摘要
为更精准地进行车内压力波动控制,需要预测高速列车通过隧道时车外隧道压力波的实时变化值。在对列车历史运行重复隧道压力波数据的分析基础上,采用工况匹配(WCM)与加权K最近邻(WKNN)算法从数据库中选取若干与本次工况相接近的运行状态数据,并根据相似程度确定数据权重,构建预测用的历史数据。分别采用差分自回归滑动平均(ARIMA)与BP神经网络(BPNN)模型对隧道压力波进行预测,并将两种预测结果并联考虑,形成ARIMA-BPNN隧道压力波组合预测模型。利用武广客运专线某隧道压力波实测数据进行仿真。仿真结果表明:与WCM-WKNN-ARIMA及WCM-WKNN-BPNN单一预测模型以及WCM-ARIMA-BPNN组合预测模型相比,所建立组合模型能有效对隧道压力波进行预测,且能够取得更高精度的预测结果。
In order to control the pressure fluctuation inside the vehicle more accurately,it is necessary to predict the real-time variation of the pressure wave outside the tunnel when the high-speed train passes through the tunnel.Based on the analysis of repeated tunnel pressure wave data of train historical operation,several running state data close to this operating condition are selected from the database by using working condition matching(WCM)and weighted K-nearest neighbor(WKNN)algorithm,and the data weight is determined according to the degree of similarity to construct the historical data for prediction.Autoregressive integrated moving average(ARIMA)and back propagation neural network(BPNN)model are used to predict tunnel pressure wave respectively,and the two prediction results are considered in parallel to form an ARIMA-BPNN tunnel pressure wave combination prediction model.The simulation is carried out by using the measured data of pressure wave in a tunnel of Wuhan-Guangzhou passenger dedicated line. The simulation results show thatcompared with the single prediction model of WCM-WKNN-ARIMA and WCM-WKNN-BPNN and thecombined prediction model of WCM-ARIMA-BPNN, the combined model can effectively predict the tunnelpressure wave and obtain more accurate prediction results.
作者
陈春俊
杨露
何智颖
周林春
CHEN Chunjun;YANG Lu;HE Zhiying;ZHOU Linchun(School of Mechanical Engineering,Southwest Jiaotong University,Chengdu 610031,China;Technology and Equipment of Rail Transit Operation and Maintenance Key Laboratory of Sichuan Province,Chengdu 610031,China)
出处
《中国测试》
CAS
北大核心
2021年第10期80-86,共7页
China Measurement & Test
基金
国家自然科学基金资助项目(51975487)
轨道交通运维技术与装备四川省重点实验室开放基金课题(2019YW003)。
