Table 2 SOH estimation performance with various methods
From: Multi-modal framework for battery state of health evaluation using open-source electric vehicle data
Model | MAPE (%) | RMSE (%) | Max error (%) | Training time (s) | |||
|---|---|---|---|---|---|---|---|
Average | Std. | Average | Std. | Average | Std. | ||
3 modalities with ResNet | 2.83 | 0.093 | 3.26 | 0.090 | 17.15 | 2.186 | 580 |
SVR model | 3.34 | 0.114 | 3.74 | 0.145 | 16.557 | 2.000 | 0.5 |
RFR model | 3.10 | 0.264 | 3.54 | 0.193 | 16.00 | 1.279 | 10 |
GPR model | 3.11 | 0.129 | 3.47 | 0.139 | 16.33 | 1.791 | 28 |
2 modalities with ResNet | 3.81 | 0.170 | 4.14 | 0.215 | 30.81 | 4.459 | 93 |
1 modality with ResNet | 4.66 | 0.498 | 4.88 | 0.491 | 28.83 | 0.942 | 31 |
3 modalities with CNN | 3.76 | 0.388 | 4.13 | 0.273 | 26.41 | 4.120 | 280 |
3 modalities with RNN | 3.09 | 0.197 | 3.89 | 0.911 | 34.33 | 27.78 | 770 |
3 modalities with LSTM | 3.29 | 0.395 | 4.04 | 0.972 | 30.34 | 24.36 | 977 |
3 modalities with FNN | 3.21 | 0.204 | 3.80 | 0.536 | 27.51 | 19.98 | 900 |
