Designing to Meet Highway Regs
Simulation, analysis ensures roadworthiness and safety

-by Richard Mandel

If they weren’t already acquainted with the concept, drivers will soon be learning the importance of having the correct tire pressure for highway driving. Accidents caused by incorrect tire pressure have made headlines and cost some automobile and tire manufacturers substantial amounts of money and credibility. Because of these statistics, the National Highway Traffic Safety Administration (NHTSA) has published the TREAD Act of 2000, declaring that tire pressure monitoring systems (TPMS) will be standard equipment in passenger vehicles and light trucks beginning November 1, 2003.

With TPMS compliance standards and specifications already defined, the auto manufacturers are awarding supply contracts to meet the NHTSA implementation schedule. The devices have to function in extreme conditions. They face temperatures varying from –40º to 125ºC, along with loading conditions that include very fast acceleration, bumpy roads, and G forces up to 3000 — created by forces generated by tire rotation.

Design considerations

Since 1987, SmarTire Systems Inc., Richmond, BC, Canada has been developing and marketing tire monitoring systems for the automotive and transportation industries. SmarTire technology uses proprietary wireless protocol to allow a driver to detect loss of tire pressure while the vehicle is in motion. The package features a custom-developed Application Specific Integrated Sensor (ASIS) that contains pressure and temperature sensing elements and a digital logic state machine which functions as the "brain" of the sensor/transmitter. This single chip is robust in design, provides a black box recording and optimizes battery life. The product features transmitter options with different installation choices for different applications. A strap-mounted transmitter attached to the wheel offers universal installation for a wide range of tire and wheel assemblies. Mounting the transmitter to the base of the tire inflation valve offers an adjustable, secure in-tire installation for specific wheel profiles. A third installation option currently under development is a bonded solution. Once installed, sensor/transmitters do not require on-going maintenance.

In addition, various modes of sensing and communication ensure faster transmission of data when problems occur. Transmitting the data from the sensor through the tough exterior of the tire and the steel frame of the vehicle is no easy task, therefore, packaged on the circuit board with the ASIS are various components supporting SmarTire's proprietary RF technology.

Meeting regulatory demand

The new TPMS regulations will create significant product demand and require suppliers to scale accordingly. SmarTire studied their earlier, limited-production tire pressure data transmitter models and observed that they had a single body for the whole device — if the transmitter broke during a tire change, it could cost $200-300 to replace the transmitter and reprogram the receiver.

The new transmitter, however, raised some design challenges — particularly regarding the materials chosen and the snap fit of the base and upper housing for the electronic components. The design team wanted the lightest possible units, and then needed to make sure thin-walled structures could stand up to the tough operating conditions. In addition, they had to meet a tight production deadline.

To further aid development, the engineers applied finite element analysis with MSC.visualNastran for Windows, from MSC.Software Corp, Santa Ana, CA. Bahaie and his team ran linear static, nonlinear and dynamic/vibration analyses to make sure the transmitter core would remain safe in extreme environments.

Simulations

Static loading provided simulation of the transmitter during a tire change. The engineers knew that in a worst case scenario the tire was likely to hit the transmitter assembly from the side. According to Kian Sheikh Bahaie, senior mechanical engineer for SmarTire, in the new model, “A cradle interfaces the transmitter core with the wheel. If the tire hits the transmitter assembly, the plastic cradle breaks but the transmitter core survives. We analyzed a huge load from the side, but found that the cradle would break before the transmitter reached a critical stress condition.” Next, they tried loads simulating huge G forces, finding that the transmitter should be completely safe at up to 400 km/h (the typical target for SmarTire transmitter performance tests) with a safety factor of two at high or low temperature.

MSC.visualNastran for Windows proved equally helpful to the design team in simulating vibration as a result of the suspension on the road. They ran a normal mode, to make sure that at normal speeds and vibration conditions the whole assembly would pass its tests.

Because of the rotation forces, the entire load on the transmitter goes upward, making it essential to avoid separation from the cradle. Strengthening the top housing to handle that load are a number of reinforcement ribs, each needing to be analyzed to make sure the weight on the rim stays low. Weight requirements meant that the design had to use light, strong plastics. “We were limited on the type of plastic and the amount of material we could use,” Bahaie says. The team went with nylon-based plastic reinforced with internal glass fibers. However, because of the part’s complex geometry, during injection molding the glass fibers don’t all have the same orientation, nor are they uniformly distributed. If the glass fibers become concentrated in a corner, they can cause stress concentration areas in the part. These in turn would cause difficulties during physical design validation tests.

Knowing this, SmarTire asked the mold maker to use its mold software to simulate the glass fiber orientation in different areas. States Bahaie, “We used the output and fine-tuned our meshing and material geometry in different areas of the part, and then performed nonlinear analysis on the cradle and top housing, which are the critical parts.”

Deadlines and production lines

Meeting the deadline presented the next challenge. “We needed three or four months for the life test on the electronics alone,” Bahaie explains. “We didn’t have time to rework the injection molds if there proved to be a problem with the part. Because of analysis, the plastic parts passed all the tests without rework.”

SmarTire typically does an initial physical design validation test run with between 200 and 250 assemblies, fixes any design problems that show up, and follow with second physical test run on 800 units. The closer to the final product the team could come in the first DV run, the easier it would be to have the product ready on time.

Bahaie comments that it was tricky to test both the material and the snap fit between the base and cover. The team performed nonlinear analysis on the corners of the base to make sure the snap fit would not break. “Plastic with glass fibers is a little brittle, and with very small snap fits we had to make sure the mechanism worked perfectly before making the mold,” he says. “It’s very difficult to rework this small area on the mold.” To reduce costs, the design doesn’t require a slider inside the mold.

The base of the housing has six holes beside each snap fit. “That means there’s less support on the bottom of the snap fit, which might cause some stress concentration in that area. MSC.visualNastran for Windows helped us monitor the problem. We did three analyses and modified the geometry three times before we got to the optimum shape and size of snap fit,” says Bahaie. “Everything was right the first time, despite our mold maker predicting that the snap fits wouldn’t work.”

According to Bahaie, analysis helped SmarTire save “about 40% on material in both price and amount. We reduced the wall thickness from 3.0-4.0 mm to 1.2 mm compared to the previous design, and reduced the base to 0.4 mm. The previous material’s specific gravity was 1.5, while the new material is only 1.16. It’s lighter, it’s much easier to inject for molding. The design passed all its tests, thanks to the good data we obtained with our MSC.visualNastran for Windows analyses.”

Exercising Restraint on a Simulated Schoolbus

Designing school bus seat and safety restraint systems that meet federal motor vehicle safety standards FMVSS No. 222 and FMVSS No. 210 has typically involved building physical prototypes for destructive testing. Since each prototype takes about seven weeks to make, every failure adds to time and cost of development. More importantly, testing of a physical prototype does not provide a complete picture of all the stresses — the stresses are known only if and where failures occur.

While seat belts have been required on passenger cars since 1968, no such law exists for large school buses weighing more than 10,000 pounds. According to the National Highway Traffic Safety Administration (NHTSA), approximately 450,000 public school buses travel approximately 4.3 billion miles to transport 23.5 million children to and from school and school-related activities every year. According to the National Safety Council, there were more than 62,000 student injuries reported from school bus accidents from 1991 to 1996.

IMMI (formerly as known as Indiana Mills & Mfg), located in Westfield, IN, has more than 35 years of experience in developing, testing, and manufacturing occupant restraints for commercial vehicles, child and infant seats and off-road machinery, and is one of the world's first seat belt manufacturers. The company’s new school bus seat and safety restraint system, introduced in the second quarter of 2002, is expected to help minimize injuries to students.

Process

By using nonlinear simulation software, IMMI proved its new design concept, combining traditional occupant containment concept as embodied in FMVSS 222 with lap and shoulder belt restraint, before investing in physical prototypes. To simulate loads and stress, IMMI designers had to deal with several critical issues:

Nonlinear Simulation — because the FMVSS 222 regulation is more of a quasi-static loading procedure, the solution times with an explicit solver would have been enormous. Besides protocols for the regulation, the seat itself has multiple contacts between the loading bar, contact between various parts of the seat, and large rotations of different members of the seat with some material nonlinearity (including plasticity) and boundary nonlinearity (including buckling). MSC.Marc, an implicit solver, allowed a larger time step with almost no numerical stability problems, delivering much faster solve times for nonlinear transient problems.

Material Properties — material property handbooks provided values for some of the yield stress and work hardening slope material properties for the plasticity loading. It was suspected that some of the handbook values were uncertain. During development, an Instron tensile test machine was purchased, enabling IMMI to generate its own material properties data. Mr. Brueggert explained, “We tested material specimens from the seat and obtained better than handbook yield and work hardening values, which we used to define the stress strain curve in MSC.Marc.”

Geometry — because there were very few solids involved and most of the mesh was defined with shell elements, IMMI’s design engineers created midplane surface geometry and exported it as IGES files. MSC.Patran was used to generate suitable meshes for the shell elements of the various parts making up the seat, including frame and cushions. The geometry was all married together and brought into MSC.Marc, where the nonlinear elastic/plastic material properties were defined and some contact bodies were defined, including two contact bodies or loading bars that pressed against the back of the seat.

Meshing — some of the early meshes were too coarse to make up the curvature that occurred as members deformed plastically and buckled locally. Based on information from the stress plots, some of the areas were re-meshed with a finer mesh. Some areas were attached to each other with spot welds (rigid constraints tying portions of the structure together). This procedure required coordination with the design engineers to ensure the precise location of spot-welds tying the structure together.

Loading — the regulations specify the loading involved. For example, FMVSS 222 specifies the loading location (elevation), cylinder diameter and length for the seat back, direction, type of cylinder pivot, onset rate and length of time. The results determine whether the simulation meets the minimum value of energy that must be absorbed during the test. FMVSS 210 is a seatbelt anchorage test. The loads are applied at the structural attachment points for the seat where the seat belts attach. A specified load factor is used for direction and magnitude of the anchorage points.

Results

The performance of the lap and shoulder belt system in restraining occupants is evaluated with a rapidly time-variant loading or deceleration pulse. It is not unusual for a seat to behave differently under this dynamic loading than under the quasi-static loading of FMVSS 222 and 210. Evaluating the seat dynamically requires more testing and iteration between quasi-static tests and dynamic tests. Marlin Brueggert, IMMI’s applied mechanics manager, explains, “Our MSC.Marc seat simulation models were followed with an explicit time-integration simulation coupled with an occupant dynamics simulation to evaluate the seat under dynamic conditions. By working in this fashion, we leveraged the simulation benefits for static testing to reduce the number of loops through the quasi-static and dynamic test procedures.”

Essentially a time-dependent load curve is conducted to determine the load vs. displacement characteristics at the loading point of the seat. The load vs. displacement curve has to fall within an upper and lower bound to meet the FMVSS 222 energy absorption requirement.

Testing a physical prototype is a pass/fail test that doesn’t indicate how close to the safety margin it is. If there is a failure when testing a physical prototype, it is usually at one point, which can be fixed. However, the design still may not work after fixing the failure, because the second or third weakest points are unknown. With the MSC.Marc simulation, weak points in the model were identified and redesigned before going to a physical prototype. Brueggert explains, “In the MSC.Marc analysis we could go back and look in the plastic strains and stresses to see how close to failure all the areas that didn’t fail were. The stresses, strains, and loads from the finite element analysis revealed a great deal of data on how close most of the members were to the margin of failure. In this manner, final testing of the product served to complement the preceding analyses. This information was invaluable in the manufacturing process, allowing engineers to assess proper tolerance levels on manufacturing deviations.”