Publication: State of Charge Estimation for Lithium-Ion Battery Using Recurrent NARX Neural Network Model Based Lighting Search Algorithm
No Thumbnail Available
Date
2018
Authors
Lipu M.S.H.
Hannan M.A.
Hussain A.
Saad M.H.M.
Ayob A.
Blaabjerg F.
Journal Title
Journal ISSN
Volume Title
Publisher
Institute of Electrical and Electronics Engineers Inc.
Abstract
State of charge (SOC) is one of the crucial parameters in a lithium-ion battery. The accurate estimation of SOC guarantees the safe and efficient operation of a specific application. However, SOC estimation with high accuracy is a serious concern to the automobile engineer due to the battery nonlinear characteristics and complex electrochemical reactions. This paper presents an improved nonlinear autoregressive with exogenous input (NARX)-based neural network (NARXNN) algorithm for an accurate and robust SOC estimation of lithium-ion battery which is effective and computationally rich for controlling dynamic system and predicting time series. However, the accuracy of recurrent NARXNN depends on the amount of input order, output order, and hidden layer neurons. The unique contribution of the improved recurrent NARXNN-based SOC estimation is developed using lighting search algorithm (LSA) for finding the best value of input delays, feedback delays, and hidden layer neurons. The contributions are summarized as: 1) the computational capability of NARXNN model which does not require battery model and parameters rather only needs current, voltage, and temperature sensors; 2) the effectiveness of LSA which is verified with particle swarm optimization; 3) the adaptability, efficiency, and robustness of the model which are evaluated using FUDS and US06 drive cycles at varying temperatures conditions; and 4) the performance of the proposed model which is compared with back propagation neural network and radial basis function neural network optimized by LSA using different error statistical terms and computational time. Furthermore, a comparative analysis of SOC estimation in proposed method and existing techniques is presented for validation of NARXNN performance. The results prove that the proposed NARXNN model achieves higher accuracy with less computational time than other existing SOC algorithms under different temperature conditions and electric vehicle drive cycles. � 2013 IEEE.
Description
Backpropagation; Charging (batteries); Ions; Learning algorithms; Lighting; Lithium-ion batteries; Particle swarm optimization (PSO); Radial basis function networks; Back-propagation neural networks; Electrochemical reactions; NARX neural network; Non-linear autoregressive with exogenous; Radial basis function neural networks; Search Algorithms; State of charge; State-of-charge estimation; Battery management systems