Research On Modeling And Control Of Proton Exchange Membrane Fuel Cell System

With the world facing the global warming problem, fuel cell technologies are viewed as one of the promising energy technologies for sustainable future due to their high energy efficiency and environment friendliness. Compared with the other types of fuel cells, a proton exchange membrane fuel cell (PEMFC) shows promising results with its advantages such as low temperature, high power density, fast response, good stability and zero emission if it runs with pure hydrogen, and it is suitable to be used in portable power supply, hybrid vehicles, and distributed power plants.The dissertation researches on modeling of PEMFC system for vehicles, modified particle swarm optimization as well as parameter identification for PEMFC model, PEMFC control problems based on H∞suboptimal control and multi-energy control strategies of hybrid vehicles based on fuel cell. The main contributions obtained in this dissertation are as follows.(1) According to the modeling principle of mechanism model and identification model of PEMFC, a mixed dynamic model of PEMFC system for vehicles which includes mechanism model and auxiliary equipment model of PEMFC is proposed. In this system model, the mechanism model part is composed of output voltage model, cathode flow model, anode flow model and membrane hydration model;the auxiliary equipment model part is composed of air compressor model, manifold lumped model, air cooler model and humidifier model. This system model overcomes the drawbacks of high complexity and too many parameters in the mechanism model of PEMFC system and the shortcoming of large amounts of experimental data needed and high cost in the identification model. This system model also considers output voltage subsystem, air supply subsystem and humidifying subsystem, etc. This proposed system model will be the foundation for the control system design in follow-up chapters.(2) Due to the problem of standard PSO wihich is difficult to jump out of local optimum and can make equilibrium point fall into logjam, causing premature convergence, a modified particle swarm optimization (MPSO) algorithm which has preferable global search ability and search speed is proposed in this dissertation. This algorithm main includes effective informed strategy; adaptive inertia weight and acceleration coefficients strategies; local search strategy based on BFGS Quasi-Newton method; randomized regrouping strategy. In this dissertation, MPSO is carried out to test function optimization for the benchmark function problem to verify the performance of proposed algorithm. Meanwhile, MPSO is compared with other wide improved PSOs. (3) A PEMFC experimental test bench is developed in Clean Energy Centre (CEC), Temasek Polytechnic, Singapore. A PEMFC designed and fabricated by CEC is used to achieve the testing of polarization characteristics with the fuel cell testing system. Based on the experimental data and combined with the proposed PEMFC model, MPSO is utilized to achieve the parameter identification for the key parameters of output voltage model under the condition of noise and absence of noise. In order to verify the advantage of MPSO, the comparisons of identified results with other improved PSOs are carried out.(4) In order to protect a PEMFC system because of oxygen excess, a novel reduced H∞suboptimal output feedback controller as 2 degrees of freedom (2DOF) controller which can maintain oxygen excess ratio (OER) better at the optimal operating point combined with a pre-compensator is designed for a nonlinear PEMFC system. Considering the existence of uncertainties and disturbances, the robust performance of the control system is verfied as dynamic running process of electrical vehicle is simulated. Meanwhile, the comprehensive comparisons with other control methods are carried out. The experimental system of Ballard 1.2kW Nexa is performed qualitatively to compare with the simulation results and also the feasibility and the validity is verified for the proposed control system.(5) In order to prevent the damage of proton exchange membrane, guarantee stable operation of a PEMFC system and prolong its working lifetime, a nonlinear H∞, suboptimal output feedback method is proposed and a PEMFC pressure control system is developed. According to a nonlinear transformation of coordinate and a dynamic extension algorithm, a nonlinear state feedback control rule could be solved and a linear and controllable Brunovsky canonical form is then obtained. Furthermore, the proposed H∞suboptimal output feedback method based on LMI is combined with state feedback exact linearization and then a nonlinear H∞suboptimal output feedback contoller is designed. To verify the disturbance restraint ability of the proposed controller, the dynamic simulation is carried out for the pressure control system as load current of fuel cell testing system with continuous step is simulated. Meanwhile, the comprehensive comparisons with other nonlinear control approaches are achieved.(6) A secondary development for electric vehicle simulation software ADVISOR is implemented with the system architecture of hybrid vehicle based on fuel cell and battery (FC+B) in this dissertation. The loading files are reconfiguration and then a model of hybrid vehicle based on fuel cell, battery and ultra-capacitor (FC+B+UC) is developed. In order to enhance the fuel economy of hybrid vehicle and increase the mileage of continuation of the journey, a fuzzy control method is utilized to design relevant multi-energy control strategy for the FC+B hybrid vehicle and the FC+B+UC hybrid vehicle. According to different standard cycle conditions, the proposed control strategy is contrasted with the power tracking control strategy which is wide adopted in ADVISOR in terms of the indexes of fuel economy and dynamic property. The rationality and the validity of the proposed control strategy on condition that satisfies the power requirement of working condition are verified for the power distribution among various power sources.