Search papers, labs, and topics across Lattice.
This study systematically compares four neural network architectures鈥擬LP, ResNet, U-Net, and FNO鈥攆ormulated as autoregressive state-transition operators to predict internal states of lithium-ion batteries based on the computationally intensive Doyle-Fuller-Newman (DFN) model. The results reveal that the U-Net architecture significantly outperforms the others, achieving a mean final-step normalized root mean square error (nRMSE) of just 3% across all internal state variables while offering a 5.38x speed-up over traditional numerical solvers. This research underscores the importance of spatial inductive bias in enhancing the performance of machine learning surrogates for real-time battery state predictions.
U-Net achieves a remarkable 5.38x speed-up over traditional solvers while maintaining 3% prediction accuracy for battery internal states, highlighting the power of spatial inductive bias.
The Doyle-Fuller-Newman (DFN) model resolves internal electrochemical states in lithium-ion batteries with high fidelity. However, the numerical solution of its governing equations is computationally prohibitive for real-time deployment, limiting scalability from individual cells to pack and fleet-scale applications. While machine learning surrogates can substantially reduce inference latency through GPU acceleration, most existing approaches learn solution approximations tied to specific operating conditions rather than learning generalizable state-evolution dynamics. This work presents a systematic comparison of four neural network architectures (MLP, ResNet, U-Net, FNO) formulated as autoregressive state-transition operators that predict full DFN internal states across a wide range of operating conditions. To ensure a controlled architectural comparison, all models are trained under a unified framework using multi-step unrolling and current-conditioning, isolating the impact of spatial inductive bias. Results demonstrate that the U-Net's multi-scale feature hierarchy achieves a mean final-step nRMSE of 3% averaged across all internal state variables after 300-step autoregressive rollouts, while providing a 5.38x speed-up over the numerical solver. These findings highlight spatial inductive bias as a critical determinant of surrogate performance, advancing the development of surrogates for internal state observability for next-generation battery management systems and digital twins.
