Back to Search
Start Over
State of charge estimation for electric vehicle power battery using advanced machine learning algorithm under diversified drive cycles.
- Source :
-
Energy . Nov2018, Vol. 162, p871-882. 12p. - Publication Year :
- 2018
-
Abstract
- Abstract State of charge estimation is one of the most critical factors to solve the key issues of monitoring and safety concerns of an electric vehicle power battery. In this paper, a state of charge estimation approach using subtractive clustering based neuro-fuzzy system is presented and evaluated by the simulation experiments using advanced vehicle simulator in comparison with back propagation neural network and Elman neural networks. Input parameters to model the state of charge estimation approach using subtractive clustering based neuro-fuzzy system are current, temperature, actual power loss, available and requested power, cooling air temperature and battery thermal factor. Data collected from 10 different drive cycles are utilized for the training and testing stages of the state of charge estimation model. Experimental results illustrated that the proposed model exhibits sufficient accuracy and outperforms both neural network and Elman neural network based models. Thus, the proposed model under different drive cycles show remarkable advancement in state of charge estimation with high potential to overcome the drawbacks in traditional methods and therefore provides an alternative approach in state of charge estimation. In addition, a sensitivity analysis is also performed to determine the importance of each input parameter on output i.e. state of charge. Highlights • An SC-ANFIS model to estimate the battery state of charge is developed. • The proposed model exhibits sufficient accuracy and outperforms other advanced models. • The model requires a simple training procedure and is evaluated using 10 diversified unseen driving cycles. • Different statistical metrics are adopted for the model state of charge estimation performance evaluation. • Sensitivity analysis is performed to determine the influence and importance of input variables on battery state of charge. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 03605442
- Volume :
- 162
- Database :
- Academic Search Index
- Journal :
- Energy
- Publication Type :
- Academic Journal
- Accession number :
- 132014797
- Full Text :
- https://doi.org/10.1016/j.energy.2018.08.071