"No Climbing," Feature Article, February 2008

no climbing

Real and virtual testing make a safe impact on a light rail design.

This article was prepared by staff writers in collaboration with outside contributors.

When Stadler Rail Group in Bussnang, Switzerland, received an order from the Netherlands for 43 of the latest generation of an articulated light rail car called GTW, the company faced a new challenge: The cars had to meet crashworthiness standards that the country had adopted in advance of their approval by the European Community. Among the requirements was that the cars provide passenger zone protection during a 36 km/h (22.4 mph) front-end collision between two units with a vertical offset of up to 40 millimeters.

Two developments drove the new requirement. First, head-on impacts could easily include a small offset if two otherwise identical cars had differing amounts of wheel wear or braking inclination. A second reason was more urgent: A recent numerical simulation of an offset collision indicated that the previous design of a crash module (a safety device on the front of the train car) might not prevent damage to the passenger zone of the vehicles during such an impact.

"Numerical simulation suggested that the crash module could undergo global shear deformation and fail at the fixation point, falling off the front structure," said Alois Starlinger, head of structural analysis, testing, and certification at Stadler. "In such a shear-mode failure, the module would not absorb any significant energy." In a worst-case scenario, one car could climb over another, severely damaging the passenger zones.

Stadler produces about 700 light and commuter rail vehicles per year. All of its products must meet stringent requirements governing safety equipment, strength of rail vehicles, and above all, protection of passengers and crew from the force of impacts.

Head-to-head meeting: A finite-element model created in Abaqus software simulates two GTW light rail cars colliding head-on.

To satisfy the new safety requirement—which is scheduled to become the standard throughout Europe in 2008—Stadler Rail designed a new crash module with an anti-climb feature. Engineers validated the module design through a combination of dynamic physical testing and simulations in Abaqus finite element analysis software from the Simulia brand of Dassault Systèmes.

The crash module is a slightly tapered rectangular tube that is 350 mm high and wide at the front, 768 mm long, and 400 mm high and wide at the rear, where it is welded to an end plate bolted onto the crash wall of the train unit. Partitions divide the module into chambers that provide stability to counter eccentric forces. On the front of the module are five horizontally aligned teeth, 70 mm apart with a depth of 40 mm, which are designed to engage the teeth of a similar module on an oncoming rail car and prevent climb.

Once the teeth have engaged, the rest of the crash module is optimized for controlled structural deformation from the front to the back. Targeted slots on the sides create intentional weak points that initiate buckling to absorb energy. In developing the design, engineers built on lessons learned while producing crash modules for previous generations of the GTW car.

For the new design, the engineers selected an aluminum alloy, AW 5754. This alloy combines low yield strength with good plastic forming characteristics, enabling it to undergo large deformations without fracture. An important engineering goal was to create modules that could absorb up to 900 kilojoules of a crash impact while decelerating the train unit at a rate of 5 g or less as far as was practicable.

To capture the material behavior of the module, Stadler extracted information from its own materials database, compiled from physical tests. Engineers incorporated the data into a digital model. They calibrated the metals simulation still further by extracting aluminum samples from a production model of a crash module and testing the samples to create stress-strain curves. By comparing these curves to results generated by the computer simulations, the engineers were able to fine-tune the behavior of the FEA analysis so that it closely matched the real-world characteristics of the aluminum alloy in a crash module.

Crash test: Stadler's crash modules demonstrate after a physical test that they can absorb sufficient energy to protect passengers.

Now the engineers were ready to build a model of the crash module and analyze its behavior on impact. Simulation of the head-on offset impact followed a number of parameters. Collision masses of the train units were estimated at 100,000 kg each. Combined closing speed was 36 km/h. Maximum energy to be absorbed by crash components of both train units was 2,230 kJ, and maximum energy to be absorbed by a single train unit was half that, or 1,115 kJ.

Because of the complexity of the analysis, engineers ran nonlinear dynamic simulations with another Simulia product called Abaqus/Explicit to observe the elastic-plastic behavior of the metal, measure progressive damage and failure of welding, analyze the large deformations of the module, and model contact and friction.

The finite element model contained 450,000 elements, and the dynamic simulations captured a period of 0.4 second broken down into 200,000 "snapshots." The engineers ran the software on an SGI Altix 350 with four Itanium processors with activated parallel processing.

Train units were modeled in 3-D with running bogies (wheel, axle, and frame assemblies) and suspension characteristics in order to capture any lift-off of the wheels and axles on impact. Contact conditions were defined between the wheels and the rails, as well as between the bogies and the body of the car. Forces applied on impact by attached articulated units were modeled axially with 1-D elements and mass elements.

Safe Arrival

Abaqus simulation results correlate very well with physical dynamic tests. The anti-climb teeth prevent either train unit from moving over the other, and the module body undergoes controlled deformation to absorb 1.1 megajoules. Aluminum buckling decelerates the train unit at an average of 1.25 g.

"Our goal was to achieve an overall compressive strength for the train unit to 1,500 kilonewtons, without undergoing any yield and deformation in the passenger structure," Starlinger said. "In fact, our crashworthiness engineering improved the compressive strength to about 3,600 kN, with only small amounts of plastic deformation in the passenger zone. And we proved out the anti-climb device against offsets as high as 80 mm."

According to Starlinger, the crash module went into production eight months after the contract was signed with the customer, Arriva Nederland, the Dutch unit of a European transportation company. "The whole GTW Arriva went into operation 10 months later, which is probably a record for starting a design from scratch in passenger train service," Starlinger said. Six other cars of the same model have been ordered by Capital Metro in Austin, Texas.

The agency has received two of the cars already and will begin light rail service in the fall.

Stadler plans to build on its experience and continue making each new train design safer than the last. Starlinger sees finite-element software as an important part of that process. "In its own way," Stadler said "FEA is now as essential to ensuring train safety as brakes are."