【摘要】 气候变化对城市的影响日益加剧，频发的极端温度事件导致城市电力系统供需不平衡问题凸显，精确的需求侧电力负荷预测成为提升电力系统适应性从而支持城市功能稳定性的关键。本文开发了一种适用于极端温度事件下负荷预测的组合模型，结合时间序列模型Prophet和机器学习模型XGBoost，有效表征极端温度影响下的电力负荷波动趋势。实验结果表明，相比传统单一模型，组合模型显著提高了极端温度事件下的电力负荷预测精度，在增强城市电力系统对气候变化适应性方面具有较强的有效性，从而为电力调度等电力系统应急管理工作提供了更可靠的支持。更多还原

【Abstract】 The impact of climate change on urban areas is intensifying, with frequent extreme temperature events leading to imbalances in the supply and demand of city electricity systems. Accurate demand-side electric load forecasting has become crucial for enhancing the adaptability of power systems and thus supporting the stability of urban functionalities. This paper develops a combined model suitable for load forecasting under extreme temperature conditions, integrating the time series model Prophet and the machine learning model XGBoost,to effectively characterize the fluctuating trends of electric load under the impact of extreme temperatures. Experimental results demonstrate that, compared to traditional single models, the combined model significantly improves the accuracy of electric load forecasting under extreme temperature events. This enhancement in prediction capability substantially aids in strengthening the adaptability of urban electricity systems to climate change, thereby offering more reliable support for emergency management tasks such as power dispatching in power systems.更多还原