Load Forecasting in Electric Utility Integrated Resource Planning
Load Forecasting in Electric Utility Integrated Resource Planning
复制标题
电力公司综合资源规划中的负荷预测
作者：
J. Carvallo;P. Larsen;A. Sanstad;C. Goldman
作者：卡瓦洛，胡安·巴勃罗；拉森，彼得·H.；艾伦·H·桑斯塔德；高盛，查尔斯·A。摘要：综合资源规划（IRP）是许多垂直一体化的美国电力公司使用的一个过程，用于确定成本/风险最低的供应和需求方资源，以满足政府政策目标和对客户（在许多情况下）对股东的未来义务。能源和峰值需求的预测是 IRP 流程的关键组成部分。对于 IRP 长期（二十年的规划期限）负荷预测性能及其与资源规划和实际采购决策的关系的定量研究（如果有的话）很少。在本文中，我们通过检查和比较 2000 年代初提交的计划与截至 2014 年的最新计划，评估了美国西部 12 家公用事业公司的负荷预测方法、假设和结果。我们发现所使用的方法和数据源有趋同性。我们还发现，最近的 IRP 中的预测通常考虑了新信息，但在研究期间仍然存在对负载增长率的系统性高估。我们将计划和采购的资源扩张与客户负载和逐年负载增长率进行比较，但没有发现直接关系。即使存在较大的预测误差，资源计划中执行的负载敏感性似乎也与以后的采购策略无关。这些发现表明，资源采购决策可能是由客户负载增长以外的其他因素驱动的。我们的结果对综合资源规划过程具有重要意义，即负荷预测的准确性对于资源采购可能并不像人们普遍认为的那样重要，负荷预测的敏感性可用于改进采购过程，负荷不确定性的管理应优先于更复杂的预测技术。
Author(s): Carvallo, Juan Pablo; Larsen, Peter H.; Sanstad, Alan H; Goldman, Charles A. | Abstract: Integrated resource planning (IRP) is a process used by many vertically-integrated U.S. electric utilities to determine least-cost/risk supply and demand-side resources that meet government policy objectives and future obligations to customers and, in many cases, shareholders. Forecasts of energy and peak demand are a critical component of the IRP process. There have been few, if any, quantitative studies of IRP long-run (planning horizons of two decades) load forecast performance and its relationship to resource planning and actual procurement decisions. In this paper, we evaluate load forecasting methods, assumptions, and outcomes for 12 Western U.S. utilities by examining and comparing plans filed in the early 2000s against recent plans, up to year 2014. We find a convergence in the methods and data sources used. We also find that forecasts in more recent IRPs generally took account of new information, but that there continued to be a systematic over-estimation of load growth rates during the period studied. We compare planned and procured resource expansion against customer load and year-to-year load growth rates, but do not find a direct relationship. Load sensitivities performed in resource plans do not appear to be related to later procurement strategies even in the presence of large forecast errors. These findings suggest that resource procurement decisions may be driven by other factors than customer load growth. Our results have important implications for the integrated resource planning process, namely that load forecast accuracy may not be as important for resource procurement as is generally believed, that load forecast sensitivities could be used to improve the procurement process, and that management of load uncertainty should be prioritized over more complex forecasting techniques.