Research On Observation And Control Of Fuel Cell

Fuel cell(FC) is a new type of power technique which converts chemical energy of fuel (the natural gas or hydrogen etc.) into electricity. Compared with the traditional way such as fossil-fueled power generation, fuel cell has the merits of high thermal efficiency, low noises and light contamination. Therefore, fuel cell is an ideal power technique in the future. The research on observation and control of fuel cell is economically and strategically important.This paper first analyzes the state of art and working principle of fuel cell and then emphasizes on discussing two problems: observation of the partial pressure of hydrogen and predictive control about the output voltage of fuel cell. The most difficult problem in fuel cell power system is the regulation of hydrogen. Insufficient hydrogen will shorten the lifetime of fuel cell, and excessive hydrogen will reduce the efficiency. We study a nonlinear adaptive observer and estimate the partial pressure of hydrogen at the input and output by the available voltage, current and total pressure measurement. We also analyze the robustness of observer in the presence of measurement error. The effect of observer is verified by simulation.Fuel cell operates under a very complicated environment, which results in the fluctuation of output voltage. This paper proposes a subspace predictive control (SPC) algorithm for the output voltage. This new algorithm merges the subspace identification method and predictive control, avoiding finding the system matrices. Subspace predictive control is a data-driven method. It merely obtains the open loop input and output data to construct the Hankel matrices. The parameters of the controller are determined by QR decomposition of these Hankel matrices. This method doesn't need the accurate model. At the end of paper, the effectiveness of this controller is validated via simulation using Matlab toolbox.