Virtual Thermal Sensing Based on Inverse Heat Conduction for Internal Temperature Estimation in Polymer Electrolyte Membrane Fuel Cell
Please login to view abstract download link
Fuel cell technology, particularly the Polymer Electrolyte Membrane Fuel Cells (PEMFCs), has gained increasing attention as a clean and sustainable energy source. Maintaining optimal operating temperature is essential to prevent material degradation and performance loss. However, direct measurement of internal temperature in PEMFCs remains challenging due to limited sensor placement accessibility and the risk of disrupting normal operation. To address this issue, a virtual thermal sensing framework based on the Inverse Heat Conduction Problem (IHCP) is developed. Because IHCPs are inherently ill-posed, a receding-horizon filtering approach is used to stabilize the inverse solution. Incorporating Computational Fluid Dynamics (CFD) results enables the use of realistic convective heat transfer coefficients and flow velocities to represent the effects of gas flow channels, thereby improving model fidelity. A model order reduction technique based on the Krylov subspace projection is further applied to achieve real-time computation. Experimental validation under various operating conditions, including controlled and uncontrolled relative humidity, and cell voltages of 0.4V and 0.6V, demonstrates excellent agreement between reconstructed and measured temperatures. The proposed method offers accurate and efficient estimation of internal heat sources and temperature distributions, even in configurations where direct sensing is impractical. Beyond PEMFCs, the technique holds strong potential for applications in electric motors, electric vehicle batteries, or thermal management of semiconductor devices.
