Skip to Main Content
In this paper, we established a soft sensor model to calculate the moisture of polymer electrolyte membrane fuel cells by artificial neural network. We trained ANN by an improved ant colony algorithm. Experimental tests indicate that the simulation results of PEMs' moisture are very close to real values, and the method possesses high precision and speed and can meet actual demands. This soft sensor model can be applied in the control of PEMFCs' moisture and temperature.