SoC Estimation of Lithium Battery Based on Fully Connected Deep Neural Network

Zhenkai Qin; Mingfu Zhu; Yizhe Zhang; Shangxin Liu; Yawei Chen

doi:10.6911/WSRJ.202502_11(2).0002

Authors

Zhenkai Qin
Mingfu Zhu
Yizhe Zhang
Shangxin Liu
Yawei Chen

DOI:

https://doi.org/10.6911/WSRJ.202502_11(2).0002

Keywords:

State of Charge; Fully Connected Deep Neural Network; Lithium Iron Phosphate Battery; 4-Fold Cross Validation Method; TensorFlow; Keras.

Abstract

In order to improve the estimation accuracy of the State of Charge (SoC) of lithium batteries, in view of the nonlinear mapping characteristics of battery data and the inadequacy of shallow neural network mapping, lithium iron phosphate batteries were taken as the research object, TensorFlow and Keras as the support of the experimental platform, and a Fully Connected Deep Neural Network (FCDNN) was proposed to establish the SOC prediction model for lithium batteries. Experimental data were obtained from the Center for Advanced Life Cycle Engineering (CALCE) at the university of Maryland. With current, voltage and resistance as the main inputs of the model, and SOC as the output, the training set was used to train the FCDNN model, and the verification set was used for verification. The training set and the verification set were divided by the 4-fold cross validation method, and the performance of the model was verified while the FCDNN model was trained. Finally, the test set is used to test the resulting model. The results show that the mean absolute error (MAE) of SOC prediction using FCDNN model is 0.954%, the model estimation error is kept within 3%, and the estimated time for a single sample is 5ms. The validity of FCDNN was verified by comparing the prediction errors of FCDNN with different hidden layers, traditional Back Propagation Neural Network (BPNN) and Ampere-hour integral method.

Downloads

Download data is not yet available.

References

[1] Xu J, Cai X, Cai S, et al. High‐energy lithium‐ion batteries: recent progress and a promising future in applications[J]. Energy & Environmental Materials, 2023, 6(5): e12450.

[2] Urquizo J, Singh P. A review of health estimation methods for Lithium-ion batteries in Electric Vehicles and their relevance for Battery Energy Storage Systems[J]. Journal of Energy Storage, 2023, 73: 109194.

[3] Fang Y, Zhang Q, Zhang H, et al. State‐of‐charge estimation technique for lithium‐ion batteries by means of second‐order extended Kalman filter and equivalent circuit model: Great temperature robustness state‐of‐charge estimation[J]. IET Power Electronics, 2021, 14(8): 1515-1528.

[4] Hannan M A , Lipu M S H , Hussain A , et al. Neural Network Approach for Estimating State of Charge of Lithium-ion Battery Using Backtracking Search Algorithm[J]. IEEE Access, 2018, 6(2018):10069-10079.

[5] Zhang X, Hou J, Wang Z, et al. Study of SOC estimation by the ampere-hour integral method with capacity correction based on LSTM[J]. Batteries, 2022, 8(10): 170.

[6] Ren Z, Du C, Wu Z, et al. A comparative study of the influence of different open circuit voltage tests on model‐based state of charge estimation for lithium‐ion batteries[J]. International journal of energy research, 2021, 45(9): 13692-13711.

[7] Ephrem Chemali, Phillip J. Kollmeyer, Matthias Preindl, et al. State-of-charge estimation of Li-ion batteries using deep neural networks: A machine learning approach[J]. Journal of Power Sources, 2018,400.

[8] Kaniewski P. Extended Kalman filter with reduced computational demands for systems with non-linear measurement models[J]. Sensors, 2020, 20(6): 1584.

[9] Kaviani S, Sohn I. Application of complex systems topologies in artificial neural networks optimization: An overview[J]. Expert Systems with Applications, 2021, 180: 115073.

[10] Lipu M S H , Hannan M A , Hussain A , et al. State of Charge Estimation for Lithium-ion Battery Using Recurrent NARX Neural Network Model Based Lighting Search Algorithm[J]. IEEE Access, 2018, 6(2018):28150-28161.

[11] Cai B, Li M, Yang H, et al. State of Charge Estimation of Lithium-Ion Battery Based on Back Propagation Neural Network and AdaBoost Algorithm[J]. Energies, 2023, 16(23): 7824.

[12] Zhang Z, Chen S, Lu L, et al. High-Precision and Robust SOC Estimation of LiFePO4 Blade Batteries Based on the BPNN-EKF Algorithm[J]. Batteries, 2023, 9(6): 333.

[13] Tao S, Jiang B, Wei X, et al. A systematic and comparative study of distinct recurrent neural networks for lithium-ion battery state-of-charge estimation in electric vehicles[J]. Energies, 2023, 16(4): 2008.

[14] Yahia A, Tamer M, Shahin M, et al. Advances in Lithium-Ion Battery SOC Prediction: A Data-Driven GRU-RNN Framework[C]//2023 11th International Japan-Africa Conference on Electronics, Communications, and Computations (JAC-ECC). IEEE, 2023: 253-258.

[15] Nishanthy J, Sivakumar N. Deep Learning based EV Battery Management System[C]//2023 International Conference on Energy, Materials and Communication Engineering (ICEMCE). IEEE, 2023: 1-5.

[16] Pattanayak S. Introduction to deep-learning concepts and TensorFlow[M]//Pro Deep Learning with TensorFlow 2.0: A Mathematical Approach to Advanced Artificial Intelligence in Python. Berkeley, CA: Apress, 2023: 109-197.

[17] CALCE. Lithium-Ion Battery Experimental Data. Accessed: Jan. 5, 2017.[Online]. Available: http://www.calce.umd.edu/batteries/data.htm.

[18] Wang P, Fan E, Wang P. Comparative analysis of image classification algorithms based on traditional machine learning and deep learning[J]. Pattern recognition letters, 2021, 141: 61-67.

[19] Karypidis E, Mouslech S G, Skoulariki K, et al. Comparison Analysis of Traditional Machine Learning and Deep Learning Techniques for Data and Image Classification [J]. arXiv preprint arXiv:2204.05983, 2022.

[20] D.P. Kingma and J. Ba. Adam: a method for stochastic optimization, CoRR abs/1412.6980, http://arxiv.org/abs/1412.6980.

[21] Wieczorek J, Guerin C, McMahon T. K‐fold cross‐validation for complex sample surveys[J]. Stat, 2022, 11(1): e454.