The rapid development of battery technology has promoted the deployment of electric vehicles (EVs). To ensure the healthy and sustainable development of EVs, it is urgent to solve the problems of battery safety monitoring, residual value assessment, and predictive maintenance, which heavily depends on the accurate state-of-health (SOH) estimation of batteries. However, many published methods are unsuitable for actual vehicle conditions. To this end, a data-driven method based on the random partial charging process and sparse Gaussian process regression (GPR) is proposed in this article. First, the random capacity increment sequences (△Q) at different voltage segments are extracted from the partial charging process. The average value and standard deviation of △Q are used as features to indicate battery health. Second, correlation analysis is conducted for three types of batteries, and high correlations between the features and battery SOH are verified at different temperatures and discharging current rates. Third, by using the proposed features as inputs, sparse GPR models are constructed to estimate the SOH. Compared with other data-driven methods, the sparse GPR has the highest estimation accuracy, and its average maximum absolute errors are only 2.88%, 2.52%, and 1.51% for three different types of batteries, respectively.

Significant progress has been made in transportation electrification in recent years. As the main energy storage device, lithium-ion batteries are one of the key components that need to be properly managed. The remaining useful life, which represents battery health, has attracted increasing attention. Because accurate and robust predictions provide important information for predictive maintenance and cascade utilization. This paper proposes a novel method to predict remaining useful life based on the optimized health indicators and online model correction with transfer learning. Gaussian process regression is used to optimize the threshold for health indicators to determine the end of life, and a usefulness evaluation strategy is proposed to assess the health indicators. Then, a combination of transfer learning and gated recurrent neural network is designed to predict the remaining useful life based on the optimized health indicators directly, which can promote online applications. The prediction model initially trained based on a relevant battery is further fine-tuned according to the early degradation cycling data of the test battery to provide accurate predictions. Moreover, a self-correction strategy is proposed to retrain the regression models so that the models can gradually reach the optimal prediction performance during the operating cycles, which could not be achieved by traditional methods. The recommended input sequence lengths for potential applications are discussed. The method is verified by experiments of a batch of batteries under fast charging conditions, and the results show that, after fine-tuning, the proposed method predicts remaining useful life with an error of fewer than 5 cycles.