Probabilistic Prediction Algorithm for Cycle Life of Energy Storage in Lithium Battery

Abstract

Lithium batteries are widely used in energy storage power systems such as hydraulic, thermal, wind and solar power stations, as well as power tools, military equipment, aerospace and other fields. The traditional fusion prediction algorithm for the cycle life of energy storage in lithium batteries combines the correlation vector machine, particle filter and autoregressive model to predict the cycle life of lithium batteries, which are subjected to many uncertainties in the prediction process and to inaccurate prediction results. In this paper, a probabilistic prediction algorithm for the cycle life of energy storage in lithium batteries is proposed. The LS-SVR prediction model was trained by a Bayesian three-layer reasoning. In the iterative prediction phase, the Monte Carlo method was used to express and manage the uncertainty and its transitivity in a multistep prediction and to predict the future trend of a lithium battery's health status. Based on the given failure threshold, the probability distribution of the residual life was obtained by counting the number of particles passing through the threshold. The wavelet neural network was used to study the sample data of lithium batteries, and the mapping relationship between the probability distribution of the residual life of lithium batteries and the unknown values were established. According to this mapping relation and the probability distribution of the residual life of lithium batteries, the health data could be deduced and then iterated into the input of the wavelet neural network. In this way, the predicted degradation curve and the cycle life of lithium batteries could be obtained. The experimental results show that the proposed algorithm has good adaptability and high prediction efficiency and accuracy, with the mean error of 0.17 and only 1.38 seconds by average required for prediction.World Electric Vehicle Journal 2019, 10, 7 2 of 17 directly measured but needs to be estimated in advance to decide whether to replace the battery to avoid some unnecessary events. The performance of batteries can be divided into two categories: electrical performance and reliability. Battery life is one of the important indicators to measure the electrical performance of batteries. Charge-discharge cycles include a charge operation and a discharge operation. The number of charge-discharge cycles that a battery can carry out while maintaining a certain output capacity is called the cycle life (service life) of the battery. For energy storage batteries, it is generally believed that the life of the battery is terminated when the available capacity of the battery decreases to 70% of the initial level [4]. Battery life includes cycle life and calendar life, where the former refers to the number of cycles of the battery from a certain charging and discharging system to the end of life and the latter refers to the time required for the battery to be stored in a certain state until the end of life. There are many complex physical and chemical reactions in the charge and discha...

Toward Group Applications: A Critical Review of the Classification Strategies of Lithium-Ion Batteries

“…References [138] adopts capacity-based statistical models for sorting batteries. Reference [139] adopts the cycle probability prediction algorithm to predict the battery cycle life. Reference [140], according to capacity distribution, order statistics, central limit theorem, and converter efficiency, optimizes energy efficiency by finding the best number of cells in a battery pack.…”