MIMO Nonlinear Identification of a PEM Fuel Cell Humidifier Using Wavelet Networks

Humidification is a key factor influencing the performance of a proton exchange membrane (PEM) fuel cell system. In order to achieve the optimal humidity by means of some control strategy, it is essential to obtain an accurate model of the humidifier, which is a multi-input-multi-output (MIMO) nonlinear system. To avoid the curse-of-dimensionality problem, a class of wavelet networks proposed by Billings was employed to identify the model of the humidifier in this work. An efficient model term selection approach was applied to solve the high dimensional problem. The model was identified based on the experimental data acquired from our test bench. The one-step-ahead predictions and the five-step-ahead predictions were compared with the real measurements respectively. It shows that the identified model can effectively describe the real system