Search

In electric vehicles, a battery management system highly relies on the measured current, voltage, and temperature to accurately estimate state of charge (SOC) and state of health. Thus, the normal operation of current, voltage, and temperature sensors is of great importance to protect batteries from running outside their safe operating area. In this paper, a simple and effective model-based sensor fault diagnosis scheme is developed to detect and isolate the fault of a current or voltage sensor for a series-connected lithium-ion battery pack. The difference between the true SOC and estimated SOC of each cell in the pack is defined as a residual to determine the occurrence of the fault. The true SOC is calculated by the coulomb counting method and the estimated SOC is obtained by the recursive least squares and unscented Kalman filter joint estimation method. In addition, the difference between the capacity used in SOC estimation and the estimated capacity based on the ratio of the accumulated charge to the SOC difference at two nonadjacent sampling times can also be defined as a residual for fault diagnosis. The temperature sensor which is assumed to be fault-free is used to distinguish the fault of a current or voltage sensor from the fault of a battery cell. Then, the faulty current or voltage sensor can be isolated by comparing the residual and the predefined threshold of each cell in the pack. The experimental and simulation results validate the effectiveness of the proposed sensor fault diagnosis scheme. [ABSTRACT FROM AUTHOR]

This paper presents a robust sliding-mode observer (RSMO) for state-of-charge (SOC) estimation of a lithium-polymer battery (LiPB) in electric vehicles (EVs). A radial basis function (RBF) neural network (NN) is employed to adaptively learn an upper bound of system uncertainty. The switching gain of the RSMO is adjusted based on the learned upper bound to achieve asymptotic error convergence of the SOC estimation. A battery equivalent circuit model (BECM) is constructed for battery modeling, and its BECM is identified in real time by using a forgetting-factor recursive least squares (FFRLS) algorithm. The experiments under the discharge current profiles based on EV driving cycles are conducted on the LiPB to validate the effectiveness and accuracy of the proposed framework for the SOC estimation. [ABSTRACT FROM AUTHOR]

• A deep learning approach is proposed to estimate the SOC of LiFePO 4 batteries. • Accurate SOC estimation can be ensured even in the case of voltage plateaus. • A closed-loop framework is developed for improving robustness. • The proposed method can quickly adapt to new scenarios via transfer learning. State of charge (SOC) estimation constitutes a critical task of battery management systems. Conventional SOC estimation methods designed for dynamic profiles have difficulties in estimating SOC for LiFePO 4 batteries due to their flat open circuit voltage characteristics in the middle range of SOC. In this study, a deep neural network (DNN) based method is proposed to estimate SOC with only 10-min charging voltage and current data as the input. This method enables fast and accurate SOC estimation with an error of less than 2.03% over the entire battery SOC range. Thus, it can be used to calibrate the SOC estimation for the Ampere-hour counting method. We also demonstrate that by incorporating the DNN into a Kalman filter, the robustness of SOC estimation against random noises and error spikes can be improved. In the case of significant disturbances, the method still maintains a root mean square error of 0.385%. Moreover, the trained DNN can quickly adapt to various scenarios, including different ageing states and battery types charged at different rates, thanks to the transfer learning nature. Compared with developing a new DNN, transfer learning can provide more accurate estimation results at less training costs. By only fine-tuning one layer of the pre-trained DNN, the root mean square error can be less than 3.146% and 2.315% for aged batteries and different battery types, respectively. When more layers are fine-tuned, superior performance can be achieved. [ABSTRACT FROM AUTHOR]

Lithium-ion batteries decay every time as it is used. Aging-induced degradation is unlikely to be eliminated. The aging mechanisms of lithium-ion batteries are manifold and complicated which are strongly linked to many interactive factors, such as battery types, electrochemical reaction stages, and operating conditions. In this paper, we systematically summarize mechanisms and diagnosis of lithium-ion battery aging. Regarding the aging mechanism, effects of different internal side reactions on lithium-ion battery degradation are discussed based on the anode, cathode, and other battery structures. The influence of different external factors on the aging mechanism is explained, in which temperature can exert the greatest impact compared to other external factors. As for aging diagnosis, three widely-used methods are discussed: disassembly-based post-mortem analysis, curve-based analysis, and model-based analysis. Generally, the post-mortem analysis is employed for cross-validation while the curve-based analysis and the model-based analysis provide quantitative analysis. The challenges in the use of quantitative diagnosis and on-board diagnosis on battery aging are also discussed, based on which insights are provided for developing online battery aging diagnosis and battery health management in the next generation of intelligent battery management systems (BMSs). • Basic aging reactions inside battery during storage and cycling were described. • Detailed classification and comparison of aging diagnosis methods were presented. • Progress and challenges of aging diagnosis in quantitative analysis and on-board applications were provided. • Evolution of dominant aging mechanism under different external factors was discussed. [ABSTRACT FROM AUTHOR]

Fractional order models have been successfully applied to estimate states and diagnose faults for lithium ion batteries. However, their order has not been identified online, which restricts their applications in battery management systems due to the intuitive nonlinearity of fractional order identification. In this study, a novel online method is proposed to identify the parameters and order of a fractional order model for lithium ion batteries using least squares and a gradient-based method, respectively. This online method is validated against both simulation and experimental results. Compared with the fixed-order method under different operation conditions, the proposed method has achieved better model accuracy and robustness of identified model parameters. Furthermore, a hardware-in-the-loop test is also used to verify the efficacy of the proposed method. Based on the analysis of the online identification results, the limitations of existing fractional order models are also pointed out, and the directions to further improve the existing models are discussed. • The order and parameters of a fractional order battery model are identified online. • The proposed method is validated by experiments and hardware-in-the-loop tests. • The proposed method is more accurate and robust than the fixed order method. [ABSTRACT FROM AUTHOR]

Automotive electrification is a main source of demand for lithium ion batteries. Performances of battery charging directly affect consumers' recognition and acceptability of electric vehicles. Study on optimized charging methods is vital for future development of a smarter battery management system and an intelligent electric vehicle. This paper starts from introducing the working principles and existing problems of simple charging methods and then elaborating various optimized charging methods along with their characteristics and applications. It demonstrates that the optimized charging methods can reduce charging time, improve charging performance and extend battery life cycle comparing with conventional charging methods. At the end, this paper also provides a four-step pathway towards the design of an optimal charging method of Li-ion batteries: determine optimization objectives, establish optimization scheme, develop matching design and implement and promote the optimal charging method. • Optimized charging methods are reviewed. • A new viewpoint on optimized charging methods is presented. • A four-step pathway towards the design of an optimal charging method is provided. • The prospects and development direction of charging methods are expounded. [ABSTRACT FROM AUTHOR]

The control system of pseudo-OCV construction. [Display omitted] • Equation between SOC error with voltage error and OCV slope is established. • Novel pseudo open circuit voltage is proposed to estimate SOC of LiFePO 4 batteries. • Accurate SOC estimation can achieved even during the plateau period. • Inter-partition correction is proposed to improve the robustness and convergency. LiFePO 4 batteries are widely used in electric vehicles and energy storage due to their high safety. However, their flat voltage characteristics in the middle state of charge (SOC) range make it difficult to correct SOC estimation errors with a feedback mechanism. This study proposes a pseudo-open circuit voltage (OCV) modeling method to improve the performance of a closed-loop feedback correction. First, the relationship between the derivative of OCV and SOC or the slope of OCV, SOC range, and estimation error is analyzed to select the SOC interval for OCV curve construction. Second, a pseudo-OCV curve is established. An optimization algorithm is developed to identify the parameters of the OCV model based on the selected SOC interval and OCV slope, which establishes a pseudo-OCV curve. Third, the pseudo-OCV is compared with the real OCV to adjust the selected interval and further determine the optimal construction interval, then the curves of the pseudo-OCV and real OCV are stitched together to obtain the optimal OCV curve for global SOC estimation. Finally, SOC estimation is carried out using the constructed full pseudo-OCV and compared with the reference SOC obtained from experiments. The results show that the SOC estimation error at the full SOC range is less than 3 %. [ABSTRACT FROM AUTHOR]

• Solid-phase diffusion equation based surface SOC is proposed. • New enhanced ECM to describe low SOC behavior more precisely was proposed. • Battery test platform was built to conduct battery tests for model validation. • Proposed model's voltage RMSE at low SOC has been reduced to 8 mV. Battery model is the basis of battery efficient and safe management. The widely used equivalent circuit model (ECM) generally shows poor behavior in predicting battery terminal voltage at low sate of charge (SOC), increasing the risk in the urgent use of a battery at low voltage greatly. To model strong nonlinearity of battery open circuit voltage (OCV), a solid-phase diffusion equation based surface SOC is proposed to characterize OCV behavior and establish the new structure of the enhanced ECM to describe low SOC behavior more precisely. Finally, a battery test platform was built to conduct battery tests for model validation. The results show that the root mean square error (RMSE) of the battery terminal voltage obtained from the proposed model at low SOC has been reduced to 8 mV compared with the RMSE of 17 mV from the traditional ECM model. It is expected that the proposed model can be employed in battery management systems to effectively improve the reliability and safety of emergency use of a battery at low SOC. [ABSTRACT FROM AUTHOR]

• An improved model with the voltage as input and current as output (VICO) is proposed. • The established VICO model is extended to an n -order VICO model. • The fault diagnosis method of current sensor is realized with the first-order VICO model. • The adaptability under different operating conditions and merit in detecting time are verified. Battery management systems (BMSs) are very important to ensure the safety of electric vehicles. The normal operation of BMSs is highly dependent on the accuracy of battery sensors. The present fault diagnosis efficiency of current sensors is much lower than that of voltage sensors due to model limitations in conventional methods. In this paper, a fault diagnosis method based on an improved model with voltage as input and current as output (VICO) is proposed to detect current sensor faults, where the least squares method combined with the unscented Kalman filter is used to estimate the fault current of current sensor. By comparing the estimated fault current with the diagnosis threshold, the fast fault diagnosis of current sensor is realized. The proposed method is verified under different operating conditions and compared with the methods based on state of charge and open-circuit voltage residuals. To highlight the importance of the proposed method, the influence and possible causes of minor faults and temperature on diagnosis are analyzed. The experimental results show that the method can detect the fault of the current sensor more accurately and quickly compared with the conventional methods, and has the ability to detect minor faults and adaptability under different operating conditions and temperatures. [ABSTRACT FROM AUTHOR]