Publication: An optimization approach for the state of charge (SOC) estimation battery based on artificial neural network
Loading...
Date
2023-07
Authors
Ng, Zi Jun
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
For every battery there is an important factor which contributes to the battery life span which is known as the State of Charge (SOC) it is used to determine the how much energy is left in the battery and usually express in percentage. It is an important factor to optimize the battery performance and lifespan. researchers are required to know how much charge will be remaining in a battery and how quickly it is being used up. Accurately measuring and managing the battery’s SOC is essential for ensuring that the electric vehicle can travel to certain amount of distance for a single full charged. Therefore, the purpose of this research is to gather data from various battery types to study the performance on each type of them. Three batteries were used in this research which are the lithium-ion, nickel-metal hydride, and lithium polymer. To calculate the SOC, some parameters such as the voltage, capacity and temperature are considered. In this research the feedforward neural network (FFNN) was used to estimate the SOC for each of the battery type. The train model was built in MATLAB and after successfully
constructing the FFNN using the Levenberg-Marquardt algorithm for estimating the SOC, analysis of variance (ANOVA) is used to find out the consistency of the accuracy performance. The performance of the FFNN produces a root mean squared error (RMSE) of 0.0124, mean squared error (MSE) of micro 0.21 and accuracy percentage of 99.23% in average.