Search

The accurate estimation of the state of charge (SOC) is crucial for the safe and reliable operation of Li-ion batteries. In this paper, novel dual-channel deep learning methods were proposed for SOC estimation during dynamic driving cycles. For dual-channel deep learning methods, each channel of the deep learning methods could be a single convolutional neural network (CNN) layer, single long short-term memory (LSTM) layer or sequentially combined structure of CNN and LSTM. After evaluation, it could be found that dual-channel deep learning methods could achieve higher accuracy than conventional single-channel method, however, the model size and floating-point operations (FLOPs) could be either reduced or increased. Among all the dual-channel deep learning methods, CNN + LSTM-CNN + LSTM method could achieve the highest accuracy and the shortest training time with sacrifices in computational complexity. CNN–CNN + LSTM method could achieve higher accuracy than the single-channel deep learning method with reduced computational complexity and model size. As the error of SOC estimation is highly correlated with voltage signals, conditional judgement was then integrated with dual-channel deep learning method to obtain higher accuracy. After integrated with conditional judgement, the averaged error of SOC estimation could be reduced by more than 45% without a substantial sacrifice in computational time. • Dual-channel deep learning methods are proposed for SOC estimation. • Two channels are responsible for processing of transient and moving averaged signals. • Model hyperparameters are optimized by Tree-structured Parzen estimator. • Cross validation method is applied to evaluate the proposed methods. • Conditional judgement is proposed to obtain higher accuracy. [ABSTRACT FROM AUTHOR]

Accurately estimating the state of charge (SOC), state of energy (SOE), and state of health (SOH) online is a critical and urgent concern in the management of lithium-ion batteries for electric vehicle applications, particularly in terms of safety and reliability. This paper develops a hybrid neural network, abbreviated as CNN-SAM-LSTM model, which combines a convolutional neural network, self-attention mechanism, and long-short term memory neural network to jointly estimate the state parameters of lithium-ion batteries, including SOC, SOE, and SOH. Additionally, a joint loss function considering homoscedastic uncertainty is developed to optimize weight adjustments for the training losses associated with the three state parameters. Experimental data collected under UDDS, BBDST and CC discharge conditions are employed to showcase the effectiveness of the proposed CNN-SAM-LSTM model. The results demonstrate that the proposed model is capable of simultaneously and accurately estimating SOC, SOE, and SOH for lithium-ion batteries under different dynamical operating conditions. Moreover, when applied to randomly segmented data, the proposed model exhibits robustness, effectively handling deviations from random discharge segments. • A CNN-SAM-LSTM hybrid neural network is developed to estimate SOC, SOE and SOH. • The loss function of the hybrid network considers homoscedastic uncertainty. • The proposed method achieves high accuracy under dynamic operating conditions. • The proposed method is strongly robust against random discharge segments. [ABSTRACT FROM AUTHOR]

Capturing the lifespan trajectory of lithium-ion (Li-ion) batteries in the early stage is critical for the operation and maintenance of battery energy storage systems (BESSs). Recently, data driven model is a promising solution to implement this task, yet the battery's early cycling stage can only provide very limited information in the training phase and the scenarios of the BESS applications are not immutable. To alleviate the above issues, this paper proposes a multi-scenario transferable learning framework with few-shot to predict the Li-ion battery lifespan trajectory. An easily constructed model is chosen to generate various pseudo trajectories from only the full lifespan trajectory of a single cell. Then, a transferred deep learning method integrating the gate recurrent unit (GRU) and one-dimensional convolutional neural network (1D CNN) is proposed to utilize the pseudo curves, whose training ability can be adjusted to a new scenario using simply the first 100 cycling data and a pre-trained strategy. Finally, the proposed framework can accurately predict the whole Li-ion battery aging trajectory for multi-scenarios in two different datasets. • A multi-scenario transferable learning framework is proposed to predict the lifespan trajectory of lithium-ion battery. • Various pseudo trajectories are generated by an easily constructed model from only one cell. • A transferred deep learning method is proposed to utilize the pseudo curves. • The proposed framework can accurately predict the battery aging trajectory in two different datasets. [ABSTRACT FROM AUTHOR]

Battery-related faults have become the most intractable problem hindering the further prosperity of fields like electric vehicle and grid energy storage. This paper is devoted to constructing a novel diagnostic framework for the faults in series battery packs, resorting to signal imaging and convolutional neural network (CNN) techniques. First, the voltage synchronicity between adjacent cells in a pack is quantified using the recursive correlation coefficient which can percept system anomalies sensitively. Then, reliant on the Gramian Angular Field (GAF) and Markov Transition Field (MTF) transformations, the correlation coefficient series is converted into pseudo images, the textures of which are full of informative details regarding system state. Finally, CNN models are employed to analyze the images for fault symptoms, thereby detecting fault occurrence, inferring fault type and evaluating fault grade. To obtain realistic dataset, different types and severities of faults are physically triggered on a li-ion battery pack. Experimental verification results indicate that the proposed framework can give accurate and reliable judgements on fault specifics, with the accuracy rates of fault type isolating and severity grading as 99.63% and 63.6% on GAF images, and as 99.75% and 58.7% on MTF images, respectively. • An improved correlation coefficient utilized to percept anomalies in series battery packs. • Signal imaging techniques employed to emphasize cross-temporal characteristics as image textures. • A CNN-based diagnostic framework proposed to infer fault type and evaluate fault grade. • Realistic dataset of different types and severities of faults to verify the framework. [ABSTRACT FROM AUTHOR]

Accurate prediction of the battery remaining useful life (RUL) at different operating conditions is critical for the battery management system to guarantee safe and efficient operation. However, because of the complicated degradation mechanisms inside the battery, it is extremely challenging to predict the battery life by measuring the external variables. Due to the sparse and random segment data in practical applications, the existing methods are difficult to be applied for online prediction. In this paper, a hybrid parallel residual convolutional neural networks (HPR CNN) model for RUL prediction is proposed. By fusing the charging data of voltage, current and temperature curves in multiple cycles, the hidden feature information of different depths is effectively extracted through the residual network. Based on the sparse data corresponding to only 20% charging capacity, combined with a cloud computing system, this method is able to achieve online prediction in various practical applications. By calculating the difference between each cycle as supplementary input data, the method is able to predict the RUL of a battery with high accuracy and reliability. Validated by a public data set and compared with other methods, the proposed method achieves a low test error of 4.15%, which is promising to be applied in the conditions of random charging process. • A method of online RUL prediction via a cloud computing system is presented. • The RUL is described by sparse segment data of battery V, I and T during charging. • A HPR CNN model is developed to extract and learn degradation features. • The model comprehensive performance is verified by comparing with different methods. • The prediction error of proposed method can be as low as 4.15% and 16.09 cycles. [ABSTRACT FROM AUTHOR]

Capacity estimation is an essential task for battery manage systems to ensure the safety and reliability of lithium-ion batteries. Considering the uncertainty of charging and discharging behavior in practical usage, this paper presents a one-dimensional convolution neural network (1D CNN)-based method that takes random segments of charging curves as inputs to perform capacity estimation for lithium-ion batteries. To improve the robustness and accuracy of the proposed 1D CNN network, a linear decreasing weighted particle swarm optimization algorithm is utilized to optimize the partial hyperparameters of neural network. Experimental data from two sets of batteries with different nominal capacities are employed for verification purpose. It is proved that the proposed method is feasible to provide accurate estimations on capacity degradation for both kinds of batteries. Furthermore, effects of length and relative position of segments on the capacity estimation are also investigated. The analysis results show that a more precise estimation of the battery capacity is prone to be obtained from the segment with a longer length or lower initial SOC. • A 1D CNN based battery capacity estimation model is developed for Li-ion batteries. • Capacity of the battery can be estimated from random segmented charging curves. • LDWPSO is utilized to optimize key hyperparameters of the 1D CNN model. • Segments with long lengths/low initial SOCs are preferred for the 1D CNN model. [ABSTRACT FROM AUTHOR]