Wang Hong, Wang Qunqun. Performance Optimization of Negative Electrode Material for Vanadium-based Nickel Metal Hydride Battery Based on Neural Network Algorithm[J]. IRON STEEL VANADIUM TITANIUM, 2020, 41(6): 60-65. doi: 10.7513/j.issn.1004-7638.2020.06.012
Citation:
Wang Hong, Wang Qunqun. Performance Optimization of Negative Electrode Material for Vanadium-based Nickel Metal Hydride Battery Based on Neural Network Algorithm[J]. IRON STEEL VANADIUM TITANIUM, 2020, 41(6): 60-65. doi: 10.7513/j.issn.1004-7638.2020.06.012
Wang Hong, Wang Qunqun. Performance Optimization of Negative Electrode Material for Vanadium-based Nickel Metal Hydride Battery Based on Neural Network Algorithm[J]. IRON STEEL VANADIUM TITANIUM, 2020, 41(6): 60-65. doi: 10.7513/j.issn.1004-7638.2020.06.012
Citation:
Wang Hong, Wang Qunqun. Performance Optimization of Negative Electrode Material for Vanadium-based Nickel Metal Hydride Battery Based on Neural Network Algorithm[J]. IRON STEEL VANADIUM TITANIUM, 2020, 41(6): 60-65. doi: 10.7513/j.issn.1004-7638.2020.06.012
The neural network model with 6×36×12×1 four-layer topological structure was used to optimize the performances of the negative electrode material of vanadium based hydrogen storage battery.The input layer parameters were titanium content,nickel content,aluminum content,chromium content,holding temperature and holding time.The output layer parameters were charge and discharge cycle stability.The results show that the model has strong prediction ability and high prediction accuracy,with an average relative training error of 4.8%and an average relative prediction error of 4.9%.Compared with the existing V3Ti Ni0.56material,the capacity decay rate of V3Ti Ni0.56Al0.3Cr0.4material optimized by neural network model after 30 charge-discharge cycles is reduced from 61%to 26%.