Intelligent Deep Learning Estimators of a Lithium-Ion Battery State of Charge Design and MATLAB Implementation—A Case Study

Abstract

The main objective of this research paper was to develop two intelligent state estimators using shallow neural network (SNN) and NARX architectures from a large class of deep learning models. This research developed a new modelling design approach, namely, an improved hybrid adaptive neural fuzzy inference system (ANFIS) battery model, which is simple, accurate, practical, and well suited for real-time implementations in HEV/EV applications, with this being one of the main contributions of this research. On the basis of this model, we built four state of charge (SOC) estimators of high accuracy, assessed by a percentage error of less than 0.5% in a steady state compared to the 2% reported in the literature in the field. Moreover, these estimators excelled by their robustness to changes in the model parameters values and the initial “guess value” of SOC from 80–90% to 30–40%, performing in the harsh and aggressive realistic conditions of the real world, simulated by three famous driving cycle procedure tests, namely, two European standards, WLTP and NEDC, and an EPA American standard, FTP-75. Furthermore, a mean square error (MSE) of 7.97 × 10−11 for the SOC estimation of the NARX SNN SOC estimator and 5.43 × 10−6 for voltage prediction outperformed the traditional SOC estimators. Their effectiveness was proven by the performance comparison with a traditional extended Kalman filter (EKF) and adaptive nonlinear observer (ANOE) state estimators through extensive MATLAB simulations that reveal a slight superiority of the supervised learning algorithms by accuracy, online real-time implementation capability, in order to solve an extensive palette of HEV/EV applications.

Investigations of Different Approaches for Controlling the Speed of an Electric Motor with Nonlinear Dynamics Powered by a Li-ion Battery – Case Study

“…Then the data set is processed for denoising, outliers removal, and data sharpness that significantly affect the performance accuracy. Essentially, these intelligent regressive neural network structures solve nonlinear time series problems using dynamic neural networks, including feedback networks [27][28][29][30]. They can be applied in open-loop, closed-loop, and open/closed-loop multistep prediction [27,28].…”

“…The design and implementation in MATLAB Simulink follow the steps inspired by [27][28][29][30] and are summarized in this subsection as follows:…”

“…Also, its robustness is better when estimating the SOC of different chemistry batteries. To achieve higher accuracy of state estimation, various intelligent algorithms based on Machine Learning (ML) and Deep Learning (DL) Artificial Intelligence (AI) models are applied to the SOC estimation and terminal voltage prediction, as those developed in [7,12,13,17,[19][20][21][22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38] easily to be adapted to all types of batteries and chemistries. The neural networks (NNs) learning techniques have a wide range of applications and are suitable for all types of batteries chemistry.…”

“…The neural networks (NNs) learning techniques have a wide range of applications and are suitable for all types of batteries chemistry. Well, these learning techniques such as machine learning (ML) and deep learning (DL) models require large amounts of accurate training data [27][28][29][30]. The estimation accuracy and the convergence speed of the Li-ion battery SOC depend on the chosen training method, architecture structures, number of hidden layers and hidden neurons, learning rate, gradient value and on number of samples and epochs [29,30].…”

“…Well, these learning techniques such as machine learning (ML) and deep learning (DL) models require large amounts of accurate training data [27][28][29][30]. The estimation accuracy and the convergence speed of the Li-ion battery SOC depend on the chosen training method, architecture structures, number of hidden layers and hidden neurons, learning rate, gradient value and on number of samples and epochs [29,30]. The flow of this research paper is organized into four sections, as follows.…”