Determining optimal discharge strategy for rechargeable lithium-ion batteries using multiphysics simulation

Abstract

Lithium-ion rechargeable batteries have enabled the proliferation of mobile electronics based on its high energy density, negligible memory effect and relatively high cycling capability. Unfortunately, while small scale deployment in consumer-level electronics has been successful, larger scale deployment for Electric Vehicles (EVs) or Hybrid Electric Vehicles (HEVs) have been handicapped by the uncertainty of its long-term reliability, power density and safety. Physical long-term testing requires months of waiting for the charge and discharge of the cells, and while the electrochemistry models for the cells are well-documented, there is a lack of modeling technique that bridges to the application level for electrical engineering designs. This thesis addresses the modeling issue by presenting a novel method to use the detailed electrochemistry-based model in real-world scenarios. The goal is to allow the simulations to be sequentially run in different states based on the change in the physical parameters of the cells, rather than switching between states using fixed time intervals. The method makes use of the highly efficient partial differential equation solver in COMSOL to simulate digital, one-way switching based on specific physical parameters. The other issue this thesis addresses is the reduction of degradation in the long-term cycling of lithium-ion batteries in larger-scale multi-cell applications. Using the detailed pseudo-2D electrochemistry models, different discharge currents were simulated in a popular cell (Sony 18650) and the optimal current, and how this optimal current decreases as the cell cycles, is presented. A 2-stage discharge method, used in conjunction with the penalty based switching algorithm, was developed based on the modeling results of the low current degradation which increases the per cycle discharge capacity by 5-10% while reducing the degradation on the anode of the cell by approximately four times when compared to non-optimized discharge methods. Overall, this research makes advances in the fields of sequential computer simulation of physical systems and electrical engineering design for lithium ion battery systems in larger-scale applications. The modeling method contributes to research opportunities in many fields, and the determination of the optimal discharge current further enables the implementation of lithium-ion batteries as the power source for EVs and HEVs.