Dynamic Modeling And State Estimation Of Water Transport In Vehicular Polymer Electrolyte Membrane Fuel Cell

The promotion of proton exchange membrane(PEM)fuel cell is significant to the energy structure optimization and air pollution control.Due to the development on the vehicular fuel cell technology for years,the gap between our country and the world advanced level is getting closer.Even though,the problems on the PEM fuel cell's performance,cost and durability are still challenging.Among of them,the water management is a key issue which affects both performance and life.An awful water management may lead to two unnormal phenomena,membrane dehydration and water flooding.The water management is also a problem which involves with several physical and chemical processes,such as fluid mechanics,heat and mass transfer and electrochemical reaction.Supported by the project of National Natural Science Foundation of China,“Research on the model and control of the high performance,long life and low cost vehicular fuel cell powertrain system” and other projects,this dissertation systematically describes and analyzes the one-dimensional water transport mathematics model,humidification optimization,dynamic model and model-based state estimation problems.This dissertation firstly developes a mathematics model to describe the water transport processes happening in the through-plane direction.The cell level model includes all the water transport models of different states at different positions.Then the dissertation proposes three regimes of water two-phase one-dimensional stable distributions in gas diffusion lays(GDLs)and uses analytical approach to get analytic solutions under different regimes.With the Iterative computation for the net water flux through the membrane,one-dimensional stable distributions of multi-state water can be obtained quickly.Secondly,based on multi-state water one-dimensional distributions,the dissertation studies the effect of the water with different states and in different components on the fuel cell output voltage,defines anode over-humidification and humidification optimization path,and presents the humidity optimization MAP to acquire better humidification.Thirdly,based on research achievements on the water transport,the dissertation developes a cathode-Inlet-Outlet(CIO)model to describe water transport dynamics of vehicular fuel cells with big active area.The CIO model considers the effect of current density distributions.Finally,based on the state space description of the CIO model,the dissertation presents a state estimation problem in which the cathode liquid saturation and the difference of current densities are two target observed states,and further designes states estimation algorithms,using Luenberger observer and Unscented Kalman Filter(UKF)respectively,to realize the online prediction of liquid water inside the cell under flooding conditions.The work in this dissertation basically deals with the water transport mechanisms and finally solves the flooding diagnosis issue of the actual vehicular fuel cell system.It is valuable for the design and optimization control of the PEM fuel cell systems.