With the development of the space industry growing, space power system as the main power supply system of spacecraft, its work efficiency problem is to get the attention of people. The solar cell array- battery power system gets widely used for its simple structure, long service life. Battery is a subsystem of the solar cell array- battery power system.It is a part of the most vulnerable to failure. The residual life prediction research also got the attention of the researchers highly.In this paper, we get Matlab/Simulink simulation of the solar cell array space-battery power supply system. The experimental results prove the validity of the established system. The space power system operating mode was described in detail, providing a good platform for the space power system research.Lithium-ion batteries gets widely used in civilian electronic products, aviation products for its high specific energy, long cycle life, low self-discharge rate. The research of Remaining Useful Life(RUL) also plays a crucial role. In this paper, we use Particle Filter(PF) algorithm to establish lithium ion battery capacity degradation model. Comparing with the polynomial regression model, this model considers the electrochemical reaction in free time between the lithium ion battery charge and discharge cycle, making the model more close to the real degradation curve of lithium-ion batteries. Using particle filter algorithm to carry on the simulation of model, the results showed having only a small error between the model and the real lithium ion battery degradation curve. Using the capacity degradation model and particle filter algorithm to predict lithium ion battery RUL. For four groups of battery data from NASA PCo E testing center, we carry on the experiment. The experimental results verify the effectiveness of the proposed method. We analysis the probability density distribution of its remaining useful life. The simulation experiment of three kinds of resampling algorithms of particle filter algorithm was carried out. The experiment results verify the system resampling algorithm has better effect. |