Intern
Computer Science XI - Modeling and Simulation

ProFlexGrid – Probabilistic flexibility optimization in controllable low-voltage grids through sector coupling

Project Description

Achieving climate neutrality requires transitioning to a low-emission, resource-efficient energy system. A crucial step in this transition is the electrification of the heat and transportation sectors, which will lead to strong coupling between them and the electricity sector. It is important that this coupling is done intelligently, especially on the low-voltage level, where new loads can easily introduce local grid overloads. Such overloads can be avoided by intelligently utilizing the inherent flexibilities of new and conventional electric loads such as electric vehicles, heat pumps, and home energy storages. The beneficial deployment of these flexibilities can be incentivized with dynamic grid fees. Additionally, if absolutely necessary, the grid operator can also directly force a load reduction of controllable units according to §14a of the Law "Energiewirtschaftsgesetz" (EnWG). The optimal usage of these diverse influence mechanisms is highly complex. The goal of the project ProFlexGrid is to create a stochastic simulation of low-voltage grids that jointly incorporates all relevant energy sectors and allows for the evaluation of novel grid control strategies. Two cases are of special interest. First, avoiding critical grid states by scheduling optimal price incentives, and second, optimally managing grid overload scenarios should they occur. A focus of the project is the explicit modeling of uncertainties with stochastic simulation and the inclusion of these uncertainties in the developed control strategies via robust optimization. Several network sections of the city of Bayreuth, Germany, serve as case studies for the developed methodologies.

Our Contribution: Transport sector

Our contribution lies in the modeling of the transportation sector. This entails the modeling of the day-to-day mobility behavior of the population that lives in or visits locations connected to the analyzed low-voltage grid. The modeling approach is based on our open-source mobility demand simulator OMoSim. The key new modeling questions in the project are how individuals will react to dynamic prices and unforeseen  control actions by the grid operator that, for example, leave their electric vehicles at a lower state of charge than expected. An additional research question is how to formulate and possibly simplify the joint probability distributions that guide mobility behavior such that they are readily usable by robust optimization.