Table 1 Comparative analysis of literature works for rul prediction of the lib.
Ref. | Model | Parameters used | Strength | Weakness |
|---|---|---|---|---|
UPF | Capacity | Low computational complexity with more accurate RUL prediction outcomes | Noise inclusion in the data was not considered which could increase the characteristics of the data | |
UKF-PF | Capacity | The RUL prediction outcome demonstrated high applicability of the proposed model | The parameters from the operating profile were not used to develop the data framework | |
SVR | Capacity and impedance | The generalization ability of the model was high | The selection of model parameters required trial and error which needed human expertise and time | |
RVM | Mean voltage falloff | The RUL prediction outcomes were accurate. The reconstitution of the capacity degradation curve is accurate | The Mean voltage falloff can be obtained only with stable operating conditions and not with a transient operation state such as EV | |
BPNN | Capacity, voltage, current, temperature | Different LIB parameters were included to train the BPNN model | The selection of the BPNN model parameters was time-consuming | |
FNN-PF | Capacity | Model parameter selection was conducted with an optimized technique | The model training was not comprehensive based on different training ratios | |
FNN-JFO | Capacity, voltage, current, temperature | Selection of features using a mathematical sampling method | -The model was not validated with other LIB datasets |
