Research of BP network based solid oxide fuel cell stack temperature model

The working temperature environment of the stack plays the key role in maintaining the high efficient, stable and secure operation of the system. This paper based on BP network model for the prediction of reactor temperature of SOFC electric power system, so as to implement thermal administration. Through repeated training on the model, selection of proper parameters and the test of the trained model, the error between test results and actual data is within ±1°C, which proves the effectiveness of this model and shows that this model can be applied in the thermal management of SOFC electricity supply system.