Abstract
Physics-based models (PBMs) such as the Doyle-Fuller-Newman (DFN) model offer high-fidelity predictions of lithium-ion battery behaviour but are often hindered by complex parameterisation requirements. This study assesses an Analytical Pulsing Protocol (APP) based workflow for DFN parameter identification using the recently proposed APP combined with minimal experimental input. APP facilitates the extraction of thermodynamic, kinetic, and transport characteristics from full-cell measurements and is used here for both parameter identification and as a structured framework for evaluating model fidelity. To the best of our knowledge, our contribution provides the first simulations of APP pulses with a physics-based model. Geometric data and scanning electron microscopy (SEM) imaging complement literature-based parameters, while fixing well-characterised values reduces optimisation dimensionality. Particle swarm optimisation (PSO) is used to optimise half-cell functions, while Bayesian optimisation with Gaussian process surrogates refines the most sensitive kinetic parameters identified through sensitivity analysis. The resulting parameter set demonstrates partial agreement with the APP pulse data and independent constant-current discharge tests, despite limitations in equilibrium assumptions and mass-transfer representation. Beyond its role in parameterisation, APP allows systematic comparison of experimental and simulated behaviour across the full state-of-charge (SOC) range, providing insight into thermodynamic, kinetic, and mass-transport limitations. This highlights APP as a potentially valuable and practical evaluation tool for assessing DFN parameter sets and diagnosing mismatches between model and experiment. The proposed methodology supports accelerated battery modelling and digitalisation efforts, reducing barriers to the adoption of physics-based models in both research and industrial applications.