Optimizing large-scale Integrated Energy Systems with Storage Facilities: A parallelized approach

Achieving the necessary energy transition efficiently and effectively hinges on designing an energy system that meets both generation and demand requirements. In addition to the sustainable supply of energy in the form of electricity, natural gas as well as hydrogen, a customized energy infrastructure should be developed. The identification of required expansion measures is frequently carried out in the form of mathematical optimization with long solution times. To address the computational challenges of Transmission Expansion Planning in integrated energy systems, this paper investigates the application of a specific mathematical optimization method, Benders Decomposition. To enable parallel optimization a certain model structure is necessary and therefore temporal coupling needs to be reduced. In addition, we investigate the computational time requirements. In all 14 test cases, the Benders Decomposition consistently outperforms direct solving of the Transmission Expansion Planning problem in terms of computation time. The time savings depend on size and complexity of the model. Reducing the temporal coupling enables parallelization, but it affects energy storage operation. This means that parallel computing and considerable time savings come at the cost of loosing the proof that the identified solution is a global optimum.