基于矩不确定分布鲁棒优化的发电自调度算法

摘要

针对发电商在电力市场中的自调度计划问题,对模型中节点边际电价的不确定性进行了研究.提出了基于矩不确定分布鲁棒优化方法(DRO-MU),用于解决含不确定节点边际电价的发电自调度问题.建立了以收益最大为目标的发电自调度模型,构建了LMPs的不确定集合,在该不确定集合下将发电自调度模型转化为DRO-MU模型,再通过拉格朗日对偶原理将模型转化为一个确定性的半正定规划进行了求解.研究结果表明,随着不确定集范围的增大,发电公司收益减少;与确定性调度计划计算结果相比,DRO-MU模型收益少,但方案安全性更高、鲁棒性更强.%Aiming at the self-scheduling problem of the generation companies in the electricity market, the uncertainty of locational marginal prices ( LMPs) was studied. A distributional robust optimization under moment uncertainty ( DRO-MU) was proposed to solve the problem. The generation self-scheduling model with the objective of maximum profit was established to construct the uncertain set of LMPs. Under the uncertain set, the self-scheduling model was transformed into the DRO-MU model, and then the model was transformed into a deterministic semidefinite programming model by Lagrangian duality. The results indicate that with the increase of the uncertainty set, the profit of genera-tion companies decreases. Compared with the results of deterministic scheduling, the DRO-MU model yields less, but the scheme is more se-cure and more robust.