Intelligent control for fuel cells
Fuel cells convert the chemical energy stored in hydrogen-rich fuels into electricity that can be used in stationary power systems and portable appliances. Hydrogen can be produced through water electrolysis using renewable energy. Not surprisingly, polymer electrolyte membrane (PEM) fuel cells emerge as one of the cleanest alternative energy sources. However, several technical issues need to be addressed to make them competitive. The EU-funded project ACOFC (Advanced controllers and observers development for fuel cell based generation systems) aimed to develop a robust control solution to deal with disturbances that affect the system operation. By their very nature, PEM fuel cells are sensitive to changes in operation conditions, such as the relative humidity of reacting gases and their temperatures. The ACOFC team demonstrated the application of higher-order sliding-mode control to PEM fuel cells. This control approach involves mapping the system's input influence matrix to a transformed system's matrix that is invertible. Among the advantages, confirmed at the Institut de Robòticai Informàtica Industrial in Barcelona, Spain on the laboratory test bench, is humidity control. In addition, the oxygen stoichiometry control has to be able to optimise the system conversion efficiency, avoiding possible performance deterioration and damages to the polymeric membranes. The experimental data under a broad range of operation conditions were also in good agreement with simulation results. Furthermore, several solutions were proposed for the measurement scheme in fuel cell-based systems. This is also an important issue for industrial applications where instrumentation needs to be minimal. Details of the controller design using higher-order sliding mode algorithms are described in a series of papers published in leading peer-reviewed journals and presented at international conferences.
Keywords
Fuel cells, hydrogen-rich fuels, polymer electrolyte membrane fuel cells, sliding-mode control, algorithm