Distribution network electricity market clearing: Parallelized PMP algorithms with minimal coordination
Distribution network electricity market clearing: Parallelized PMP algorithms with minimal coordination
复制标题
配电网电力市场清算：最小协调的并行 PMP 算法
DOI：
--
复制
发表时间：
2014
期刊：
影响因子：
--
通讯作者：
M. Caramanis
中科院分区：
文献类型：
--
作者：
Elli Ntakou;M. Caramanis
社会最优的电力市场清算问题与不同的，复杂的公用事业结构的参与者在分布水平，提出了计算的挑战，加剧了相关的非凸潮流约束。我们调查的变化和扩展的近端消息传递算法在文献中提出，并实现类似的，但更简单，社会福利函数的实例。数值结果表明，（i）与解决计算要求高但精确的集中式市场清算问题相比，所提出的PMP算法扩展可以显著提高计算效率，（ii）与通过求解集中式公式获得的基准结果的比较表明，PMP算法可以实现出色的准确性，以及（iii）与已有的PMP算法相比，该算法大大降低了子问题协调和收敛性验证的通信需求，并能实现更快收敛的异步子问题迭代。
The socially optimal power market clearing problem with diverse, complex-utility-structure participants at the distribution level, poses computational challenges that are exacerbated by the associated non-convex load flow constraints. We investigate variations and extensions of Proximal Message Passing algorithms proposed in the literature and implemented on similar, though simpler, social welfare function instantiations. Numerical results demonstrate that (i) in comparison to solving a computationally demanding, yet exact, centralized market clearing problem, significant computational improvements are possible with the proposed PMP algorithm extensions, (ii) comparison to the benchmark results obtained by solving the centralized formulation reveals that excellent accuracy is attainable by the PMP algorithms and (iii) in comparison to the PMP algorithms existing in the literature, the proposed extension reduces significantly the communication requirements for sub-problem coordination and convergence verification and enables faster converging asynchronous sub-problem iterations.