Developing a Better Bus

It’s called the Altoona Bus Research and Testing Center, or ABTC among the regulars. The facility is under the auspices of the Vehicle Systems & Safety Program of the Pennsylvania Transportation Institute at Penn State. At the ABTC, buses are undergoing a variety of tests to determine their viability for municipal transport. Safety. Structural integrity and durability. Reliability. Maintainability. Fuel economy. These and other tests are carried out at the ABTC.

In effect, when it comes to buses, make it in Altoona, or you won’t make it.

And it is there that what started out as a piece of marketing material for a topology optimization software package may end up in the not-too-distant future. That’s right: what started out as a piece of marketing material for a software package is becoming a bona-fide, full-size city transit bus.

As Timothy R. Smith, director, Design Engineering, Altair Product Design (altairproductdesign.com), recalls, in 1994, Altair Engineering (altair.com) originally released OptiStruct, a finite-element based technology that is within the company’s larger HyperWorks suite. Essentially, OptiStruct is deployed early in the design process for a product. Parameters about the package space, design intent, and manufacturing process are entered into the system. The OptiStruct algorithms then work to provide a design that is both lightweight and structurally sound. It works at both the component level and at the systems level. It takes away material where it isn’t needed but helps assure that what’s left meets the functional requirements (e.g., strength, stiffness).

As the company is born and based in metro Detroit, it isn’t surprising that Altair has been doing considerable work with the auto industry. But (1) those firms that were using OptiStruct early on were using it for future programs, and weren’t about to let Altair talk about it and (2) Altair was thinking about reaching people in other industries. So in 1999, they started thinking about creating some marketing collateral for OptiStruct, and they realized that they couldn’t use examples from existing customers since that information was confidential, and besides, most of that was automobile-related and they were looking for those additional markets.

Which led them to think about buses. “So we said, ‘Let’s create a bus structure. Not just pictures, but the packaging.’ So we used our software and some new methods to map out the structural space for a transit bus.”

It was there that they deployed the new software. “Our philosophy with topology optimization is to only define the areas of the system that cannot move.” So in the case of a bus things like the overall dimensions—length, width, height, wheelbase—are established, as is the passenger capacity. Then things like the axle location, wheels and tires, engine—things that cannot move—are established.

So imagine, in effect, a solid rectangular block of material (the shape of a bus) from which the interior is removed (to accommodate the passengers and the engine and the axles and the like). The rules and constraints are applied, and the software generates the optimal way to layout the structure of—in this case—a bus.

Smith points to a 1:10 model that they were to build in July 2009 for the purpose of explaining to potential suppliers what they were looking for. “Nothing is very orthogonal when you look at it in terms of an existing bus structure, where everything is very perpendicular and parallel. The structure isn’t symmetrical”—after all, if you think of the front door of the bus and the rear door, the front door is close to the front axle and the rear door is more amidships—“so based on the asymmetry of the environment, this is the structure.”

The structure of a conventional bus is the sort of thing that you could design with a T-square and a right triangle (to use old-school drafting tools). Sure, the skin that wraps it may have curves and bulges, but the underlying elements are rather uniform. But the bus developed at Altair is, as Smith describes it, “very organic.” What the software did was create a structural model that provides the optimal load paths. Consequently, there are X-shapes in different sections and each section is different from the other. As they ran iterations (looking at bending constraints, then, say, torsional constraints, then combining them) they let the software find the location for the windows, which contributes to the variation in design throughout the structure.

Smith admits that the model resulted in as many as 145 different profiles for each of the structural members. And the normal state of manufacture in the bus industry is to use standard plates, tubes, C-sections, and the like.

Realize that what they were working toward was not an avant-garde design for a bus, but for a bus that would have as much as 25% of its mass removed through the topology optimization.

But product development is more than just creating design. So Smith and his colleagues went out on the road with their digital model and visited bus manufacturers in North America. And what they discovered surprised them. Whereas the Altair designers and engineers were thinking about a holistic design, the standard approach is essentially to create buses with off-the-shelf components, with the componentry (e.g., front axle, rear axle, driver’s seat, air conditioning system) being specified by the transit authority that will be receiving the bus. The specs go right down to the particular vendors and associated serial numbers. Not only are the bus manufacturers “working on the thinnest of thin margins,” but, Smith goes on to explain, the thought at those companies is “Why create an optimal bus when the municipality will change it?”

And here were these guys from Detroit with a structural shape that looked, comparatively speaking, like a Frank Gehry building compared with a suburban ranch house.

During their visits they also learned that the federal government pays some 80% of the sticker for transit buses with the local transit authorities picking up the remaining 20%.

So they decided that if they were going to create a whole new specification for transit bus design, they needed to go to Washington, which they did in 2005. They managed to receive federal appropriations from the previous and current administrations, given an abiding interest in things like mass transit and buses.

“As engineers, we would continue to refine things in CAD,” Smith said. But in March 2009, James R. Scapa, chairman of Altair, told them they had to “get physical and to get physical fast.” He wanted the development to appeal not only to politicians but to venture capitalists, so they had to create more than the CAD models—and more than a 1:10-scale model. They were going to build a bus. And they had a timetable: a demonstration vehicle that would be under power by September 2010.

Remember: Altair is a design and engineering company, not a vehicle manufacturer. Yet they moved forward on what has become BUSolutions LCO-140. Yes, a bus.

Smith says that they’d developed an optimal diesel driveline for the bus, one with shift points that would allow the engine to run in its sweet spot. Diesel engines, of course, are common in city buses. But again, Scapa told them in April 2009 that they had to do something different. He pointed out that there was an increasing interest in alterative powertrains, a great deal of activity in electric vehicles and hybrid electrics, but that he wanted them to do something different.

So they began looking for powertrain alternatives, and investigated the series hybrid hydraulic. That’s right: a hybrid that is based on fluid power. “A bus is the exact kind of vehicle that a hydraulic hybrid system is designed for: one that starts and stops a lot,” Smith says. That is, when the vehicle brakes, there is compression of the fluid into a high-pressure accumulator, and about 70% of the energy is recaptured, which is considerably more than the amount of power captured during regenerative braking in battery-based systems. This fluid—on the order of 5,000 to 6,000 psi—is then used to power the vehicle to its next stop. They checked out a number of developmental systems, but opted to work with Parker Hannifin (parker.com), which has a mature system (it introduced it in 2006). Given that they were facing the September 2010 target, they figured that it was better to work with something that was developed rather than something that was still being developed. (It should be noted that the hybrid also uses a 280-hp Cummins ISB series diesel.)

When it came to other elements of the bus-in-becoming—from the tires to the seats—they also opted to work with vendors with commercial credibility within the bus world (e.g., Michelin). It’s also worth noting that the program had significant partnerships with companies including ArvinMeritor and Carrier, companies familiar to those at municipal bus depots.

So there they were. Given the mandate to build a bus. Told to switch from a diesel to something non-traditional. And there was one more thing.

The bus-as-developed was going to be made with stainless steel. The rationale was that stainless would address corrosion issues, given that a city bus is generally on the road for about 20 years.

Smith says that they talked to people in the steel industry about the availability of sheet and tubes, and learned that both were accessible. But then there was a big BUT. In the optimized structural design, there are 26 unique profiles. The Altair engineers were told that they could get custom tubes rolled for them. No problem. All they had to do was buy 5,000 lb. of material. Which was fine by them, because 5,000 lb. was the approximate weight of the structure. BUT that wasn’t what the steel vendors meant: They meant 5,000 lb. for each of the 26 profiles. That meant 65 tons of steel for a 2.5-ton need.

So they asked about buying the tooling for rolling the profiles. And were told that it would cost them $100,000 per tool.

Clearly, a rethink was in order. One idea was to build the demonstration bus out of common steel and to paint it to resemble stainless. “Then when we showed it to people, we could wave our arms and say that when we build the real thing, it will be stainless,” Smith recalls. But then Scapa suggested that they ditch the steel and go with an aluminum structure.

Imagine the consternation with that.

However, this brings us back to the topology optimization. According to Smith, “The beauty of topology optimization is that whether your structure is made out of bubble gum or cast iron, the load paths will be the same.” So because they had the analytic model of the stainless-steel structure, it was a matter of scaling up the thickness of the aluminum pieces by a factor of three to account for the difference in density between the two materials. While they didn’t take out any mass through the material switch, they were able to (1) make significant cost savings, as in a tool for aluminum extrusion coming in at $3,000, or $97,000 less and (2) create shapes that would allow them to simplify the manufacturing: in some cases it was a matter of creating interfaces for mating pieces that would eliminate compound miter cuts; in others it was using the extrusion to provide an exposed surface, thereby foregoing the use of cladding.

One issue that had to be addressed was the effect of welding on aluminum, as welding degrades the mechanical properties of the material (which explains the extensive use of mechanical fasteners and adhesives on aluminum structures). They considered the possibility of building the bus with no welding, but determined that given the way it was designed, it was going to be necessary. So they went back to analysis software to determine the effects of welding. “The endurance limits for 6061 aluminum is about 90 MPa. So we set the parameter at 45 to see whether there were any points where that would be a problem. We found that because the structure is uniformly stressed, there weren’t huge areas that had to be addressed. There was the need for some gussets or reinforcements, but no big issues,” Smith says.

There are some areas where there are fasteners used, but only if these are components that may need replacement, such as brackets. And there is some steel used, mainly where the suspension attaches to the vehicle.

The bus as developed by Altair is said to be 10% lighter than competitive products. Annual cost-of-ownership is calculated to be $25,000 less. The fuel economy of the hybrid system is two times better than an average diesel bus (7.3 mpg vs. 3.5 mpg). And there are additional benefits.

But Smith admits, “We’ve taken the opportunity and the liberty to rewrite the bus specifications.”

Still, it is a long way from a piece of marketing material as was originally envisioned. And, yes, although their present funding takes them only to the completion of the demonstrator, they’re hoping that whether it’s government officials or venture capitalists, they’ll get the money they need to take it to Altoona.