Open AccessSimulation and estimation of lithium-sulfur battery charge state using fuzzy neural network
Author(s) -
Д. В. Колосницын,
Alexandra A. Savvina,
L.A. Khramtsova,
Elena Kuzmina,
Е. В. Карасева,
В. С. Колосницын
Publication year - 2021
Publication title -
èlektrohimičeskaâ ènergetika
Language(s) - English
Resource type - Journals
eISSN - 1680-9505
pISSN - 1608-4039
DOI - 10.18500/1608-4039-2021-21-2-96-107
Subject(s) - state of charge , battery (electricity) , adaptive neuro fuzzy inference system , artificial neural network , lithium (medication) , voltage , lithium battery , computer science , control theory (sociology) , fuzzy control system , simulation , fuzzy logic , engineering , chemistry , control (management) , artificial intelligence , electrical engineering , power (physics) , thermodynamics , ion , physics , medicine , organic chemistry , endocrinology , ionic bonding
The possibility of determining the charge state of lithium-sulfur batteries using the ANFIS model was estimated. Easily measurable in practice physical quantities were used as input parameters of the model. They are the battery voltage, the rate of its change and the number of previous cycles. The analysis of ANFIS models with various parameters (the number and type of membership functions) was carried out. It was shown that ANFIS is a model that makes it possible to estimate the charge state of a lithium-sulfur battery with the accuracy of more than 95%. The proposed type of models can be used in control and monitoring systems, together with digital aggregated twins, for additional training of models based on real data and increasing the accuracy of estimating the charge state of lithium-sulfur batteries.