LG 18650HG2 Li-ion Battery Data and Example Deep Neural Network xEV SOC Estimator Script
Description
The included tests were performed at McMaster University in Hamilton, Ontario, Canada by Dr. Phillip Kollmeyer (phillip.kollmeyer@gmail.com). If this data is utilized for any purpose, it should be appropriately referenced. A brand new 3Ah LG HG2 cell was tested in an 8 cu.ft. thermal chamber with a 75amp, 5 volt Digatron Firing Circuits Universal Battery Tester channel with a voltage and current accuracy of 0.1% of full scale. these data are used in the design process of an SOC estimator using a deep feedforward neural network (FNN) approach. The data also includes a description of data acquisition, data preparation, development of an FNN example script. The test data, or similar data, has been used for some publications, including: C. Vidal, P. Kollmeyer, M. Naguib, P. Malysz, O. Gross, and A. Emadi, “Robust xEV Battery State-of-Charge Estimator Design using Deep Neural Networks,” in Proc WCX SAE World Congress Experience, Detroit, MI, Apr 2020 C. Vidal, P. Kollmeyer, E. Chemali and A. Emadi, "Li-ion Battery State of Charge Estimation Using Long Short-Term Memory Recurrent Neural Network with Transfer Learning," 2019 IEEE Transportation Electrification Conference and Expo (ITEC), Detroit, MI, USA, 2019, pp. 1-6.