电动汽车智能充电路径规划算法研究

发布时间：2019-04-03 16:52

【摘要】：随着汽车产业的快速成长,传统的燃油发动机汽车带来的环境污染与能源危机问题日渐凸显,新能源汽车成为未来汽车工业的重要发展方向之一。以电动汽车为代表的新能源汽车要代替传统内燃机汽车,尚存在电池续航能力等问题有待处理。现有电动汽车充电时间长,而公共充电桩资源有限,长时间在外运行的电动汽车会出现何时前往何地充电的路径规划问题。为此,本文针对“充电焦虑”问题,主要研究在有限的充电设施网络之上,如何为电动汽车提供智能充电路径规划,以期提高充电效率,降低排除时间,促进新能源汽车的发展。在现有研究成果基础之上,本文首先针对电动汽车剩余电量(SOC)的估算及其与可行驶距离的关系进行了分析,建立起了支持向量机(SVM)预测模型,同时基于信息粒化,将实际驾驶过程中人为因素与交通状况等影响SOC的外在因子也一并考虑,从而得出更为精准可预测SOC波动范围及其与可行驶距离的关系模型。在此基础上,对实时工况下电动汽车的充电路径规划问题抽象为实时路况下的旅行商问题(TSP),通过对时间的离散化,将动态路况简化为阶段性静态路径规划,给出了基于“Dijkstra+模拟退火”的路径规划算法,保证在当前时间段内路径规划最短,同时采用“模糊+精确”两段式计算,以缓解充电桩的排队问题。最后,建立起仿真实验平台,通过对相关算法的对比实验,对相关算法的有效性进行了验证。
[Abstract]:With the rapid growth of automobile industry, the problems of environmental pollution and energy crisis caused by traditional fuel engine vehicles are becoming more and more prominent. New energy vehicles have become one of the important development directions of the future automobile industry. In order to replace the traditional internal combustion engine vehicle, the new energy vehicle represented by electric vehicle still has some problems such as battery life ability and so on. The existing electric vehicle has a long charging time, but the public charging pile resource is limited, and the long-running electric vehicle will have the problem of when to go and where to charge the electric vehicle. Therefore, aiming at the problem of "charging anxiety", this paper mainly studies how to provide intelligent charging path planning for electric vehicles on the limited charging facility network, in order to improve the charging efficiency and reduce the elimination time. Promote the development of new energy vehicles. On the basis of the existing research results, this paper firstly analyzes the estimation of the residual power (SOC) of electric vehicles and its relationship with the driving distance, and establishes a (SVM) prediction model based on support vector machine, at the same time, based on the information granulation. The human factors in the driving process and the external factors that affect the SOC, such as traffic conditions, are also considered, so as to obtain a more accurate and predictable fluctuation range of the SOC and the relationship model between the fluctuation range and the driving distance. On this basis, the charging path planning problem for electric vehicles under real-time working conditions is abstracted as the traveling salesman problem under real-time conditions. By discretization of the time, the dynamic road condition is simplified to the stage static path planning by (TSP), and the dynamic road condition is simplified to the stage static path planning. This paper presents a path planning algorithm based on "Dijkstra simulated annealing", which ensures the shortest path planning in the current time period and adopts the "fuzzy and accurate" two-stage calculation to alleviate the queuing problem of charging piles. Finally, the simulation platform is set up, and the validity of the related algorithm is verified by the comparison experiment of the related algorithms.
【学位授予单位】：山东理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.8

【参考文献】