Decentralized DLMPs with synergetic resource optimization and convergence acceleration

Dynamic pricing has been proposed for scheduling and controlling distributed energy resources (DERs) to mitigate operational problems in distribution networks. In this paper, we employ distribution locational marginal prices (DLMPs) to optimally schedule DERs, considering both line and voltage constraints.

DLMPs are calculated in an iterative, decentralized manner to respect user privacy, using dual decomposition and the subgradient method. The inclusion of generating DERs in the problem formulation, and the nature of the subgradient method, may significantly increase the amount of required iterations.

Recognizing the importance of convergence speed in such a mechanism, we act upon this by proposing a novel concept to identify and remove redundant constraints of the problem. The method is based on network topology observations in radial distribution networks. Redundant constraints are mapped to the respective Lagrange multipliers and are set to zero, thereby significantly increasing convergence speed.

We validate our method in a 33 and a 136 bus system, showing the synergies of the co-optimization of generating and consuming DERs, and reducing the number of iterations at least threefold.