ABSTRACT
Lithium-ion batteries (LIBs) are widely used in portable devices, such as mobile phones and laptops. The validity of the data and the amount of data are regarded as the most important part in the field of machine learning. Additionally, a huge amount of molecule data certainly levels up the performance of the machine learning model. According to the schematic method, each of the additive molecules should undergo gas and solvent state optimization and charged system optimization in both states as well. Back-propagation network, a supervised learning process, was used to effectively achieve the best classification with different sort of weights and usage of activation and transfer functions by given input and output. A functional new electrolyte and additive that can be stable at a high working voltage instead of decomposing has been required to fulfill the development of the higher working voltage and higher energy density LIBs.
