Semi-physical modeling and control of a centrifugal compressor for the air feeding of a PEM fuel cell

Air compressor which is to used to deliver the air (including about 21% oxygen) to the cathode channel for electrochemical reaction is a crucial component for the polymer electrolyte membrane (PEM) fuel cell. The working characteristics of the compressor greatly influences the fuel cell performance as well as the durability. In this paper a semi-physical modeling method is adopted to analyze the operating property of a centrifugal compressor which is more suitable for automotive fuel cells because of its compactness. This model includes many physical and empirical parameters which are very difficult to determine. Interior-point optimization method based on Newton iteration is used to identify those parameters. The result shows that the modeled compressor map has a good agreement with the experimental data. Meanwhile, the compressor efficiency is analyzed and compared with the measurement to further validate the developed model. This compressor is thus adapted to a validated 10 kW fuel cell model. A dynamic feedforward controller is proposed based on the load torque to control the air mass flow, eliminating the disturbance produced by the compressor load in transient. The simulation results show that this compressor could satisfy requirements of the fuel cell under dynamic load situations while keeping both the fuel cell and compressor with high efficiencies.