A Multi-Stage Model Order Reduction Framework for Efficient Simulations of Parametrized Lithium-Ion Battery Cells

In this thesis, we present a new multi-stage model order reduction framework for efficient simulations of a parameterized pseudo three-dimensional battery model. We first introduce the continuum battery model for a general intercalation battery cell on the basis of non-equilibrium thermodynamics. In order to efficiently evaluate the resulting parameterized non-linear system of partial differential equations, we apply the hierarchical model reduction method with two hierarchies. We construct reduced spaces in the radial (pseudo) and the macroscopic dimension. First, we reduce the radial dimension using the standard hierarchical model reduction method, a tensor-product approach that tackles the dependence of the battery problem on the various spatial directions in a different manner. We obtain a system of equations which depend only on the macroscopic scale and are independent of the radial direction. Afterwards, we employ the reduced basis method. The reduced basis method is a model order reduction technique on the basis of an incremental hierarchical approximate proper orthogonal decomposition approach and empirical operator interpolation. The reduction framework is particularly well suited to investigate and quantify parameter-depended degradation effects of battery cells. Several numerical experiments are given to demonstrate the scope and the efficiency of the multi-stage model order reduction framework.