Fuzzy-based modelling technique for PEMFC electrical power generation systems emulation

A fuzzy-based technique for modelling both static and dynamic responses of the voltage in proton exchange membrane fuel cell energy power generation systems (PEMFC-EPGS) is presented. Based on dynamical effects present on the PEMFC, the technique requires lower calculation capacity than complex physical models. This modelling technique is validated by developing a fuzzy-based model of a user-designed PEMFC-EPGS comparing its accuracy and calculation time to a previously parameterised physical model. The technique is intended to support the design and simulation of control strategies and the development of PEMFC-EPGS models and hardware emulators. The simple structure of the models obtained with the technique results in a small calculation processing demand, this being an effective option for real-time and embedded implementations. Finally, the development and experimental validation of a fuzzy-based model in a low-cost emulator is presented. This emulator permits to evaluate control strategies and the behaviour of power systems interacting with the PEMFC-EPGS dynamics. Within the study and evaluation of efficient and renewable sources of clean electrical energy, fuel cell systems are under intense research and development efforts. This paper focuses on the proton exchange membrane fuel cells or PEMFC [۱]. In the PEMFCs, the fuel (hydrogen) is provided pressurised at the anode, arriving to a catalytic layer. In this layer, the hydrogen dissociates into protons and electrons. A PEMFC requires some support systems in order to produce electric power. These systems provide the reagents and regulate the pressure ratio between them and their temperature. In the same way, they regulate the stack temperature and manage the water produced. Also, the support systems inject water vapour to regulate the membrane humidity. The optimisation and control of these support systems are subjects of intense research and development. The solutions presently available permit the production of electric power in a useful form. A scheme of a PEMFC-EPGS is presented in Fig.۱,This topology is commonly used in electrical and hybrid automobiles, such as Ford P۲۰۰۰ [۲], and in experimental PEMFC prototypes. Some of them are commercially available as the Nexa Power Modules from Ballard or the PowerPEMTM-PS۲۵۰ PEMFC system from Hpower Co. Other PEMFC EPGS systems are user-designed implementations as the one presented by Friede et al. [۳]. The simplicity, small processing time and wide knowledge of implementation techniques make fuzzy and neural models appropriate choices for the development and simulation of control strategies, and also for the implementation of simulators and emulators of fuel cells and, in particular, of PEMFC-EPGS whose physical models are very complex and require large calculation times. In the test and evaluation of support hardware, control systems and fault detection strategies for mechanical systems and electrical power interfaces, experimentation on the physical systems is required. Some of these systems are expensive, and the required equipment and safety procedures are not easily provided by most control and power electronics laboratories. Examples of these systems are mechanical loads, wind turbines and PEMFC-EPGS. These difficulties have been addressed using emulators. In this way, Rodic et al. [۴] developed a mechanical load emulator, Han et al. [۵] developed an electrical load simulator, and Lin et al. [۶] and Wai et al. [۷] developed wind turbine emulators. In this paper, a fuzzy based modelling technique for the dynamic and static behaviour of a PEMFC-EPGS output voltage is described. An application to a user-designed PEMFC-EPGS is presented. This technique allows to minimise the required computational resources for its implementation. Both fuzzy-based and physical models are implemented in a real-time system, and a comparison of the computational loads is performed. The fuzzy-based model is designed to implement a PEMFC-EPGS emulator using a low-cost digital processing device that, in spite of its computational limitations, satisfies the real-time processing requirements. One of the main objectives of the proposed fuzzy-based modelling technique is the development of low-cost PEMFC-EPGS emulators. This paper presents a fuzzy basedmodelling technique for PEM fuel cells electric power generation systems. The proposed implementation of the technique require a small computation time, being suitable for real-time and embedded systems, like simulators and emulators. This fuzzy-based modelling technique allows a simple description of the dynamic effects that affects the PEMFC voltage and power. As a consequence of the simplicity, the required data for the identification of the fuzzy models are easily generable. The technique has been applied to a particular system, where the temperature, humidity and pressures ratio have been regulated. The application to this technique to more general cases is in progress. Following the fuzzy-based technique, a model of a user designed PEMFC-EPGS has been developed and compared with a previously parametrised physical model and experimental data. The results validate the fuzzy-based modelling technique proposed. Finally, the performed experiments demonstrate that the fuzzy emulator reproduces satisfactorily the behaviour of a user-designed PEMFC-EPGS. The emulator can be used in a preliminary step to develop control strategies and electronic power systems that later will be used in more expensive real PEMFC-EPGS