Battery electric vehicle (BEV) has been thought as a key factor in decreasing global greenhouse gas emissions and energy conservation. Therefore, recent developments in BEVs have heightened the need for energy consumption prediction. In this paper, an effectively model in virtue of the accurate velocity prediction approach is proposed. Prior to commencing the study, road information and historical driving data are sought from electronic map and realistic road tests respectively. Following processing for these raw data, a semi-physical and semi-empirical model is introduced to tackle the problem for energy consumption calculation. For the velocity prediction, a Markov-chain-based method in conjunction with road information is proposed, which endows energy consumption model with precise velocity profile as input to obtain final results. The feasibility and precision of this method were validated on various road types with acceptable results exhibiting a mean error of less than 2%, highlighting its anticipated preferable performance.

电动汽车（BEV）被认为是减少全球温室气体排放和节约能源的关键因素。因此，BEV的最新发展增加了对能耗预测的需求。本文提出了一种基于精确速度预测方法的有效模型。在开始研究之前，分别从电子地图和现实道路测试中寻求道路信息和历史驾驶数据。在处理这些原始数据之后，引入了半物理和半经验模型来解决能耗计算问题。对于速度预测，提出了一种基于马尔可夫链的结合道路信息的方法，该方法赋予能量消耗模型以精确的速度分布作为输入以获得最终结果。