Probabilistic forecasts provide a distribution of possible outputs and so can capture the uncertainty and variability of Variable Renewable Energy (VRE). However, taking advantage of uncertainty information has practical challenges that make it difficult to integrate probabilistic forecasting into control room decision-making. This paper proposes a novel use-case for probabilistic forecasts by incorporating them into the hour-ahead operations for situational awareness via a risk-averse multi-stage stochastic program. We employ a Markovian representation of the probabilistic forecasts that enables the formulation of the multi-stage problem and avoids a scenario generation phase. We test the model on a realistically sized system to assess risk and showcase the capability of using probabilistic renewable forecast as input to produce probabilistic output forecasts of future system states. The results show that the model can capture time consistency in the reserves and Area Control Error (ACE) forecast. The solution times are adequate for risk profiling in hour-ahead timescales.

中文翻译：

---

可再生能源的马尔可夫表示的多阶段随机风险评估

概率预测提供了一个 可能输出的分布，因此可以捕捉可变可再生能源 (VRE) 的不确定性和可变性。然而，利用不确定性信息存在实际挑战，这使得将概率预测集成到控制室决策中变得困难。本文提出了一种新的概率预测用例，通过风险规避的多阶段随机程序将概率预测整合到用于态势感知的提前一小时操作中。我们采用概率预测的马尔可夫表示，从而能够制定多阶段问题并避免情景生成阶段。我们在实际大小的系统上测试模型，以评估风险并展示使用概率可再生预测作为输入来生成未来系统状态的概率输出预测的能力。结果表明，该模型可以捕捉储量和区域控制误差（ACE）预测中的时间一致性。解决时间足以进行提前一小时的风险分析。