Designing the Future with DFMA
Design analysis and costing software help evaluate designs for automotive hydrogen fuel cell systems.
How do you create and cost a detailed, realistic design using components and processes that don’t yet exist? That was the challenge facing Directed Technologies, Inc. (DTI), Arlington, Va., when the Department of Energy (DOE) contracted with them to design automotive hydrogen fuel cell systems—as they might be built in 2006, 2010, and 2015.
DTI is a technical consulting firm that evaluates new and emerging technologies for the government and industry. The company has researched hydrogen infrastructures and applications since the early 1990s. Previous projects have included reformer technology and cost-effective fuel cells.
In March of 2006, DTI began a two-year study for the DOE as part of an ongoing effort to explore alternative fuels. The goal of the study was to estimate production costs for direct hydrogen proton exchange membrane (PEM) fuel cell systems for use on board automobiles. The government also set technical targets for each of the 2006, 2010, and 2015 versions to meet. “The targets didn’t directly drive our analyses,” says Jeff Kalinoski, an engineer at DTI. “Instead, we used our work to evaluate how realistic the targets were.”
To create the three systems, the engineers at DTI employed process modeling software, computer-aided design tools, and Design for Manufacture and Assembly (DFMA) software from Boothroyd Dewhurst, Inc., Wakefield, R.I.
Fuel cells and the future
Engineers on the DTI team spanned a wide range of expertise, including chemical, aerospace, and mechanical engineering. Communication between disciplines was made easier by a carefully laid-out work process. “Our design and analysis methodology was rigorous and time-tested,” Kalinoski comments.
Work for modeling the fuel cell system fell into four stages:
1. Research. This involved reviewing literature, looking up patents, and interviewing and talking with engineers and experts. During this phase, the team examined hundreds of assumptions behind the construction and the materials selection of a complete hydrogen power system. The research phase will continue throughout the life of the project.
2. System modeling. The team modeled the 2006 system in a process-modeling package called HYSYS. The software enabled DTI to establish a detailed model based on technical parameters and to establish specifications and a bill of materials for the system.
The engineers then extrapolated technology and process improvements from the same model to come up with bills of materials for the 2010 and 2015 systems.
3. Component designing. The next step was to use the system model and bill of materials as a basis for modeling all components in 3-D, with enough detail for accurate costing. “Since we were costing a design, not building it, we didn’t have to address every single consideration,” Kalinoski points out. “We designed pipes to handle flow, but we didn’t worry about whether flow was turbulent or laminar. Knowing the diameter, thickness, shape, and materials was enough.”
4. DFMA. In a final step, the design team applied DFMA software and methodology to cost their system thoroughly and to select optimum designs and processes for manufacturing individual components.
DFMA and design costing
DFMA software is based on two interlocking approaches: design for assembly (DFA) and design for manufacturing (DFM). DFA guides engineers to evaluate the functional purpose of each assembly component in a conceptual design, helping them to simplify the design. DFM identifies and calculates the cost drivers associated with manufacturing and finishing parts in alternative processes.
In DFM, concurrent costing tools enable engineers to determine the most cost-effective process and, in some cases, to alter design features and further reduce costs. The software contains an extensive library of data for varied materials and processes, including secondary machining. A key benefit of DFM software is the quick generation of an initial cost estimate at any stage of design in just a few simple steps.
A DFM analysis was particularly important in choosing a manufacturing process for the bipolar plates of the fuel cells. “We had to cost out the manufacturing process carefully,” Kalinoski says, “since there are nearly 800 plates in each hydrogen system.” A previous study that also used DFMA software had concluded that injection-molded composite plates were the best process. For the new study, the team re-examined injection molding closely with the software. “We looked at the process in detail, re-designing the plates, changing the number of platens in the mold, and generating costs for different batch sizes,” Kalinoski says.
This time around, stamped steel plates seemed an attractive possibility. Steel is ideal for stamping because of its strength, high strain rate, and low cost. The DTI team did press-force calculations based on the geometry of the plate, and they used the results to size presses for different stamping methods. Once the engineers had selected press types, they used DFMA software to compare multi-stage, compound, turret, and progressive dies at multiple batch sizes and annual rates.
The results suggested that a four-stage progressive die was the most cost-effective option at all production levels considered. There are two shearing stages for intake and exhaust manifolds, a shallow forming stage for flow paths, and a part-off stage to cut off the finished part. The team also performed similar calculations based on their own costing methodologies, then consulted with two different outside manufacturing companies to ensure that their results were accurate.
The choice of process guided alterations in the final part design. The plate now has straight, parallel sides and is stamped from a steel roll the exact width of the plate so that the final die stage cuts it with zero wastage. The progressive die can produce up to 80 plates per minute.
The design and process analyses yielded some surprises. For instance, the proton exchange membrane required layers of catalyst on either side of it. The team explored coating the membrane with catalyst on both sides at once for manufacturing efficiency at high-volume production. Kalinoski points out, “When it came to reviewing low-end production, we had a $750,000 coating machine that we were only running at 12% capacity. But DFMA calculations showed us that even at that rate, it was still more efficient than the other processes we examined.”
The road ahead
The study is two-thirds complete and is going well. “We came close to the cost targets the DOE specified,” Kalinoski says, “though there is more to be done.” The DOE set a 2006 fuel cell stack cost target of $70/kW, and DTI estimates a cost of $67/kW. So far, the 2015 stack stands at $25/kW. The DOE was hoping for $15.
Between now and February 2008, the DTI team will continue tweaking its designs to see if there are any other costs they can remove from their future fuel cell stacks and from the overall system. For example, they will examine alternate manufacturing methods for the proton exchange membrane and the gas diffusion layers, both significant cost elements. Also, they will explore replacing the pure platinum catalyst with a less expensive platinum compound.
DTI plans to perform annual design and costing updates to their hydrogen system designs, possibly through 2011. “I’m optimistic that we can continue to improve on our system designs and manufacturing techniques each time we update the cost estimates,” says Kalinoski. “DFMA will remain an integral part of that work.”
--Jeff Kalinoski, Engineer
Directed Technologies, Inc.
Resources:
Boothroyd Dewhurst, Inc., Wakefield, R.I., 401-783-5840, www.dfma.com
DTI is a technical consulting firm that evaluates new and emerging technologies for the government and industry. The company has researched hydrogen infrastructures and applications since the early 1990s. Previous projects have included reformer technology and cost-effective fuel cells.
In March of 2006, DTI began a two-year study for the DOE as part of an ongoing effort to explore alternative fuels. The goal of the study was to estimate production costs for direct hydrogen proton exchange membrane (PEM) fuel cell systems for use on board automobiles. The government also set technical targets for each of the 2006, 2010, and 2015 versions to meet. “The targets didn’t directly drive our analyses,” says Jeff Kalinoski, an engineer at DTI. “Instead, we used our work to evaluate how realistic the targets were.”
To create the three systems, the engineers at DTI employed process modeling software, computer-aided design tools, and Design for Manufacture and Assembly (DFMA) software from Boothroyd Dewhurst, Inc., Wakefield, R.I.
|
Engineers on the DTI team spanned a wide range of expertise, including chemical, aerospace, and mechanical engineering. Communication between disciplines was made easier by a carefully laid-out work process. “Our design and analysis methodology was rigorous and time-tested,” Kalinoski comments.
Work for modeling the fuel cell system fell into four stages:
1. Research. This involved reviewing literature, looking up patents, and interviewing and talking with engineers and experts. During this phase, the team examined hundreds of assumptions behind the construction and the materials selection of a complete hydrogen power system. The research phase will continue throughout the life of the project.
2. System modeling. The team modeled the 2006 system in a process-modeling package called HYSYS. The software enabled DTI to establish a detailed model based on technical parameters and to establish specifications and a bill of materials for the system.
The engineers then extrapolated technology and process improvements from the same model to come up with bills of materials for the 2010 and 2015 systems.
3. Component designing. The next step was to use the system model and bill of materials as a basis for modeling all components in 3-D, with enough detail for accurate costing. “Since we were costing a design, not building it, we didn’t have to address every single consideration,” Kalinoski points out. “We designed pipes to handle flow, but we didn’t worry about whether flow was turbulent or laminar. Knowing the diameter, thickness, shape, and materials was enough.”
4. DFMA. In a final step, the design team applied DFMA software and methodology to cost their system thoroughly and to select optimum designs and processes for manufacturing individual components.
DFMA and design costing
DFMA software is based on two interlocking approaches: design for assembly (DFA) and design for manufacturing (DFM). DFA guides engineers to evaluate the functional purpose of each assembly component in a conceptual design, helping them to simplify the design. DFM identifies and calculates the cost drivers associated with manufacturing and finishing parts in alternative processes.
In DFM, concurrent costing tools enable engineers to determine the most cost-effective process and, in some cases, to alter design features and further reduce costs. The software contains an extensive library of data for varied materials and processes, including secondary machining. A key benefit of DFM software is the quick generation of an initial cost estimate at any stage of design in just a few simple steps.
A DFM analysis was particularly important in choosing a manufacturing process for the bipolar plates of the fuel cells. “We had to cost out the manufacturing process carefully,” Kalinoski says, “since there are nearly 800 plates in each hydrogen system.” A previous study that also used DFMA software had concluded that injection-molded composite plates were the best process. For the new study, the team re-examined injection molding closely with the software. “We looked at the process in detail, re-designing the plates, changing the number of platens in the mold, and generating costs for different batch sizes,” Kalinoski says.
This time around, stamped steel plates seemed an attractive possibility. Steel is ideal for stamping because of its strength, high strain rate, and low cost. The DTI team did press-force calculations based on the geometry of the plate, and they used the results to size presses for different stamping methods. Once the engineers had selected press types, they used DFMA software to compare multi-stage, compound, turret, and progressive dies at multiple batch sizes and annual rates.
The results suggested that a four-stage progressive die was the most cost-effective option at all production levels considered. There are two shearing stages for intake and exhaust manifolds, a shallow forming stage for flow paths, and a part-off stage to cut off the finished part. The team also performed similar calculations based on their own costing methodologies, then consulted with two different outside manufacturing companies to ensure that their results were accurate.
The choice of process guided alterations in the final part design. The plate now has straight, parallel sides and is stamped from a steel roll the exact width of the plate so that the final die stage cuts it with zero wastage. The progressive die can produce up to 80 plates per minute.
The design and process analyses yielded some surprises. For instance, the proton exchange membrane required layers of catalyst on either side of it. The team explored coating the membrane with catalyst on both sides at once for manufacturing efficiency at high-volume production. Kalinoski points out, “When it came to reviewing low-end production, we had a $750,000 coating machine that we were only running at 12% capacity. But DFMA calculations showed us that even at that rate, it was still more efficient than the other processes we examined.”
The road ahead
The study is two-thirds complete and is going well. “We came close to the cost targets the DOE specified,” Kalinoski says, “though there is more to be done.” The DOE set a 2006 fuel cell stack cost target of $70/kW, and DTI estimates a cost of $67/kW. So far, the 2015 stack stands at $25/kW. The DOE was hoping for $15.
Between now and February 2008, the DTI team will continue tweaking its designs to see if there are any other costs they can remove from their future fuel cell stacks and from the overall system. For example, they will examine alternate manufacturing methods for the proton exchange membrane and the gas diffusion layers, both significant cost elements. Also, they will explore replacing the pure platinum catalyst with a less expensive platinum compound.
DTI plans to perform annual design and costing updates to their hydrogen system designs, possibly through 2011. “I’m optimistic that we can continue to improve on our system designs and manufacturing techniques each time we update the cost estimates,” says Kalinoski. “DFMA will remain an integral part of that work.”
--Jeff Kalinoski, Engineer
Directed Technologies, Inc.
Resources:
Boothroyd Dewhurst, Inc., Wakefield, R.I., 401-783-5840, www.dfma.com
Use of this website is subject to its terms of use.
Privacy Policy
