Unit Commitment Without Commitment: A Dynamic Programming Approach for Managing an Integrated Energy System Under Uncertainty
提出一种动态规划框架，通过随机拉格朗日乘子作为代理电价，将系统级问题分解为机组级问题，在不确定性下平衡供需，降低运营成本2%-5%。
Building Flexibility into Energy System Dispatch The growing use of renewable energy is forcing power system operators to grapple with increasing uncertainty and intermittency in their energy supplies and demands. In “Unit Commitment without Commitment: A Dynamic Programming Approach for Managing an Integrated Energy System Under Uncertainty,” Brown and Smith develop a dynamic programming (DP) framework for balancing supply and demand over time. The approach introduces stochastic Lagrange multipliers as surrogate energy prices, which reduces the system-wide DP to a collection of unit-specific DPs, where each unit is managed to maximize its expected profit over a long time horizon, given these uncertain prices. Real-time dispatch decisions are then made using the unit-specific value functions to capture the longer-term impacts of the dispatch decisions. Using data from the Duke Energy Carolinas and Progress systems, this new approach reduced operational costs by 2% in current systems and 4%–5% in example future scenarios with increased solar and storage capabilities. Strikingly, the proposed methods performed within 0.2%–0.3% of plans based on perfect foresight, across a wide variety of scenarios.