Intelligent hybrid system for the prediction of the voltage-current characteristic curve of a hydrogen-based fuel cell

Due to some reasons like sustainability and energy strategy, there is a clear trend using new ways to obtain energy, more efficient and, usually, renewables. In addition, with other different objectives, many researchs are being carried out on energy storage systems; one of the most promising, in terms of capacity and mobility, is hydrogen-based. In the present work a model is obtained to predict the dynamic behavior of a hydrogen fuel cell, which will improve its control. The variables used in this research have been extracted from a test bench, where a fuel cell is monitored under several load conditions with a programmable load connected to its output. To perform this model, a hybrid intelligent model was chosen. This kind of models use clustering techniques to divide the data set and, after that, intelligent regression algorithm with artificial neural networks are used for each group. The proposal has been tested with two validation data set, obtaining highly satisfactory results.