A fuel cell combines oxygen from the air with hydrogen to produce electricity, heat, and water. Since the fuel is converted directly into electricity, a fuel cell can operate at much higher efficiencies than internal combustion engines, extracting more power in the form of electricity from the same amount of fuel. With no moving parts, it is a quiet and reliable source of power, offering long-term prospects in many industries, not least for cars of the future.
Since computational fluid dynamics (CFD) was first introduced into automotive engineering, this sophisticated technology for the simulation of fluid flow and heat transfer has found applications in almost all areas of vehicle design from ergonomics to fuel injection. CFD techniques, including the multiphase aspects that are intrinsic to fuel cell operation, are now being used to understand and improve fuel cells for automotive applications. Computer simulation is much less expensive than physical testing, and this is a major advantage of the CFD technique.
Nearly all the major automotive manufacturers are developing fuel cells for entry into the commercial market. Honda began leasing fuel cell vehicles (FCVs) in America and Japan at the end of last year, while Nissan has brought forward the launch of its commercial FCV from 2005 to 2003. Fuel cell technology is generally regarded as the most promising candidate to replace the internal combustion engine because it is fuel-efficient and clean.
The type of fuel cell used in automotive applications involves a proton-exchange membrane (PEM). At one side, a catalyst splits the hydrogen into its component electrons and protons. The protons pass through the exchange membrane, while the electrons travel through an external circuit, producing electrical current. At the other side of the PEM, the hydrogen ions combine with oxygen and the electrons to complete the process and form water. CFD has made it possible to develop and optimise the relationships between the component parts of the fuel cell and the surrounding flows of fuel and energy. By trying different simulations, researchers have been able to predict and improve the efficiency of the components of the fuel cell.
One major application of CFD has been to simulate the moisture flow. In order to work effectively, the PEM must be hydrated at all times. The simplest solution, from a theoretical point of view, would be to humidify the fuel and oxidiser at the inlet so that everything is wet, but that would require a bulky external humidifier to be bolted on. That would not be a good option for a mobile application, so automotive manufacturers favour relying on the water produced internally by the reaction of the hydrogen and oxygen. A design difficulty exists in trying to retain the water within the fuel cell without losing it to exhaust. CFD flowfield simulations have been important in demonstrating the best way to retain the water and distribute it uniformly across the membrane.
In order to create enough power, many single PEM fuel cells need to be connected together with bipolar plates, but understanding how geometry and operating conditions affect the performance of single cells is the first step towards high performance designs. Researchers from the Centre for Fuel Cell Research at the University of South Carolina have been working with computational fluid dynamics software specialists, the CD adapco Group, to try to understand the behaviour of single PEM cells.
The University of South Carolina is a centre of excellence for fuel cell research and development. It was recently selected by the US National Science Foundation as the first Industry and University Cooperative Research Centre for Fuel Cells. The centre's mission is to help industry advance the technology of PEM fuel cells through its on-going research.
John Van Zee, a professor of chemical engineering at the centre, explained the thrust of their work: 'Our role is to assist industry, wherever possible, in moving toward the eventual commercialisation of fuel cell technology, and in educating well-qualified engineers and scientists. We employ a team of the world's fuel cell experts and use the very latest research methods to aid us in our work. CFD helps us with this mission by allowing us to understand our experimental data. Using accurate simulation software enables us to speed up our research and reduce the time we spend building physical models.
'USC has been using CFD to understand fuel cell behaviour since 1998,' continued Van Zee. 'With the internal combustion engine, methods exist to simulate the flow of multiple reactants. However, in fuel cells you have the extra electrochemical component, for which you need to add a charge conservation equation and transport of ions. The collaboration between our experts and those of CD adapco has helped to move our research on enormously.'
The critical aspect of a PEM fuel cell is that the protons require water in the membranes to facilitate transport to the cathode. Water moves from anode to cathode, with the protons, through a mechanism known as electro-osmotic drag. If the membrane is thin, and the concentration of water on the cathode is higher than on the anode, water can move through the membrane from cathode to anode by diffusion. Recent advances in polymers and composite membranes allow for very thin membranes with maximum structural integrity to be used in PEM fuel cells. The key to optimum performance is a membrane that is wet enough for maximum conductivity, but dry enough around the electrodes so that the transport of gas is not limited.
'This is an ideal problem for 3D CFD simulations to solve,' explained Van Zee. The group runs simulations using CD adapco's Star-CD software, 'to help us understand the distribution of water in the membrane and at the electrode surfaces. CFD provides predictions of water, temperature, and current density distributions inside the fuel cell. These predictions are difficult to achieve by purely experimental means, due to the complexity of the system and the small length scales involved. The computational results allow us to predict the local current density, species transport in a serpentine flow path, temperature distribution and water phase change. In other words, Star-CD can simulate a complete picture of the behaviour of the fuel cell, enabling us to begin the crucial design optimisation phase.'
The humidifier temperature of the fuel cells holds the key to design optimisation. Changes to the temperature affect the behaviour of the water within the cell, and consequently impact on the voltage produced. 'At lower humidifier temperatures, say between 55 and 65° centigrade, the membrane is not sufficiently wet and the current oscillates in a very unstable manner,' explained Van Zee. 'The oscillations decrease when the humidifier temperature increases and the maximum current occurs for humidifier temperatures of 85°C at the anode and 75°C cathode. When the temperatures are raised to 95/85°C, the current becomes very stable but the measured current is lower. We attribute the lower current at the higher temperatures to water accumulation on the cathode side of the part of the cell, called the membrane and electrode assembly (MEA). Higher humidity generates membrane flooding on the cathode side and also in the first channel of the anode. The flooding creates higher resistance for access of the oxygen and hydrogen to the catalyst on the membrane surface. It is these scenarios that CFD simulation can predict and ultimately help to resolve.'
The model of a fuel cell is complicated. Physical models are time consuming and costly to produce, but simulations need to be accurate if the research is to be successful. Fuel cells are made up of a gas channel flow-field plate and a thin membrane, sandwiched between anode and cathode diffusion layers. There are 20 serpentine passes in the flow path, increasing the complexity of the CFD model required.
The centre calculates numerical predictions so that comparisons can be drawn with the CFD results. Van Zee concluded: 'The agreement between the measured and predicted fluxes of water across the membrane at the different temperatures, indicates that the transport parameters and the water condensation phenomena are accurately described by the CFD simulations.
'We are satisfied that the use of 3D CFD simulations provides us with the necessary accuracy to understand the experimental behaviour within a laboratory scale PEM fuel cell. In turn, our confidence in these predictions allows us to scale up the results to full size models, simply by using large-scale geometry. What this means for us is finally being able to break the old paradigm of "build and test" and advance the development of this technology through computer-aided design, ultimately leading to the introduction of more environmentally friendly technologies as soon as possible.'
