"I would like to have a 3-D model of a locomotive on a workstation that I can take to the customer, modify during a live session to suit his requirements, and then transfer directly to the engineering office and from there to the shop floor," said Richard Dunteman, general director of component and locomotive-platform-development engineering.

In a marketplace dedicated to heavily engineered and customized complex products, Electro-MotiveÑwith its partners in the process, EDS Unigraphics in Maryland Heights, Mo., and EDS in Plano, Tex.--has targeted the product-development process itself as a critical competitive weapon that can ensure business success. Specifically, the company has embraced a completely digitally driven product-development process--what EDS Unigraphics calls virtual product development (VPD)--to move its concepts electronically through design, analysis, digital prototyping, and manufacture, and even to the customer's desktop for review, input, and approval.

The VPD strategy at Electro-Motive is more than a future vision. The company has already made significant progress toward its goal of creating a fully digital representation of its product.

"One of the major projects we're working on right now is what we call our platform locomotive," said Dunteman. "Almost all the critical parts have been entered into our 3-D database, and the digital model lets us make and analyze changes quickly and know that things are going to work right the first time."

Revolutionary Engine

The Electro-Motive Division, which originated the diesel electric locomotive, racks up more than $1 billion in revenues annually by serving a worldwide market with its line of passenger and freight locomotives. GM employs about 5,000 people at its engineering and manufacturing facility in LaGrange and at its Diesel Division assembly plant in London, Ontario.

Electro-Motive's drive to improve productivity and maintain quality for its extremely complex products led the company to a partnership with EDS Unigraphics to implement VPD. "There's a lot of design work involved in these complex products," Dunteman said. "We've got to stay away from paper because it really slows down the product-development effort. What we want to be able to do is design a 3-D model, maintain it online, interact with it online, and take it to manufacturing and then to the customer in that form."

Another goal is to eliminate physical mock-ups. "It's very time-consuming to prototype, to make pieces," Dunteman said. "The more work that can be done on-screen, the more these complicated designs can be verified in advance, the fewer the hands that will have to touch them, and the faster they will be produced."

One of the first projects Electro-Motive has been working on using the VPD strategy is GM's revolutionary 6,300-horsepower GM16V265H engine for the locomotive marketplace. "No one in this industry has ever delivered a brand-new engine from a clean sheet of paper in three years," Dunteman said. "We plan to achieve that through VPD. In fact, we have built the first engine in 18 months from the start of the program, and we are slightly ahead of our very aggressive schedule."

On the engine project, Electro-Motive is linking analysis, design optimization, and machine-tool and fixture programming to the solid modeling. "By doing the up-front analysis simultaneously with creating the model," Dunteman said, "we believe we will be able to verify the reliability of the engine in much less time than it has taken us in the past, even with more-stringent reliability goals."

In the future, the Electro-Motive Division's VPD strategy will allow customers to visualize almost all aspects of a locomotive design, down to the last button and dial in the cab, before it has been manufacture

Integrated Modeling and Analysis

The engine product-development process revolves around a team of engineering and manufacturing staff working simultaneously on product design, analysis, manufacturing, fixture design, and serviceability around the virtual model. In particular, Electro-Motive is linking analysis and modeling tightly throughout the project.

"There's a whole host of analysis work that has been done off the master model," Dunteman said. "The piston crown has been analyzed for mechanical and thermal properties in the 3-D database. The connecting rod has been fully analyzed as a model, along with all the bearings. We have performed very sophisticated stress and flow analysis on the turbocharger. We were able to perform some 30 iterations on the cylinder head design in order to make the part as a casting rather than a steel fabrication, and we analyzed and optimized the casting process before we ever poured a piece.

"As a result of our analysis, the first block casting and the first head castings were used to build prototype engines," he added. "That is unheard of in engine development, where engineers expect several rounds of casting process refinement before good pieces are poured. Basically, the engine has been fully stress-analyzed and optimized before it was ever built."

The Electro-Motive Division expects a significant compression in engine qualification time as a result. "We ran a three-year field test on a previous engine before it went into production, but we donÕt plan to do that here," Dunteman said. "We intend to qualify the reliability and durability of this engine in half the time, both in our own power-generating sites and in locomotives in the field."

The enormous horsepower of the engine will allow railroads to run half as many locomotives as before. That raises the stakes in terms of reliability. "With half as many units, each unit will have to be twice as reliable," Dunteman said. "Otherwise, the mission reliability of the train will suffers."

In addition to analysis, manufacturing is also tightly linked to modeling within the VPD implementation. "We have a pilot project where we used the 3-D model to generate the code for computer-numerical-control machining on a large, flexible machining center," Dunteman said. "We also transferred information electronically from the master model to make pattern equipment for the cylinder head and other components."

Checking Interference Online

The digitally defined locomotive product provides Electro-Motive with a flexibility it never had before. "We have configured our platform locomotive to accept three different engines, ranging in traction horsepower from 4,300 to 5,000 to 6,000, without having to make any drawing changes," Dunteman said. "This in essence allowed us to produce three models from one, at a savings of over 50 percent in design and drafting effort."

Electro-Motive is also doing its interference checking online. "A locomotive is a complicated structure--technologywise, it's as challenging as an airliner--with pipes and cables running through a dense packing of mechanical components," he added. "Drawing the structure in 2-D just guarantees that you're going to have all kinds of interference problems." Catching those interferences represents a major opportunity for the company. "The interferences we catch daily on the 3-D CAD screen have reduced scrap, rework, and lost time," Dunteman said.

Working in a "virtual" manner also lets Electro-Motive perform multiple operations on the 3-D locomotive model simultaneously. "We're doing the analysis work live, while we were procuring components," he said.

Electro-Motive has extended the VPD process beyond the physical boundaries of its headquarters in LaGrange. One of GM's engineering groups in London, working off the same 3-D database, conducts simultaneous design sessions live at both locations, transferring information back and forth over a high-speed telecommunications line.

Electro-Motive has also extended the digital link to vendors. "We have integrated key suppliers into the VPD loop, transferring data electronically to remotely located manufacturing shops for the machining and pattern work," Dunteman said.

Top-Down Modeling

The results are evident in another Electro-Motive project: export locomotive platforms that can be quickly customized for varying customer requirements in the international marketplace. "We are trying to design platforms that will cover a broad range of the world," Dunteman said. "We are doing it with small, dedicated teams and a top-down approach to the solid model."

That approach is the logical culmination of the VPD strategy. "It means the team basically defines the big picture--how long it's going to be, its general shape, and major components--and then refines it down," Dunteman said. "It's the opposite way from how a locomotive is generally developed, by designing all the components and then fitting them together."

The top-down approach will reduce the number of problems that are discovered deep in the product-development process. "The idea is to have the overall structure and know everything will fit before working out all the details at the lower level," Dunteman said. "If the team finds an interference with a pipe--or something else that just won't fit--18 months into a 24-month program, it will have a major problem and have to do a lot of scurrying around, rework, expediting, and field modification as a result."

There are obstacles to be faced, however, en route to virtual product development. "We're at the very leading edge of this technology, and we have made a deliberate decision that we were going to use the master model and VPD approach when we moved to the Unigraphics 3-D solid modeling system," Dunteman said.

"One of the first things we realized when we began was that we have very good skills here in the drafting area," he added. "But spreading the concept that we're really working on an electronic database--and need the discipline to follow the rules so we can generate parametric models that we can modify through various iterations of the model, rather than just using the system as a drawing tool--has been a major challenge."

Another challenge is the specialist skill set of Electro-Motive engineers. The division's engineers typically are assigned specialized hardware and/or subsystems and work at "arm's length" from drafting. "We have a vision of nonspecialist engineers working in teams to create the 3-D model as well as do all the analysis on that model," Dunteman said. "That's a significant value-added approach for the future, but we're not sure how quickly we can really get there."

He said there's also a significant challenge on the manufacturing side. "Manufacturing engineering is just beginning to be able to fully utilize the data without re-creating their own drawings or layouts. This is, again, an enterprise-skill-level and systems issue."

Another issue involves assembly modeling. "In this locomotive, we have really pushed the envelope of the workstation hardware and the software because of the size and complexity of our models," Dunteman added.

Dunteman emphasized the importance of Electro-Motive's partnership with EDS Unigraphics and EDS to help get over the inevitable bumps in the road. "Both EDS Unigraphics and EDS have invested heavily in this effort," he said. "With such a wide-reaching, enterprise-level product-development implementation, a company needs to have someone who understands and shares its vision, someone who can help it move forward toward realizing that vision through things like assisting in determining training needs and then supplying that training expertise.

There are organizational issues," Dunteman said, such as "how does an organization address the architecture of the master database on which everything is going to be based when people aren't used to thinking in those terms? There are system issues: how to avoid manually feeding data, in order to have things work automatically.

"We've had a lot of growing pains together, but we have gotten things done that we couldn't have gotten done, and we've gotten them done faster, because of the partnership."

EDS and EDS Unigraphics agree that Electro-Motive has moved quickly toward implementing the virtual product development vision. "When you consider how complex a locomotive is and the large number of parts that are involved, what Electro-Motive and Diesel have done is a remarkable story," said John Ferreri, an EDS engineering-services manager on-site at LaGrange, providing a wide range of information technology services to Electro-Motive.

"We are extremely impressed at the level where they are today, and at how far they have come in the last three years since they began migrating from a 2-D environment," he added. Ferreri credited a lot of that to a strategic perspective at the company.

"The Electro-Motive and Diesel divisions have developed their vision at the strategic level, and that is what has made their success possible," he said. "A core group of managers and users are now turning that vision into reality."

Ready for revenue service in 1997, the GM16V265H engine delivers 6,300 horsepower at 1,000 rpm with ac traction, and can replace two 3,000-horsepower unit

VPD and the Customer

The success of the VPD strategy so far and its projected future benefits have justified Electro-Motive's long-term commitment. In the future, the top-down model will enable Electro-Motive to take the VPD modeling paradigm a step further, outside the boundaries of the enterprise, by bringing the digital, virtual representation of the product to the customer. "It will take significant advances in portable computing--over the next two or three years, at least--to be able to take these models out to the customer," Dunteman said.

Electro-Motive has accomplished that to a limited extent in two projects--a new North American passenger model and the export platform locomotive--by bringing customers to its facility to see the 3-D model. "We developed the 3-D model and confirmed the styling of the locomotive live with the customer, during the iterative bid process," Dunteman said.

"A locomotive is a very complicated product, with a lot of customer extras designed into it," he added. "It made an enormous difference to the project that the customer could see what the unit was going to look like and that it would meet all major contract requirements beforehand."

Electro-Motive's goal, once it fully develops and refines its VPD capability, is to let that accumulated knowledge drive future projects so that it becomes a true "learning" organization. "We're looking for our virtual, electronic system to capture the knowledge, capture all the history of locomotive design, and then "guide" engineering generalists working in teams on the design and modification of future locomotives through that knowledge base and visualization capability," Dunteman said.

Toward that goal, the division has created a challenge for itself. "We want to create the first fully digital process and product to come out of the General Motors engineering stables," Dunteman said.

home | features | weekly news | marketplace | departments | about ME | back issues | ASME | site search

© 1996 by The American Society of Mechanical Engineers