About the Possibility of Simulation the Discharge Characteristics of Lithium–Sulfur Batteries Using Fuzzi Neural Networks

Abstract

In this work we considered the possibility of simulation of changes in the characteristics of lithium-sulfur batteries during cycling using an Adaptive Neuro-Fuzzy Inference System, ANFIS. The discharge profiles and the curve of decrease of discharge capacity of lithium-sulfur cells during cycling have been simulated. Neural network training was performed on every 5th cycle from the first to 95 cycles. It was shown that the simulated discharge profiles of lithium-sulfur cells are in good agreement with the experimental discharge profiles. The forecast depth of the decrease in the discharge capacity of lithium-sulfur cells during cycling with an accuracy of $$ \gg $$ 5% was 45 cycles. Simulation time of one discharge profile lasts 4.5 seconds, which makes it possible to use this approach in the development of control and monitoring systems for batteries (Battery Management System, BMS).