摘要
研究了基于误差逆传播网络(BPNN)和长短时记忆神经网络(LSTM)的组合式预测模型及其在脱轨系数预测中的应用。通过在SIMPACK多体力学仿真软件上建立的车-轨系统仿真场景,形成深度学习训练数据集。使用Python语言在TensorFlow框架上开发单项网络,基于仿真数据集,对单项预测模型BPNN和LSTM展开训练。评估单项模型的预测精度,使用加权平均法对单项模型进行组合,建立组合式脱轨系数预测模型并分析模型预测性能。结果表明,相较于单项模型,组合式预测模型能更准确地预测出脱轨系数的变化趋势。
This paper studies the combined prediction model based on error back propagation network(BPNN)and long short-term memory neural network(LSTM)and its application in the prediction of derailment coefficients.A deep learning training data set is formed through the vehicle-track system simulation scene established on the SIMPACK multi-body mechanics simulation software.Use the Python language to develop a single-item network on the TensorFlow framework,and train the single-item prediction models BPNN and LSTM based on the simulation data set.Evaluate the prediction accuracy of the single model,use the weighted average method to combine the single models,establish the combined derailment coefficient prediction model and analyze the model prediction performance.The results show that the combined prediction model can predict the change trend of derailment coefficients more accurately than the single model.
作者
张卜
刘怡伶
张文静
Zhang Bu;Liu Yiling;Zhang Wenjing(School of Mechanical and Automobile Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)
出处
《农业装备与车辆工程》
2022年第7期52-56,共5页
Agricultural Equipment & Vehicle Engineering
关键词
列车
脱轨系数
组合式
预测模型
误差逆传播网络
长短时记忆网络
train
derailment coefficient
combined type
prediction model
back propagation network
long and short-term memory network
