Solving the optimal power flow problem in a cooperative way between power generators and industrial users can reduce electricity costs more effectively than price-based demand side management. In previous work, we recently developed a decomposition-based algorithm based on Benders-type cuts to solve cooperative optimal power flow problems, which allows to protect sensitive load information. However, the algorithm suffers from unfavorable scaling behavior when the number of cooperating load entities increases. Herein, we improve the quality of the cutting planes that cooperating electricity users generate for the grid in lieu of sharing their dynamic process models. After comparing different cutting strategies and combinations thereof, we develop a tailored cutting strategy. This strategy improves the scaling behavior of our decomposition-based algorithm with multiple cooperating electricity user locations drastically. We obtain speedups factors up to almost 60 compared to our algorithm with the initial cutting strategy.

与基于价格的需求侧管理相比，以发电商和工业用户之间的合作方式解决最优潮流问题可以更有效地降低电力成本。在之前的工作中，我们最近开发了一种基于 Benders 型切割的基于分解的算法来解决协作最优潮流问题，从而保护敏感的负载信息。然而，当协作负载实体的数量增加时，该算法会受到不利的缩放行为的影响。在这里，我们提高了合作电力用户为电网生成的切割平面的质量，而不是共享他们的动态过程模型。在比较不同的切割策略及其组合后，我们制定了量身定制的切割策略。这种策略极大地改善了我们基于分解的算法与多个合作电力用户位置的缩放行为。与采用初始切割策略的算法相比，我们获得了高达近 60 的加速因子。