Automatic dimensioning of energy system components for building cluster simulation

In this paper, we present an approach on automatic energy system modeling and simulation. We develop two different methods to dimension the components of energy systems: one approach is an easy-to-use and to-adapt rule-based method, where the size of components is based e.g. on the heat demand of the buildings. The second approach is to dimension components with a genetic algorithm with a target function to reduce total annual cost. We apply and compare the methods for two different system designs to a building cluster case study in Germany. One system is a decentral heat pump system with back-up gas boiler, thermal and electrical storage and PV, the other system is a hydrogen-based central fuel cell system with electrolyzer, thermal and electrical storage and PV. We compare both systems based on current (2020) and future (2050) framework conditions.Through the application of the genetic algorithm a reduction of equivalent annual cost of up to 27% (2020 scenario) and 21% (2050 scenario) is achieved for the heat pump system. The hydrogen system with optimized dimensioning becomes economically viable under 2050 conditions due to reduced fuel prices and a higher electricity feed-in tariff compared to 2020. Depending on the use case, both approaches have their merit: rule-based dimensioning can quickly simulate a variety of different scenarios, the genetic algorithm can achieve economically ideal system designs, but with longer computation time, especially with unfamiliar framework conditions.