Estimation Of Lithium-ion Batteries Research Articles

Constant current charging time based fast state-of-health estimation for lithium-ion batteries

The state of health (SOH) estimation is critical for a battery management system's safe operation. Considering feature extraction, time-consuming, model/calculation complexity problems, a battery SOH estimation method based on constant current charging time (CCCT) is proposed in this paper. Unlike previous works, it is proved that CCCT can perfectly replace incremental capacity peak area. Since no filtering process is required in this method, the validity of the feature is maximally preserved. The random forest regression is combined to form accurate and fast SOH estimation. The proposed method is validated with the Oxford and CALCE datasets, collected from different batteries under different conditions. The average root-mean-square error of 8 cells for SOH estimation is 0.52%. Compared with the incremental capacity analysis (ICA)-based SOH estimation method, the prediction accuracy of the proposed method is improved by 41.6%, and fewer data are utilized. Besides, the time needed for the model training and prediction of the proposed method is less than 1 s. Additionally, the proposed method is proved to have good adaptability to different voltage ranges and charging/discharging conditions.

State-of-Charge Estimation of Lithium-ion Batteries Using LSTM Deep Learning Method

State-of-Charge Estimation of Lithium-ion Batteries Using LSTM Deep Learning Method

A Novel Trigger Mechanism for a Dual-Filter to Improve the State-of-Charge Estimation of Lithium-Ion Batteries

Abstract State-of-charge (SOC) estimation is essential in the energy management of electric vehicles. In the context of SOC estimation, a dual filter based on the equivalent circuit model represents an important research direction. The trigger for parameter filter in a dual filter has a significant influence on the algorithm, despite which it has been studied scarcely. The present paper, therefore, discusses the types and characteristics of triggers reported in the literature and proposes a novel trigger mechanism for improving the accuracy and robustness of SOC estimation. The proposed mechanism is based on an open-loop model, which determines whether to trigger the parameter filter based on the model voltage error. In the present work, particle filter (PF) is used as the state filter and Kalman filter (KF) as the parameter filter. This dual filter is used as a carrier to compare the proposed trigger with three other triggers and single filter algorithms, including PF and unscented Kalman filter (UKF). According to the results, under different dynamic cycles, initial SOC values, and temperatures, the root-mean-square error of the SOC estimated using the proposed algorithm is at least 34.07% lower than the value estimated using other approaches. In terms of computation time, the value is 4.67%. Therefore, the superiority of the proposed mechanism is demonstrated.

A Variable-length scale Parameter Dependent State of Charge Estimation of Lithium Ion Batteries by Kalman Filters

A Variable-length scale Parameter Dependent State of Charge Estimation of Lithium Ion Batteries by Kalman Filters

Applying Set Membership Strategy in State of Charge Estimation for Lithium-ion Battery

As state of charge is one important variable to monitor the later battery management system, and as traditional Kalman filter can be used to estimate the state of charge for Lithium-ion battery on basis of probability distribution on external noise. To relax this strict assumption on external noise, set membership strategy is proposed to achieve our goal in case of unknown but bounded noise. External noise with unknown but bounded is more realistic than white noise. After equivalent circuit model is used to describe the Lithium-ion battery charging and discharging properties, one state space equation is constructed to regard state of charge as its state variable. Based on state space model about state of charge, two kinds of set membership strategies are put forth to achieve the state estimation, which corresponds to state of charge estimation. Due to external noise is bounded, i.e. external noise is in a set, we construct interval and ellipsoid estimation for state estimation respectively in case of external noise is assumed in an interval or ellipsoid. Then midpoint of interval or center of the ellipsoid are chosen as the final value for state of charge estimation. Finally, one simulation example confirms our theoretical results.

Physics-Informed Recurrent Neural Network With Fractional-Order Gradients for State-of-Charge Estimation of Lithium-Ion Battery

As a typical machine learning algorithm, neural networks (NNs) has been designed and developed for battery management system (BMS) with artificial intelligence. State of charge (SOC) estimation of lithium-ion battery (LIB) is the basis of BMS so as to widely employ NNs, and recurrent neural network (RNN) is usually selected to describe the time-series characteristics of LIB. However, RNN is a data-driven statistic black box, which cannot reveal electrochemical principle and learn inner Knowledge of LIB. This paper introduces fractionalorder gradients for RNN to improve its backpropagation process, so that network updates weights instructed by the fractionalorder characteristics of LIB. Our work provides two backpropagation patterns with fractional-order gradient descent and momentum for RNN, respectively, both resulting in a physicsinformed RNN for SOC estimation of LIB. The proposed physicsinformed RNN can conduct training in which the gradients and the loss of network is informed by the physical fractional-order laws of LIB. Experimental results under operation conditions of federal urban driving schedule (FUDS) are presented with satisfying SOC estimation accuracy. Furtherly, physics-informed RNN proposed in this paper is not limited to SOC estimation, but also other state estimation or even fault prognosis for LIB.

Hfcm-Lstm: A Novel Hybrid Framework for State-of-Health Estimation of Lithium-Ion Battery

Hfcm-Lstm: A Novel Hybrid Framework for State-of-Health Estimation of Lithium-Ion Battery

Distributed State-of-Charge Estimation for Lithium-Ion Batteries with Random Sensor Failure Under Dynamic Event-Triggering Protocol

Distributed State-of-Charge Estimation for Lithium-Ion Batteries with Random Sensor Failure Under Dynamic Event-Triggering Protocol

An Error Covariance Correction-Adaptive Extended Kalman Filter Based on Piecewise Forgetting Factor Recursive Least Squares Method for the State-of-Charge Estimation of Lithium-Ion Batteries

An Error Covariance Correction-Adaptive Extended Kalman Filter Based on Piecewise Forgetting Factor Recursive Least Squares Method for the State-of-Charge Estimation of Lithium-Ion Batteries

Constant Current Charging Time Based Fast State-of-Health Estimation for Lithium-Ion Batteries

Constant Current Charging Time Based Fast State-of-Health Estimation for Lithium-Ion Batteries

A Novel Relevant Long Short-Term Memory-Squared Gain Extended Kalman Filter for the State of Charge Estimation of Lithium-Ion Batteries

A Novel Relevant Long Short-Term Memory-Squared Gain Extended Kalman Filter for the State of Charge Estimation of Lithium-Ion Batteries

A comprehensive review on the state of charge estimation for lithium‐ion battery based on neural network

Implementing carbon neutrality and emission peak policies requires a high-level electric vehicle field. Lithium-ion batteries have been considered an essential component of electric vehicle power batteries. Effective state of charge (SOC) estimation for lithium-ion batteries is a critical problem that needs to be addressed at present. With the feature extraction and fitting capability, the neural network can achieve accurate SOC estimation without considering the internal electrochemical state of the battery. This article overviews the definition of SOC and the relationship with battery aging state. Then, by examining recent literature on estimating the SOC of Lithium-ion batteries using neural network methods, the methods are classified into three categories: feed-forward neural network method, deep learning method, and hybrid method. The progress of neural network methods in SOC estimation applications is systematically reviewed, including principles, advantages, disadvantages, current status, and estimation errors. Possible recommendations for next-generation intelligent battery management systems and SOC estimation are also presented. This review's highlighted insights will inspire researchers in the battery field and point the way to developing electric vehicles.

A Novel State-of-Health Estimation for Lithium-Ion Battery via Unscented Kalman Filter and Improved Unscented Particle Filter

This paper aims to improve the rapidity and accuracy of the State of Health (SOH) assessment for lithium-ion battery. By integrating the Unscented Kalman Filter (UKF) and Improved Unscented Particle Filter (IUPF) algorithm, SOH can be effectively evaluated. The UKF algorithm is used to estimate the state of charge (SOC), the IUPF algorithm is employed to identify the ohmic internal resistance. The novelty of the proposed strategy relies on the 4-dimensional IUPF filter that is split into a 3-dimensional UKF filter and a 1-dimensional IUPF filter. Experimental results demonstrate that more accuracy and a faster rate of SOH estimation can be achieved via the UKFIUPF algorithm compared to the IUPF approach.

An adaptive fractional‐order extended Kalman filtering for state of charge estimation of high‐capacity lithium‐ion battery

The state of charge (SOC) estimation of lithium-ion battery is a crucial portion of the battery management system (BMS). The high-precision estimation is the foundation of BMS safety and efficiency. To that extent, a fractional-order algorithm with time-varying parameters model is proposed to ensure the accuracy of the SOC. Since the battery state changes slowly and is related to the state in the past, this study proposes a memory factor M containing the battery state in the past to estimate the SOC. Moreover, by comparing the experimental results of different orders, the most appropriate fractional order is determined. In order to eliminate the influence of noises introduced into historical data processing, an adaptive noise factor is added to the algorithm. The experimental results confirm that the maximum error of the adaptive fractional-order extended Kalman (AFEKF) estimation is less than 2%, which indicates that the estimation method provides a higher accuracy than the extended Kalman filter.

A fast sliding-mode-based estimation of state-of-charge for Lithium-ion batteries for electric vehicle applications

This paper present a 2- order fast embossing non-singular terminal sliding mode observer (2-FENSTSMO) for the state of charge (SoC) estimation of lithium-ion batteries. As these batteries have nonlinearities and uncertainties in their dynamic model, to estimate their SoC accurately a robust and nonlinear estimator is required. The sliding mode observers have chattering phenomena and prolong convergence time in their performance but the proposed observer tackles these problems and provides less chattering and convergence time. First, a second-order fast non-singular terminal manifold based on estimation error is defined, and then, according to the Lyapunov stability theory, in order to make a Lyapunov function descending, an appropriate control law is selected. An equivalent circuit model (ECM) for the battery including uncertainty is considered to design the proposed FNTSM observer. The stability of the proposed method is provided and the simulation and experimental results confirm the effectiveness of the proposed method.

A Method for Estimating State of Charge of Lithium-Ion Batteries Based on Deep Learning

State of charge (SOC) estimation of lithium-ion batteries is a problem of time series. In deep learning methods, both convolutional neural network (CNN) and recurrent neural network (RNN) can be used to solve such problems. In this paper, based on deep learning, a hybrid neural network model is proposed to estimate the SOC of lithium-ion batteries by taking the sequence of sampling points of voltage, current and temperature as input. The model is mainly composed of three modules, namely, convolutional module, ultra-lightweight subspace attention mechanism (ULSAM) module and the gated recurrent unit (GRU) module. Convolutional module and ULSAM module are responsible for extracting the feature information from the sequence of sampling points and outputting feature maps. GRU module is responsible for processing the sequences of the feature maps and outputting the value of SOC. The proposed model is tested on the public NASA Randomized Battery Usage dataset and Oxford Battery Degradation dataset. The experimental results show that the proposed model can obtain a relatively accurate SOC estimation at unknown aging state and complex operating conditions.

Accurate and reliable state of charge estimation of lithium ion batteries using time-delayed recurrent neural networks through the identification of overexcited neurons

Various neural network models have been adopted for lithium ion battery state of charge (SOC) estimation with good accuracy. However, problems for battery states estimation from neural networks were usually not reported, which is mainly due to the lack of effective solutions other than a trial and error training process. This paper firstly proposes time-delayed recurrent neural network for lithium ion battery modeling and SOC estimation. Both exceptional performances and unexpected overfitting or poor performances are reported with in-depth analysis of the root cause. With explicit formulation of the network, each hidden neuron’s output is examined. It is discovered that overexcited neurons could be the root cause for unexpected poor performances of the neural network. Without overexcited neurons, expectational SOC estimation accuracy is consistently obtained with estimation error being less than 1% for lithium ion magnesium phosphate (LiFeMgPO4) batteries considering a fair comparison in literature.

A joint moving horizon strategy for state-of-charge estimation of lithium-ion batteries under combined measurement uncertainty

Measurement uncertainty is a common problem for state of charge (SOC) estimation of lithium-ion batteries in real applications. In this paper, to mitigate the negative effect of unseen measurement uncertainty, a joint moving horizon estimation (joint-MHE) approach is proposed. First, the equivalent circuit model (ECM) is constructed for battery modeling. Then, on the basis of ECM, the augmented state space model is formulated, in which the bias current is treated as an additional state and the measurement noises are summarized in covariance matrices. Finally, by integrating joint-MHE strategy with augmented model, the SOC estimation under measurement uncertainty condition is implemented. The effectiveness of proposed method is conducted under three uncertainty issues, including current bias, combined current uncertainty, and combined current and voltage uncertainty, and compared to the conventional MHE and the joint-extended Kalman filter (EKF) thoroughly. The results demonstrate that the joint strategy is an effective way to suppress the uncertainties in measurements. Furthermore, although two joint methods both can reduce the negative effect of unseen measurement uncertainty, the joint-MHE could provide better convergence speed and SOC estimation accuracy, and is much less sensitive to different uncertainty sources. Under the combined measurement uncertainty, the RMSE by joint-EKF is 5.32% during the whole applied DST range, while that by joint-MHE is only 1.46%. It thus indicates that the joint-MHE is a potential promising approach to tackle the measurement uncertainty problem, which would greatly assist in improving the feasibility of ECM-based SOC estimation approach in commercial BMS.

Enhancing Stability and Robustness of State-of-Charge Estimation for Lithium-Ion Batteries by Using Improved Adaptive Kalman Filter Algorithms

The traditional Kalman filter algorithms have disadvantages of poor stability (the program cannot converge or crash), robustness (sensitive to the initial errors) and accuracy, partially resulted from the fact that noise covariance matrices in the algorithms need to be set artificially. To overcome the above problems, some adaptive Kalman filter (AKF) algorithms are studied, but the problems still remain unsolved. In this study, two improved AKF algorithms, the improved Sage-Husa and innovation-based adaptive estimation (IAE) algorithms, are proposed. Under the different operating conditions, the estimation accuracy, filter stability, and robustness of the two proposed algorithms are analyzed. Results show that the state of charge (SOC) Max error based on the improved Sage-Husa and the improved IAE is less than 3% and 1.5%, respectively, while the Max errors of the original algorithms is larger than 16% and 4% The two proposed algorithms have higher filter stability than the traditional algorithms. In addition, analyses of the robustness of the two proposed algorithms are carried out by changing the initial parameters, proving that neither are sensitive to the initial errors.

Deep learning approach towards accurate state of charge estimation for lithium-ion batteries using self-supervised transformer model

Accurate state of charge (SOC) estimation of lithium-ion (Li-ion) batteries is crucial in prolonging cell lifespan and ensuring its safe operation for electric vehicle applications. In this article, we propose the deep learning-based transformer model trained with self-supervised learning (SSL) for end-to-end SOC estimation without the requirements of feature engineering or adaptive filtering. We demonstrate that with the SSL framework, the proposed deep learning transformer model achieves the lowest root-mean-square-error (RMSE) of 0.90% and a mean-absolute-error (MAE) of 0.44% at constant ambient temperature, and RMSE of 1.19% and a MAE of 0.7% at varying ambient temperature. With SSL, the proposed model can be trained with as few as 5 epochs using only 20% of the total training data and still achieves less than 1.9% RMSE on the test data. Finally, we also demonstrate that the learning weights during the SSL training can be transferred to a new Li-ion cell with different chemistry and still achieve on-par performance compared to the models trained from scratch on the new cell.