Cloud-to-edge based state of health estimation method for Lithium-ion battery in distributed energy storage system
Ji Wu, Xingtao Liu, Jinhao Meng, Mingqiang Lin
Abstract
Of the key parameters in the battery management system, the state of health is the most vital one concerning the distributed energy storage system's safety. Due to the limitation of the computing power of the battery management system in the actual application, a cloud-to-edge based state of health estimation method is proposed in this paper, where the battery management system is used to measure and pre-process the voltage and current data of the battery in the edge side. The cloud platform is utilized to estimate the state of health with an advanced data-driven algorithm on the cloud side. To reduce the complexity of the estimator, save the network traffic and decrease the impacts from measurement noises, a 3-round feature selection approach is developed to extract the measured battery data from the charging process. Afterward, a random forest regressor is applied to build the battery degradation model with the selected features and then estimate the state of health. Experimental results show that using the selected features may have a sufficient estimating accuracy while costing fewer traffic data and calculations. The proposed method also has a solid ability to reduce the effects of the noises. Moreover, experiments with different lithium-ion batteries are conducted to demonstrate the universality of the proposed method.