Hydrogen fuel cell technology for automobiles holds a great deal of promise-pollution-free vehicles, a nearly unlimited fuel source and an alternative to dependence on foreign oil. But designers of fuel cell-powered cars face great challenges-everything from designing a new fuel tank to planning the gradual conversion of the fueling infrastructure from gasoline to hydrogen. One of the largest questions is how to get hydrogen to the fuel cell in the first place.

"There are already gas stations throughout the world," says Gregory Ariff, a senior engineer at Directed Technologies Inc. (Arlington, VA). "A sensible place to start was finding out what it would cost to provide those stations with the ability to offer hydrogen." The new hardware required for these stations includes a reforming and purification system to create hydrogen from natural gas, a compressor to pressurize the gas, storage vessels and a pump to dispense hydrogen gas into cars.

Directed Technologies, a technical consulting firm, has studied the potential of hydrogen as an automotive fuel since the early 1990s. In 2002, the firm conducted a cost comparison of two natural gas reformer technologies-autothermal reforming (ATR) and steam methane reforming (SMR)-for the Department of Energy (DOE).

To create hydrogen from natural gas, both methods expose natural gas to a catalyst, usually nickel, at high temperature and pressure. The autothermal reformer burns a portion of the natural gas within the reforming vessel to provide heat for the reaction. The steam methane reformer uses an external source of hot gas to heat tubes containing a mixture of steam and methane.

The catalytic reaction in the autothermal reformer takes place in one large vessel. By contrast, the reaction in the steam reformer takes place in more than a hundred metal tubes mounted in parallel inside a heated vessel. A wide range of difficult-to-process metals-from stainless steel to expensive high-temperature alloys, such as Inconel or Haynes 556-is used in building the equipment for both ATR and SMR.

"The autothermal reformer was known to be a lower capital-cost system, but the steam methane reformer was more efficient," Ariff says. "The question was, at a fueling station scale, which technology would produce lower-cost hydrogen for the consumer?"

"An accurate comparison of ATR and SMR could not be performed by simple rules-of-thumb," says Ariff. "We needed a careful methodology for this study, and we needed to apply a rigorous, metric costing approach that would enable us to evaluate two competing technologies." Directed Technologies used Design for Manufacture and Assembly (DFMA) software from Boothroyd Dewhurst Inc. (Wakefield, RI) for the cost comparison. Analyses derived from the software ultimately showed that, even though it is more costly, a well-designed steam methane reformer could produce hydrogen more cost-effectively than an autothermal reformer. But the path to this conclusion involved complex analysis.

DFMA software consists of interlinked Design for Assembly (DFA) and Design for Manufacture (DFM) software. DFA software guides engineers to evaluate the functional purpose of each assembly component in a conceptual design and to rate each component on its ease of orientation and assembly. Finally, the software generates estimates of total assembly time and costs. DFM software identifies the major cost drivers associated with manufacturing and finishing the parts. Concurrent costing tools aid engineers in choosing the most cost-effective shape-forming process for a part and to consider how individual part features might be modified to optimize manufacturing costs.

Ariff says that DFMA gave the company's engineering teams a common objective language, for quick and accurate decision-making during the design process. "We have mechanical engineers, chemical engineers and some with backgrounds in aerospace, so DFMA is an excellent tool for providing a framework that we can all converse in," he says.

At the beginning of each reformer study, chemical engineers used simulation software to create an ideal step-by-step model of the hydrogen extraction process. This established a baseline for assessing the efficiency of subsequent design iterations. Then they created a rough bill of materials for the initial reformer design and set a hypothetical production volume of 250 units.

"Once we had a bill of materials, Ariff says, "we used DFMA to compare potential manufacturing and machining methods for parts. We also established assembly times and labor costs for subassemblies. From that point, it was a matter of identifying the best possible combination of design and assembly processes."

For the steam methane reformer the shell-and-tube vessel, in which hydrogen extraction takes place, accounted for a large part of the assembly cost. "Materials choices build in costs you may not expect," Ariff says. "For instance, high-temperature resistant alloys can be five to nine times more expensive than stainless steel. But they also involve additional manufacturing costs because they are more difficult to work with." The number of assembly tasks also made the shell-and-tube vessel expensive. Each of the reformer tubes is welded at both ends to a tube sheet, and there are from 134 to 180 tubes, depending on operating pressure.

"Using DFMA to review small changes in dimensions and different materials was crucial," Ariff says. "It helped us to resolve questions of whether to use lesser amounts of expensive, high-temperature materials or greater amounts of less-expensive conventional alloys for particular components." In the end, a design change as simple as selecting orbital welding for the tube ends helped reduce assembly costs significantly, even for the high-temperature alloys.

The final report delivered to DOE justified the careful analysis. The steam methane reformer, which produced hydrogen more efficiently but was more costly to manufacture, could be designed as a cost-effective choice over the autothermal reformer. The per-unit cost of a steam methane reformer came to $123,545. And it operates at ten atmospheres

While this was still $20,000 more expensive than a comparable autothermal reformer design, the steam methane reformer more than made up the difference in operating efficiency. "In the end, the capital cost savings for an ATR design couldn't offset its high operating expenses compared to the SMR," Ariff says. "Without DFMA, we might not have established that."

"There is still a lot of work to do before hydrogen is as cost-effective as fossil fuels," Ariff says, "But we've come a long way. We will continue to use DFMA and other design analysis tools to simplify and improve existing technology."

For more information on technology consulting, call 703-243-3383 visit www.directedtechnologies.com or Reply 1.

For more information on DFMA software, call 401-272-1510 or visit www.dfma.com.