Control and design of PEM fuel cell-based systems

Abstract

The use of fuel cell systems based on hydrogen is advantageous because of their high efficiency in the energy conversion and null emissions. In this thesis, an extensive study about the control and design of electrical generation systems based on fuel cells is performed. The main focus is in hybrid systems composed of fuel cells and supercapacitors as energy storage elements, oriented to automotive applications. The determination of the hybridization degree (i.e. the determination of the fuel cell size and the number of supercapacitors) is performed through a proposed methodology with the objective to fulfil the conductibility requirements and to consume the lowest amount of hydrogen.The process of design starts with the determination of the electrical structure and utilizes a detail model developed using ADVISOR, a MATLAB toolbox for modelling and studying hybrid vehicles. The energy flow between the vehicle components is analyzed when the vehicle is tested with different Standard Driving Cycles, showing how the losses in each component degrade the efficiency of the system and limit the energy recovery from braking.With regard to the energy recovery, a parameter to quantify the amount of energy that is actually reused is defined and analyzed: the braking/hydrogen ratio.To control the energy flow between the fuel cell, the energy storage system, and the electrical load in Fuel Cell Hybrid Vehicles (FCHVs), three Energy Management Strategies (EMSs) based on the fuel cell efficiency map are presented and validated through an experimental setup, which is developed to emulate the FCHV. The resulting hydrogen consumptions are compared with two references: the consumption of the pure fuel cell case, a vehicle without hybridization, and the optimal case with the minimum consumption. The optimal consumption for a given vehicle is determined through a methodology proposed that, unlike other previous methodologies, avoids the discretization of the state variables.To operate the fuel cell system efficiently, the system is controlled through a proposed control technique, which is based on Dynamic Matrix Control (DMC). This control technique utilizes the compressor voltage as control variable and also a new proposed variable: the opening area of a proportional valve at the cathode outlet. The control objectives are the control of the oxygen excess ratio at the cathode and the fuel cell voltage. The advantages of this new control variable are analyzed both in steady state and transient state. Simulation results show and adequate performance of the controller when a series of step changes in the load current is applied.On the other hand, the diagnosis and fault-tolerant control of the fuel cell-based system is approached. A diagnosis methodology based on the relative fault sensitivity is proposed. The performance of the methodology to detect and isolate a set of proposed failures is analyzed and simulation results in an environment developed to include the set of faults are given. The fault-tolerant control is approached showing that the proposed control structure with two control variables has good capability against faults in the compressor when the oxygen excess ratio in the cathode is controlled.