针对高速公路服务区充电负荷直接受到车辆到达过程约束，与城市场景存在差异的特点，以北京市某高速公路服务区为研究对象，开展小时级充电负荷预测研究。基于真实运行数据，构建融合历史负荷与外生变量的 TimeXer 预测框架，并重点引入驶入电动汽车数量变量；进而通过候选变量筛选、逐步消融和基线对比，分析不同外生变量对预测性能的影响。结果表明，在当前样本条件下，驶入电动汽车数量对预测精度提升最为显著，相较仅使用历史负荷序列的模型，进京方向和出京方向的预测误差均明显下降，而天气变量带来的增益相对有限。与 Historical Average、ARIMA、LSTM、Transformer 及仅使用内生变量的 TimeXer 等基线模型相比，所构建模型在进出京两方向上均取得更优结果。研究表明，在本案例场景下，结合车辆到达信息组织外生变量输入能有效提升高速公路服务区小时级充电负荷预测性能，可为服务区充电设施运行管理提供参考。

Considering that charging load at expressway service areas is directly constrained by vehicle arrival processes and differs from that in urban charging scenarios, this study investigates hourly charging load forecasting by taking an expressway service area in Beijing as the case study. Based on real operational data, a TimeXer-based forecasting framework integrating historical load and exogenous variables is developed, with particular emphasis on the number of arriving electric vehicles as a traffic-arrival-related variable. Through candidate-variable screening, stepwise ablation, and baseline comparison, the effects of different types of exogenous variables on forecasting performance are analyzed. The results indicate that, under the current sample conditions, the number of arriving electric vehicles contributes most significantly to improving prediction accuracy. Compared with the model using only historical load sequences, the prediction errors for both the inbound and outbound directions toward Beijing decrease significantly, while the gains from weather variables are relatively limited. Compared with baseline models including Historical Average, ARIMA, LSTM, Transformer, and TimeXer with only endogenous variables, the proposed model achieves better performance in both the inbound and outbound directions. The findings indicate that, in this case study, organizing exogenous variable inputs by incorporating vehicle arrival information can effectively improve hourly charging-load forecasting performance for expressway service areas and provide a reference for the operation and management of charging facilities in such areas.