Due to the constraints of manufacturing and materials

high-power plants cannot rely on only one solid oxide fuel cell stack. A multi-stack system is a solution for a high-power system

which consists of multiple fuel cell stacks. A short lifetime is one of the main challenges for the fuel cell before large-scale commercial applications

and prognostic is an important method to improve the reliability of fuel cells. Different from the traditional prognostic approaches applied to single-stack fuel cell systems

the key problem in multi-stack prediction is how to solve the correlation of multi-stack degradation

which can directly affect the accuracy of prediction. In response to this difficulty

a standard Brownian motion is added to the traditional Wiener process to model the degradation of each stack

and then the probability density function of the remaining useful life (RUL) of each stack is calculated. Furthermore

a Copula function is adopted to reflect the dependence between life distributions

so as to obtain the remaining useful life for the whole multi-stack system. The simulation results show that compared with the traditional prediction model

the proposed approach has a higher prediction accuracy for multi-stack fuel cell systems.