HIL simulation of an optimized energy management incorporating an experimental biocatalytic methanation reactor
In this contribution, a model-based method for analyzing and optimizing energy systems comprising the electrical, thermal and chemical domain is presented. The method is a variant of the Hardware-in-the-Loop (HIL) simulation where virtual components are combined with real experimental components of the evaluated system. In order to integrate the real components with minimal instrumentation efforts, measured quantities are included as information flows, only, while the physical power flows are connected to local supply structures, like the electric grid or gas distribution system. This contribution incorporates a biocatalytic methanation reactor as an experimental component to convert hydrogen and carbon dioxide into methane. Compared to the well-known Sabatier process, this reactor operates at lower temperature levels and does not need pure carbon dioxide. This allows a dynamic operation and makes it more flexible regarding the carbon dioxide source whose availability is often critically discussed. The virtual energy components are represented by real-time capable models describing their physical behavior. In a test scenario, the electrical energy supply of residential quarters is investigated where photovoltaic data and a modeled fuel cell system are included beside the real experimental methanation process. For the dynamical management of energy and operating gases, electrical and chemical storage units are considered as virtual components, as well. The previous described energy system allows various strategies regarding the operation of the components, especially the storage units. Therefore an optimized energy management is reasonable, based on a designated criterion, e.g. minimal operating costs or maximum energy efficiency. In order to find the global optimum, the method of dynamic programming is used to determine the optimal control sequence for an assumed operation case, e.g. given by the photovoltaic yield of the considered day. Finally, the found solution is tested in real-time by the proposed HIL simulation.