A lightweight deep learning framework for accurate and efficient battery State-of-Health (SoH) monitoring. PACE combines temporal convolutional networks with physics-informed features from equivalent circuit models and chunked attention mechanisms to achieve superior performance while maintaining computational efficiency.
This work was developed at Singapore Institute of Technology. The top three main contributors are:
Sara Sameer (sara.sameer@singaporetech.edu.sg)
Zhang Wei (wei.zhang@singaporetech.edu.sg) [Corresponding author]
Kannan Dhivya Dharshini(dhivyadharshini.kannan@singaporetech.edu.sg)
Data for PACE can be obtained from the Google Drive.
We develop three battery-specific modules, including dilated temporal blocks for efficient temporal encoding, chunked attention blocks for context modeling, and a dual-head output block for fusing short- and long-term battery degradation patterns. Together, the modules enable Pace to predict battery health accurately and efficiently in various battery usage conditions
| Model | #Params ↓ (×10³) | FLOPs ↓ (×10⁶) | RMSE ↓ (1-Cycle) | MAE ↓ (1-Cycle) | η ↑ (1-Cycle) | RMSE ↓ (30-Cycle) | MAE ↓ (30-Cycle) | η ↑ (30-Cycle) | RMSE ↓ (50-Cycle) | MAE ↓ (50-Cycle) | η ↑ (50-Cycle) |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Transformer | 2559.5 | 133.2 | 0.014 | 0.009 | 27.9 | 0.016 | 0.010 | 24.4 | 0.017 | 0.009 | 23.0 |
| TCN | 47.0 | 1.7 | 0.036 | 0.022 | 580.2 | 0.053 | 0.039 | 401.8 | 0.057 | 0.042 | 373.6 |
| BCA | 173.5 | 5.0 | 0.034 | 0.018 | 169.5 | 0.037 | 0.020 | 155.8 | 0.038 | 0.020 | 150.5 |
| Pace (ours) | 70.9 | 5.1 | 0.023 | 0.010 | 613.4 | 0.033 | 0.014 | 427.5 | 0.035 | 0.015 | 403.1 |
Install Python 3.6, PyTorch 1.9.0.
python main.py --mode train --epochs 100 --batch_size 32 --lr 1e-3 --input_window 100 --output_window 30 --num_channels 32 64 64 --kernel_size 3 --num_runs 3 --model_dir models --wandb_project battery_soh --use_wandb --scale_data --attention_type multi --chunk_size 16
Accepted at ACM Symposium On Applied Computing (SAC) 2026