Conference Agenda

To efficiently transform residential building portfolios, residential building portfolio modernization roadmaps (RBPMRs) are used to provide information about which specific action to take in which building and which time period. Precisely, RBPMRs constitute transitions of individual building energy systems as well as building renovations over a given time horizon while collectively respecting resource constraints at the building portfolio level. We define the problem of determining optimal RBPMRs and present a MILP model for the problem. To accelerate the solving process, we first introduce a Benders decomposition that projects out building energy system operations into Benders subproblems, and subsequently embed a Benders cut separation algorithm into the branch-and-cut process of the commercial MILP solver Gurobi. Our approach employs various acceleration techniques, including constraint modifications, a tailored Benders optimality cut strengthening technique, and a primal construction heuristic for the Benders relaxed master problem. Additionally, we employ an in-out method to enhance the dual bound at the root node. Using data from the German Census, we construct test instances to evaluate the algorithmic implementations on residential building portfolios. Finally, we demonstrate the practical applicability of our optimization approach through a case study involving data from the residential building stock of a German city.

The security-constrained optimal transmission switching problem aims to minimize the operational costs of a power system while maintaining N-1 security. Busbar reconfiguration additionally increases the amount of possible topologies by representing the topological behavior of substations. We consider the linearized DC formulation of the problem, with minimal filtering of switching actions or contingencies. We present a Benders decomposition approach to efficiently solve the problem. In the master problem, we solve an optimal transmission switching problem, optimizing switching actions and generator re-dispatches to reduce branch overloads, without considering contingencies. Instead, for each contingency, a power flow subproblem is solved. To that end, switching and generator decisions from the master, as well as the outaged branches of the contingency, are incorporated into a power flow problem. There, the objective is to minimize the severity of branch overloads. Due to the number of subproblems, it can be beneficial to select a subset of all subproblems to solve as well as constraints to add to the master problem. This can reduce the number of subproblems solved, and limits the size of the master problem, while maintaining optimality. We discuss metrics to measure anticipated subproblem and cut relevance, and how to incorporate them into the solving process. The performance of our approaches is evaluated on modified versions of the publically available pglib-opf instances.

The Alternating Current Optimal Power Flow (AC OPF) problem has garnered significant attention in recent years. One promising approach to address this challenge is the Semidefinite Programming (SDP) relaxation of the OPF. In our study, we focus on a hierarchical system of OPF problems related to transmission and distribution networks. Communication between levels involves computing both lower and upper bounds on available power. Specifically, we explore issues arising when computing upper bounds, which involves solving a problem with a maximization objective rather than the standard minimization one. Finally, we propose potential modifications to mitigate these challenges.