The Study Of Li-ion State Of Health Prediction Based On Elman Neural Network

Li-ion battery is internationally recognized as a ideal chemical energy source. As the latest secondary battery, it is popular due to its extraordinary performance. The performances of Li-ion battery keep improving needs battery testing technology matching the pace, the life span of Li-ion battery is a key issue in battery testing system design. In order to estimate life and health status of the battery, State of Health is proposed. Therefore, this paper establishes a state of health prediction model of Li-ion battery, providing a effective way to monitor state of health in real time.At the present stage, the evaluation method of state of health is normally on external feature of the battery, and its prediction result is unsatisfied. This paper utilizes internal resistance parameter identification to analyze the relativity between the internal feature and state of health of the battery, and proposes a new state of health prediction method. This method adopts internal resistance to predict battery state of health, and the prediction accuracy improves, at the same time, fix the problem exists in tradition method.This paper takes ironic phosphate Li-ion power battery as study object. After analyze the characteristics of Ohm internal resistance and polarization resistance of battery, a simple equivalent circuit model of the battery is established, and the parameter in the model is identified with recursive least square method. Apply the internal resistance identify model to charging-discharging cycle life experiment to analyze the relation between internal resistance and state of health. A state of health prediction model for Li-ion battery is established using Elman neural network, which take internal resistance as input parameter. The prediction results meet the expected effect. Hereditary algorithm is adopted for weigh optimization of the prediction model. Simulation results verify that the prediction results becomes better.The study in this paper indicate that, in Li-ion battery research, take internal resistance as a parameter of state of health prediction is feasible, also provide a new direction for battery state of health prediction...