A non-centralized model predictive control (MPC) scheme for solving an economic dispatch problem of electrical networks is proposed in this paper. The scheme consists of two parts. The first part is an event-triggered repartitioning method that splits the network into a fixed number of non-overlapping sub-systems {(microgrids)}. The objective of the repartitioning procedure is to obtain self-sufficient microgrids, i.e., those that can meet their local loads using their own generation units. However, since the algorithm does not guarantee that all the resulting microgrids are self-sufficient, the microgrids that are not self-sufficient must then form a coalition with some of their neighboring microgrids. This process becomes the second part of the scheme. By performing the coalition formation, we can decompose the economic dispatch problem of the network into coalition-based sub-problems such that each subproblem is feasible. Furthermore, we also show that the solution obtained by solving the coalition-based sub-problems is a feasible but sub-optimal solution to the centralized problem. Additionally, some numerical simulations are also carried out to show the effectiveness of the proposed method.
翻译:本文提出了解决电力网络经济发送问题的非集中模型预测控制(MPC)计划。 计划由两部分组成。 第一部分是将网络分裂成固定数量的非重叠子系统{(microgrims)}。 重新分割程序的目标是获得自给自足的微格, 即那些能够利用自己的发电单位满足本地负荷的微格, 但是, 由于算法不能保证所有由此产生的微格网络都自给自足, 因而不能自给自足的微格网络必须随后与其某些相邻的微格网组成联盟。 这个过程成为计划的第二部分。 通过执行联盟的形成, 我们可以将网络的经济发送问题分解为基于联盟的子问题, 这样每个子问题都是可行的。 此外, 我们还表明,通过解决基于联盟的子问题而获得的解决办法是可行的,但是, 无法自给自足的微格网必须随后与一些相邻的微格网组成联盟。 这个过程将成为计划的第二部分。 通过进行联合组建, 我们可以将网络的经济发送问题分解成基于联盟的子问题。 此外, 我们还表明, 通过解决以联盟为基础的子问题而获得的解决办法是可行的, 并且也是对中央问题进行模拟的一些数字模拟的方法。