The IMPETUS® origin story - how a systematic simulation discrepancy led to a new dynamic test method
A material property is only as good as the test that produced it. This sounds self-evident. In practice, it is routinely ignored, because the test method itself is rarely questioned. It is assumed to be correct by default, simply because it is standard.
This assumption failed in a specific, documented case. A major German automotive OEM needed dynamic material data for plastic components subject to new passive safety regulations. The components were tested. The material cards were built from standard tensile data. The simulations did not match component-level crash behavior. Not by a small margin, systematically, and in a direction that pointed to something more fundamental than a modeling error.
Where the discrepancy came from
The cause was not a flawed simulation. It was a blind spot in the test method itself. A standard tensile test characterizes a polymer under one loading condition: tension. But polymers are not symmetric in how they respond to load. Most thermoplastics are measurably stiffer and stronger in compression than in tension. A tensile test cannot see this asymmetry, because it never puts the material into compression. It measures one half of the material’s actual behavior and reports it as the whole.
Real components rarely load a material in pure tension. A bending load, the load state that dominates in impact scenarios, casings, brackets, and structural plastic parts, puts one face of the material into tension and the opposing face into compression, simultaneously, within the same test. Three-point bending, unlike a tensile test, cannot avoid this asymmetry. It is forced to capture both sides of the material’s behavior at once, which is exactly what a tensile-only characterization cannot do.
Building a test around the problem, not around convention
Once the asymmetry was identified as the root cause, the engineering task changed. It was no longer “how do we run a tensile test faster,” but “what test setup captures the loading condition that actually governs the component’s failure.” That question led to a pendulum-based dynamic test system, purpose-built to combine bending and compression loading at the impact speeds relevant to passive safety, roughly 0.5 to 4.4 m/s, at energies up to 50 J, without requiring the infrastructure of a conventional high-speed servo-hydraulic testing machine.
Speed vs. strain rate
This speed range is where a common misunderstanding sits. A customer requesting material data for a pedestrian impact, or a specific occupant head impact scenario, often assumes the test itself has to run at the specific speed of the full scale test. What actually determines whether the data is representative is the local strain rate the material experiences at the point of failure, not the vehicle-level impact speed. Regulatory headform tests illustrate the numbers involved: pedestrian head protection under UN Regulation No. 127 / GTR No. 9 specifies a headform impact velocity of 9.7 ± 0.2 m/s, and interior occupant head protection under FMVSS 201 / ECE R21 tests at 24 km/h (6.7 m/s). The pendulum’s bending-and-compression setup reaches the same strain rates as these full-scale impacts through its contact geometry and specimen constraint, even though the absolute velocities are not matching. Matching strain rate, not matching speed, is what makes the material card valid.
Desktop-scaled solution
This is a detail worth pausing on. Conventional dynamic test rigs typically need reinforced floors, dedicated hydraulic supply, and continuous maintenance. Because the system that became IMPETUS® is desktop-scale and runs on standard building power, it can sit in a development lab or at a supplier’s site, not only in a specialized test facility. That is not a convenience feature. It is what makes it practical to generate the right kind of data early enough in a development program to matter, rather than late enough to only confirm a problem that already shipped.
Technical specifications: dimensions, mass, and electrical supply requirements – standard building power, no dedicated infrastructure.
One system, two loading modes
The pendulum system was built for dynamic bending, but it is not limited to that. It also performs dynamic tensile tests, which means a single desktop-scale device covers both loading modes at impact-relevant strain rates.
On the tensile side, results from the pendulum system were compared directly against a conventional servo-hydraulic testing machine across multiple unreinforced and reinforced plastic types. The stress-strain curves from both systems overlapped closely across the tested materials and strain rates.
Two loading modes, one measurement chain, one evaluation, and tensile results that line up with what an established testing machine produces.
From a test method to a validated component prediction
The value only becomes visible at the component level, where the choice of test data decides whether a simulation is usable. This is where the two loading modes stop being interchangeable. A material card identified from dynamic tensile data describes the material in tension only. A card identified from dynamic bending contains both sides of the response (tension as well as compression) as measured. Fed into the same CAE model, the two material cards do not produce the same answer.
In one documented case, a plastic component’s crash response was simulated using both approaches, a tensile based material card as well as a bending based material card. The resulting CAE model using the bending based card matched the measured impact test data, including the shape of the acceleration response over time, closely enough to serve as a validated basis for the design decision that depended on it — the same class of prediction that had failed at the start of this story, when the card came from tensile data alone.
Use Case:
Pedestrian Head Impact on a Bumper
A pedestrian head impact on a plastic bumper (≈40 km/h) is a bending-dominated load case. Both a tensile-based and a bending-based material card capture the first impact peak reasonably well, but only the bending-based card reproduces the second peak accurately, where the tensile-based card underpredicts the response.
If the component bends in the real application, start with bending data.
The underlying point
The story is not really about a piece of equipment. It is about what happens when a systematic discrepancy between simulation and physical test is treated as information rather than noise. The tensile test was not “close enough.” Building a new test around the actual loading condition, rather than defending the convenience of the existing one, was what closed the gap.
You can stay up to date by following us on social media: