Research On Health Prognostic Methods For Lithiumion Batteries In Satellite

Lithium-ion battery is a potential type of secondary battery that was originally used in mobile electronic products such as digital cameras,MP3/4,mobile phones,etc.With the development and maturity of technology,its advantages of high energy density,high safety,low pollution,and light weight have been continuously revealed,and it has gradually been applied to heavy industrial products such as electric vehicles,satellites,and green energy storage.Most of the loads on satellites are power-consuming products,and satellites periodically appear in the sunlit and shadow areas in orbit environments.In the sunlit area,the solar photovoltaic panels can continue to generate electricity and supply the satellite loadings.However,in the shadow areas these loadings completely depend on the energy storage system in the satellite.Secondary batteries compose the key constitute part of the satellite energy storage system.The life and health status of secondary batteries determine the normal operation time duration of satellite loadings,and thus determine the failuretime of a satellite.In the history,artificial satellite adopted nickel-cadmium and nickel-hydrogen batteries as the energy storge system,which played a significant role in a period of time.Howerer,these batteries have the drawbacks of low energy density,short cycle life and high self-discharge rate,and are difficult to meet the high reliability requirements of new-generation aerospace products.Lithium-ion batteries,as a new type of secondary battery with mature technology and bright prospects,are increasingly being applied to space products such as satellites and space stations.In this paper,researches on satellite lithium-ion battery health prediction methods are conducted to solve practical engineering problems such as battery life prediction,reliability assessment,and online residual life prediction.The research results lay a theoretical foundation and model basis for the design of satellite BMS(battery management system,BMS).This article starts from the battery failure mechanism analysis.Though exploring the internal failure mechanisms and external aging performances of the lithium-ion batteries,the key degradation parameter is determined and the health prediction models are built.The main contributions and innovations of this paper are as follows.(1)Off-line prediction techniques for satellite lithium-ion batteries.Taking the NASA laboratory battery cycle life test data as dataset,the framework of satellite lithium-ion battery off-line health prediction was established.Combining the failure mechanism analysis and the trend analysis results of the original experimental data,the capacity is modeled as a key feature to measure battery health status.Based on the failure-time data and capacity fading data,the battery life distribution model and reliability assessment model are established respectively,and the effectiveness of the two methods is compared from multiple angles.The battery reliability interval estimation method based on capacity degradation data is presented.Similarly,the failure data and capacity data are used to establish battery residual life prediction models.Results show that the capacity degradation data based method yields to more robust and accurate health prediction results.(2)In-orbit prediction techniques for satellite lithium-ion batteries.The framework of satellite lithium-ion battery lifetime prediction in online environments was established.Under laboratory conditions,battery capacity can be accurately measured by the full discharge method,and then the offline helath prediction(including reliability estimation and residual life prediciton)can be carried out on the basis of capacity degradation model.For in-orbit satellites,these batteries mainly experience partial charge and discharge,whose usable capacity cannot be directly obtained.Therefore,how to predict battery capacity accurately is the core work of online satellite battery health prediction.By analyzing the characteristics of online charging curves(voltage and current)and battery aging trends,two online capacity prediction approaches are proposed respectively to deal with the two cases of deep and shallow discharge.Finally,the online health prediction of satellite lithium ion batteries based on the predicted capacity results.(3)Multi-parameter prediction techniques for satellite lithium-ion batteries.During the charging and discharging process,there are several correlated features degrade simultaneously,and jointly determine the failure time of a battery.In this paper,a two-parameter prediction model for satellite lithium-ion battery is proposed.Two linear Wiener processes are used to describe the fade law of battery capacity and energy over cycle numbers.The Frank Copula function is used to describe the correlation between two degradation processes.The parameters in the degradation model are divided into fixed effect parameters and random effect parameters.The offline data are used to realize fixed parameters estimation,and the online data is used to perform random parameters updating with the particle filtering algorithms.Finally,the analytical expression of battery residual life distribution is derived under the Bayesian framework.