基于神经网络的汽车行驶工况识别方法研究  

Research on Driving Cycle Recognition Based on Neural Network

作　　者：钱星[1,2] 张春英 陈圆圆[3] 王勇 靳玉刚 江杰[1,2] 丁波 QIAN Xing;ZHANG Chunying;CHEN Yuanyuan;WANG Yong;JIN Yugang;JIANG Jie;DING Bo(FAW Wuxi Fuel Inject Equipment Research Institute,Wuxi 214063,China;Advanced Technology Research Institute,FAW Jiefang Engine Bus iness Division,Wuxi 214063,China;Zhejiang Xinchai Co.,Ltd.,Xinchang 312500,China)

机构地区：[1]中国一汽无锡油泵油嘴研究所,江苏无锡214063 [2]一汽解放发动机事业部前瞻技术研究院,江苏无锡214063 [3]浙江新柴股份有限公司,浙江新昌312500

出　　处：《现代车用动力》2022年第2期7-12,共6页Modern Vehicle Power

摘　　要：混动车辆能量管理实现自适应控制策略的前提是对汽车行驶工况的正确识别。针对此问题提出了一种基于神经网络的车辆行驶工况识别方法。首先通过采集车辆实际行驶路谱,整理出3种典型工况。在此基础上,运用方差分析工具,筛选提取10项特征,用于路谱的辨别,并且整理成用于训练的数据。采用Matlab建立目标神经网络,经过训练和测试,模型识别率达到98.5%。经过综合工况的验证,模型识别率良好,达到预期效果。In order to realize the adaptive control strategy of hybrid vehicle,the driving cycle recognition is required.A neural network algorithm is proposed to recognize the driving cycle.Firstly,three typical driving cycles are obtained by collecting the actual driving cycles.The 10 characteristic parameters of driving cycle are extracted by the variance analysis,and the training data is compiled.The target neural network was established by the Matlab,the model recognition rate reached 98.5%after training and test.Through the validation of the comprehensive driving cycle,the model has the high identification effect and reaches the expectation.

关 键 词：商用车 增程式 混合动力 GT-SUITE 中国汽车行驶工况循环 

分 类 号：TK423[动力工程及工程热物理—动力机械及工程]