Topology Construction Based on Graph Theory for SOC Balancing in Dynamic Reconfigurable Battery System

The inherent heterogeneity of lithium-ion battery cells induces systemic performance degradation through the bucket effect. This paper addresses performance degradation in heterogeneous lithium-ion battery systems through a graphtheoretic control framework for dynamic reconfigurable networks (DRBNs). A hybrid physics-constrained model integrates secondorder RC electrochemistry with switched Kirchhoffian network dynamics. This paper proposes an operation control strategy for DRBN based on graph theory. An optimization algorithm for DRBN balance is designed, using SOC variance as the optimization variable and combining dynamic programming to derive the transfer cost, which significantly improves SOC balance and operating time. To address the time-consuming problem of dynamic programming, the genetic algorithm is used for optimization, effectively enhancing efficiency and achieving a good balance effect. It is determined that a moderate scale can meet practical needs.