autonomous train localization and train integrity inspection.Before the field application of GNSS-based EOT(End of Train)equipment,there are some harsh situations such as difficulties to reproduce EOT equipment working scenario,GNSS signal cannot be controlled at will,etc.Therefore,it is necessary to test the GNSS localization function and performance metrics under the given working environment in laboratory using simulation methods before EOT equipment is put into practical application,getting rid of the test requirements in the field.The main research of this thesis is:(1)The modeling method of EOT GNSS signal reception constrained environment scenario is studied.Based on the real GNSS signal propagation environment of EOT,the characteristics of visible satellites for EOT is analyzed.The modeling method of EOT GNSS signal reception constrained environment scenario is proposed using Apriori algorithm.The GNSS signal reception receiving scenarios of EOT along the track are clustered using hierarchical clustering algorithm,and the scenario database of EOT is established.The model is validated by field test data.The simulation method of GNSS localization scenario based on Skyplot database for EOT is proposed.(2)The train integrity inspection method based on BP neural network is studied.Considering the influence of GNSS positioning error on the train integrity inspection,the estimation deviation of train length is calculated using error ellipse,the variations of velocity difference between HOT(Head of Train)and EOT is studied using moving time window.Train integrity inspection model is designed based on BP neural network,HOT and EOT GNSS localization data key parameters are extracted to train the samples.The result shows accuracy of the model can reach 99.55%.(3)GTIS(GNSS-based Train Integrity Inspection Software)is designed and developed.GTIS includes train spatial-temporal scenario simulation module,GNSS navigation signal control module,EOT GNSS signal reception constrained environment scenario generation module and train integrity inspection module.GTIS can control the GNSS signal simulator generating GNSS signal for HOT and EOT,especially using the EOT Skyplot database to generate the signal reception constrained environment.GTIS collects the HOT and EOT GNSS localization data and demonstrate train integrity inspection algorithm proposed in this thesis.(4)Based on the Haoji heavy-haul railway field test data,the simulation scenario experiment is designed to validate the simulation function and performance metrics of the constrained EOT environment and the train integrity inspection method proposed in this thesis.There are 78 pictures,27 tables and 50 references.