Lithium-ion batteries are widely used as effective energy storage and have become the main component of power supply systems. Accurate battery state prediction is key to ensuring reliability and has significant guidance for optimizing the performance of battery power systems and replacement. Due to the complex and dynamic operations of lithium-ion batteries, the state parameters change with either the working condition or the aging process. The accuracy of online state prediction is difficult to improve, which is an urgent issue that needs to be solved to ensure a reliable and safe power supply. Currently, with the emergence of artificial intelligence (AI), battery state prediction methods based on data-driven methods have high precision and robustness to improve state prediction accuracy. The demanding characteristics of test time are reduced, and this has become the research focus in the related fields. Therefore, the convolutional neural network (CNN) was improved in the data modeling process to establish a deep convolutional neural network ensemble transfer learning (DCNN-ETL) method, which plays a significant role in battery state prediction. This paper reviews and compares several mathematical DCNN models. The key features are identified on the basis of the modeling capability for the state prediction. Then, the prediction methods are classified on the basis of the identified features. In the process of deep learning (DL) calculation, specific criteria for evaluating different modeling accuracy levels are defined. The identified features of the state prediction model are taken advantage of to give relevant conclusions and suggestions. The DCNN-ETL method is selected to realize the reliable state prediction of lithium-ion batteries.

No abstract

Summary To accurately evaluate the state of charge (SOC) and state of health (SOH) of Li‐ion battery, the second‐order RC equivalent‐circuit model is used to characterize the battery performance, a novel dual adaptive Kalman filtering algorithm is presented by adding double cycles and noise adaptive steps to realize the joint estimation of the SOC and internal resistance. The state variables can be corrected with each other as go through the cycle under three operating conditions. The accuracy of the SOC estimation method proposed in this paper is significantly improved compared with the extended Kalman filtering and the unscented Kalman filtering algorithm. Under three operating conditions, the average error and the maximum error decreased obviously. An equation for calculating the SOH in terms of internal resistance increase was built. The estimation result of the SOH effectively simulated the actual situation, compared with the actual result, the maximum error under the three operating conditions are within a lower level than the improved unscented Kalman filtering algorithm. The convergence effect of the algorithm has obvious advantages over that of the algorithm used for comparison, which could effectively track the state change of the battery.

For the lithium battery management system and real-time safety monitoring, two issues are of great significance, namely, the ability to accurately update the model parameters in real time and to accurately estimate the state of charge and health. In this context, this thesis adopts the second-order RC equivalent circuit model and the forgetting factor recursive least squares -double extended Kalman filtering (FFRLS-DEKF) algorithm with multi-time scales and low-pass filter. Forgetting factor recursive least squares is applied to conduct online parameter identification, and the traditional double extended Kalman filtering algorithm is optimized to evaluate the state of charge and model parameters in the micro-scale and macro-scale. In this way, the error caused by two different characteristics is reduced, and a low-pass filter is added to optimize the fluctuation problem of the estimated value of the model parameters. According to the experiment results, the maximum error between the model simulation value and the actual value of the terminal voltage is 0.0459 V. If the initial value of the state of charge deviates from the actual value, the maximum errors under BBDST and HPPC conditions record 0.0235 and 0.0048, respectively, the forgetting factor recursive least squares -double extended Kalman filtering algorithm with multi-time scales and low-pass filter is able to track the true value within 40 s. Furthermore, the lithium-ion battery state of health both reaches 98% under the two conditions. In summary, the experimental analysis shows that the algorithm helps reduce the influence of initial values on the results, thereby reducing error accumulation and improving the robustness.

He M, Cao W. Novel coestimation strategy based on forgetting factor dual particle filter algorithm for the state of charge and state of health of the lithium-ion battery. Int

No abstract

Summary The remaining useful life (RUL) is a core parameter of the battery management system. To realize accurately predict the RUL, the paper takes the National Aeronautics and Space Administration battery test data set as the research object, and a battery capacity degradation model based on an exponential growth model is built to characterize the battery aging process. A novel cuckoo search optimization particle filtering algorithm is proposed for the RUL prediction by transferring the particles in the prior distribution region of the particle filtering algorithm to the maximum likelihood region. The initial cycle numbers are set differently, the capacity decay process of the four groups of batteries can be predicted completely, and the RUL of the batteries can be obtained. Compared with the commonly used particle filtering and unscented particle filtering algorithms, the results show that the proposed method has obvious advantages in the relative error of prediction and resampling rate under the four datasets. For the B0005 dataset, the relative errors are 4.8%, 3.2%, and 0.8%, respectively under 30, 50, and 70 cycles, corresponding to the particle filtering algorithm errors are 10.3%, 7.9%, and 4.8%, and the unscented particle filter algorithm errors are 6.3%, 4.8%, and 2.4%, respectively. In addition, the method for RUL prediction present in the paper has a high confidence level, and low resampling rate, which plays an important role in promoting the further application of lithium‐ion batteries. Highlights A novel cuckoo search optimization particle filtering algorithm is proposed. Different battery aging cycle number is distinguished to predict. Multiple battery RUL prediction is realized. RUL prediction has high prediction accuracy and high confidence.

No abstract