This paper presents an operating model of a station that charges a single-depot and homogeneous fleet of electric vehicles (EV) performing deliveries. Both the charging station and the fleet of EVs are owned by a single delivery company. The operating model, whose goal is to find the optimal EV battery charging schedule, is based on a clustering technique that keeps track of the number of EVs with a specific battery state of energy (SoE), while considering battery degradation, variable (dis)charging efficiency and nonlinear charging speed. The proposed operating model can be used both for the day-ahead scheduling and for the intraday model-predictive-control-based adjustments. Due to its relatively low capacity as compared to other market participants, the charging station is considered to be a price taker in both markets. The price uncertainty is considered using the robust uncertainty budget. The paper also evaluates inefficiency of a commonly used charging policy, i.e., the baseline model, where every delivery vehicle is charged to at least a predetermined SoE before its departure. The presented model is evaluated using multiple case studies and sensitivity analysis. As opposed to the non-clustered baseline model, the proposed approach scales well for very large fleets. Our analysis confirms that the results of the baseline model depend on the preset SoE at departure, while the proposed model provides optimal solution without assumptions on the departing SoE.

中文翻译：

---

基于集群的电动汽车单车厂充电模型

本文介绍了一个车站的运营模型，该车站为执行交付的单一仓库和同质电动汽车 (EV) 车队充电。充电站和电动汽车车队都由一家送货公司所有。该操作模型的目标是找到最佳的 EV 电池充电时间表，它基于一种聚类技术，该技术跟踪具有特定电池能量状态 (SoE) 的 EV 数量，同时考虑电池退化、变量 (dis)充电效率和非线性充电速度。建议的操作模型既可用于日前调度，也可用于基于预测控制的日内模型调整。由于与其他市场参与者相比，其容量相对较低，充电站被认为是两个市场的价格接受者。使用稳健的不确定性预算来考虑价格的不确定性。该论文还评估了常用收费政策的低效率，即基线模型，其中每辆送货车辆在出发前至少按预定的 SoE 收费。所提出的模型是使用多个案例研究和敏感性分析来评估的。与非集群基线模型相反，所提出的方法适用于非常大的车队。我们的分析证实，基线模型的结果取决于出发时的预设 SoE，而所提出的模型提供了最佳解决方案，而无需对出发 SoE 进行假设。每辆运载工具在出发前至少按预定的 SoE 收费。所提出的模型是使用多个案例研究和敏感性分析来评估的。与非集群基线模型相反，所提出的方法适用于非常大的车队。我们的分析证实，基线模型的结果取决于出发时的预设 SoE，而所提出的模型提供了最佳解决方案，而无需对出发 SoE 进行假设。每辆运载工具在出发前至少按预定的 SoE 收费。所提出的模型是使用多个案例研究和敏感性分析来评估的。与非集群基线模型相反，所提出的方法适用于非常大的车队。我们的分析证实，基线模型的结果取决于出发时的预设 SoE，而所提出的模型提供了最佳解决方案，而无需对出发 SoE 进行假设。