Development of chemistry-specific battery energy storage system models using combined multiphysics and reduced order modeling

Abstract

Batteries are pivotal towards the decarbonization of energy systems as they address the intermittent nature of renewable energy technologies. The techno-economic feasibility of deploying batteries in microgrids is often analyzed using energy systems modeling tools. These often utilize the idealized battery model, which is simpler but neglects electrochemical phenomena. Reduced-order models, which are derived from continuum-scale physics-based models, have improved accuracy but are yet to be applied in the cost optimization of microgrids. In this work, we proposed a novel methodology for generating reduced-order models of lithium-ion lithium iron phosphate, nickel cobalt aluminum oxide, and nickel manganese cobalt chemistries for use in the techno-economic optimization of microgrids. We simulated previously reported multiphysics models of Li-ion batteries in COMSOL Multiphysics® and reduced them into equivalent circuit models. These were implemented in Island Systems LCOEmin Algorithm (ISLA), our in-house energy systems modeling and optimization tool. We then simulated and optimized renewable energy systems in ISLA to determine the discrepancy between the reduced-order equivalent circuit models versus the idealized battery model. The results revealed that the idealized battery model can miscalculate the SOC by >5 % SOC due to the assumption of constant voltages, while the optimum sizes differed by >5 % when there are sharp peaks in demand or if non-renewable sources are competitive. The idealized battery model is therefore not a valid approximation under these scenarios. This work demonstrated a novel multi-scale framework from the continuum-scale multiphysics modeling of batteries to macroscale energy systems optimization.