Optimal Battery Configurations and Control Strategies under Dynamic Loads

Lithium-ion batteries gain widespread application in automotive, energy, and electronics industries. Management of large-scale lithium-ion battery systems become imperative and challenging especially under the presence of battery variations. Recently battery array reconfiguration has been proposed to improve energy utility with respect to a specific implementation architecture. However the feasibility issue of meeting heavy power load requirement under battery variations and shortages is rarely addressed. In this thesis, building upon a battery discharge trace model we study a general principle of battery array configuration for a given power load requirement without assuming a priori an implementation architecture. Observing that optimal configurations for heavy and light loads are completely different, we devise two configuration modes: one to maximize power delivery and the other to maximize energy delivery. We further proposed a strategy switching between the two modes depending on the power requirement from a dynamic load. We also propose a control strategy for bypassed/enabled architecture, a simplest but most practical reconfigurable architecture. Experiments on power load patterns from an electric vehicle road test and from daily electricity usage of a business unit show the unique benefits and effectiveness of our proposed methods.