2024/10/17 by Le Xu, Xu, Le, Simone Fasolato +3
Chemistry · Engineering · Mathematics · Medicine · #Advanced Battery Technologies Research #Applied mathematics #Battery (electricity) #Chemical engineering #Chemistry #Composite material #Core (optical fiber) #Engineering #FOS: Electrical engineering #Geology #Lithium (medication) #Lithium iron phosphate #Materials science #Mathematics #Medicine #Observability #Organic chemistry #Particle (ecology) #Phosphate #Physics #Systems and Control (eess.SY) #Thermodynamics #Volume (thermodynamics) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2410.14032
openalex publication_date 2024/10/17 · openalex created_date 2024/11/02 · openalex updated_date 2026/07/28
The increasing adoption of Lithium Iron Phosphate (LFP) batteries in Electric Vehicles is driven by their affordability, abundant material supply, and safety advantages. However, challenges arise in controlling/estimating unmeasurable LFP states such as state of charge (SOC), due to its flat open circuit voltage, hysteresis, and path dependence dynamics during intercalation and de-intercalation processes. The Core Shell Average Enhanced Single Particle Model (CSa-ESPM) effectively captures the electrochemical dynamics and phase transition behavior of LFP batteries by means of Partial Differential-Algebraic Equations (PDAEs). These governing PDAEs, including a moving boundary Ordinary Differential Equation (ODE), require a fine-grained spatial grid for accurate and stable solutions when employing the Finite Difference Method (FDM). This, in turn, leads to a computationally expensive system intractable for the design of real-time battery management system algorithms. In this study, we demonstrate that the Finite Volume Method (FVM) effectively discretizes the CSa-ESPM and provides accurate solutions with fewer than 4 control volumes while ensuring mass conservation across multi ple operational cycles. The resulting control-oriented reduced order FVM-based CSa-ESPM is experimentally validated using various C-rate load profiles and its observability is assessed through nonlinear observability analysis. Our results reveal that different current inputs and discrete equation numbers influence model observability, with non-observable regions identified where solid-phase concentration gradients are negligible.