Hybrid modeling in thermochemical heat storage

Abstract

Storing heat energy has been under long-standing investigation for prospective applications, such as the capturing of excess heat from industrial processes or storing energy in concentrated solar power plants. In this contribution, thermochemical heat storage (THS) in the SrBr2-System is investigated, which provides a large energy capacity and next to perfect reversibility. However, the upscaling of these reactors is hampered by structural changes through mechanical and physical alteration of the powder bed, as well as changes in the microstructure, leading to degradation of the reactive powder bed during cycling. Even though modeling these effects can be done in principle, developing and parametrizing these models is difficult, due to the substantial structural changes happening on multiple scales in the reacting bed. In this contribution we try to tackle these challenges by way of hybrid modelling, i.e., the combination of physical and data-driven methods. To this end, experimental work is carried out on the macroscale (cm) to determine material properties, such as permeability and thermal conductivity. Further, imaging of the microstructure is done on the microscale (µm) using optical microscopy and µCT, which can be used to compute effective transport parameters. Then, the available data is used to build a hybrid model, combining data-driven techniques and physical simulations. This is done by constructing a surrogate model from the full physical simulation model and incorporating experimental data. This can be done either by training a neural network on a combination of simulated and experimental data, or by using techniques, such as model order reduction, where the non-linearities are handled by a neural network. In this contribution we will show the use of hybrid modeling techniques for the simulation of THS and give a short outlook on the application of our model to the optimization of the geometry of heat transfer fins in a reactor.