Efficient management of the power and energy output of a high voltage battery pack requires a precise estimation of the State of Energy (SOE). For the accurate estimation of SOE, this work presents two data-driven methods as Deep Neural Network (DNN) and a regression model, i.e. Support Vector Regression (SVR). The effectiveness of the SOE estimation was compared, analysed, and studied through these models under similar conditions. For performance enhancement of estimation, a modified algorithm based on the grid search of optimized hyperparameters was proposed and evaluated in both the models. For training of the model at subsequent thermal ranges, two case studies were performed using US06, UDDS, LA92, and HWFET drive cycles and at four different temperature levels (−10, 0, 10, and 25 ℃), for each cycle. The results indicate that the DNN method has provided enhanced performance for State of Energy Estimation as compared to the regression models of ML, i.e. SVR. This work highlights the prevailing challenges in the industry and proposes the potential recommendation for Battery Management System (BMS) development and SOE estimation in next-generation EV applications.

Kumar, P., Rafat, Y., Cicconi, P., Saad Alam, M. (2022). State of Energy Estimation of Li-Ion Batteries Using Deep Neural Network and Support Vector Regression. In Computational Modelling in Industry 4.0: A Sustainable Resource Management Perspective (pp. 299-324) [10.1007/978-981-16-7723-6_16].

State of Energy Estimation of Li-Ion Batteries Using Deep Neural Network and Support Vector Regression

Paolo Cicconi;
2022-01-01

Abstract

Efficient management of the power and energy output of a high voltage battery pack requires a precise estimation of the State of Energy (SOE). For the accurate estimation of SOE, this work presents two data-driven methods as Deep Neural Network (DNN) and a regression model, i.e. Support Vector Regression (SVR). The effectiveness of the SOE estimation was compared, analysed, and studied through these models under similar conditions. For performance enhancement of estimation, a modified algorithm based on the grid search of optimized hyperparameters was proposed and evaluated in both the models. For training of the model at subsequent thermal ranges, two case studies were performed using US06, UDDS, LA92, and HWFET drive cycles and at four different temperature levels (−10, 0, 10, and 25 ℃), for each cycle. The results indicate that the DNN method has provided enhanced performance for State of Energy Estimation as compared to the regression models of ML, i.e. SVR. This work highlights the prevailing challenges in the industry and proposes the potential recommendation for Battery Management System (BMS) development and SOE estimation in next-generation EV applications.
2022
978-981-16-7722-9
Kumar, P., Rafat, Y., Cicconi, P., Saad Alam, M. (2022). State of Energy Estimation of Li-Ion Batteries Using Deep Neural Network and Support Vector Regression. In Computational Modelling in Industry 4.0: A Sustainable Resource Management Perspective (pp. 299-324) [10.1007/978-981-16-7723-6_16].
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11590/440847
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? ND
social impact