A Neural Network-Based Evaluation Model for Longitudinal Forces in Heavy-Haul Trains
[Purposes]The longitudinal force between vehicles can directly indicate the magnitude of the longitudinal impulse of the train.However,the traditional longitudinal force mechanism model is highly coupled and complex to solve.Therefore,a neural network-based evaluation model for longitudinal force of heavy-haul trains is proposed.[Methods]The inter-vehicle longitudinal forces can directly reflect the magnitude of the longitudinal impulses.However,each vehicle's state is highly coupled and computa-tionally complex under the traditional coupling force mechanism mode.Therefore,this article proposes a neural network-based longitudinal force evaluation model,summarizes various factors related to longitu-dinal force which were designed as sample features and constructs a longitudinal force evaluation model for heavy haul trains.[Findings]The final simulation demonstrated that when the error between the evalu-ated longitudinal force values and the actual values was less than 100 kN,the accuracy of the model reached 92.4%.[Conclusions]Using neural network as the evaluation model of the longitudinal force of the heavy-haul train can effectively and accurately predict the coupler force and provide reference for the safe and stable operation of the heavy-haul train.
heavy-haul trainlongitudinal forceneural networkevaluation model
NETL
NSTL
万方数据
