Even when a part withstands the stress of a single, ultimate load without breaking, it can still fail by "tiring" after repeated cyclic stress.
"Off-highway vehicles are usually subjected to severe load specifications, requiring careful durability analysis,” said Neil Bishop, Durability and Fatigue Analyst for MSC Software.
Today’s emphasis on fuel efficiency is making fatigue analysis more important. “Vehicles are generally becoming lighter and using more specialized materials. This is making durability analysis a more complicated process,” he says.
How do parts fail in fatigue? As Dr. Michael Hack, Product Line Manager for durability simulation from LMS International explained, fatigue means failure due to incremental growth of cracks through repeated loading. Crack initiation is typically seen as a change in the stiffness of the part. These cracks then grow to a complete break or failure of the part. “Fatigue can be considered an element or subset of durability, although you often hear these terms used incorrectly as describing the same thing," Hack said.
Simulating fatigue through analysis requires an FE model, materials and manufacturing data, and loads data as input. These supplied loads could be measured directly from a prototype, partially predicted using some measured data combined with analytics or predicted using a multibody simulation such as LMS Virtual Lab Motion.
Historically, loads were supplied from physical testing alone. “Multibody simulations and 1-D simulations from hydraulics are much more efficient today,” said Hack, enabling simulation of load data. However, physical testing remains a key element. Although it varies from customer to customer, he observes every vehicle program measures loads at least once.
“Solely relying on predicted loads is not done as standard practice yet, but it is getting close,” he said. “Getting the load right is still a challenge and is more time-consuming than calculating fatigue. Fatigue accuracy results are mainly influenced by the quality of loads and local stress-strain histories.”
Considering the accuracy of the predictions, the quality of the inputs may actually be more important, agreed Dr. Peter Heyes, Technical Specialist responsible for fatigue life prediction methods for HBM nCode.
“In particular, the understanding of the service loads,” he said. “Accuracy of service loads probably yields the greatest benefit to improved quality of prediction.”
Heyes also noted that in recent years improvements in computer performance and modeling software have enabled analysts to use ever larger FE models and to capture geometry and nonlinearities with greater fidelity. While providing a set of tools for managing and analyzing service data and fatigue analysis, nCode also provides accelerated testing. This allows developing shaker table tests, matching fatigue and shock loads in service.
“Fatigue of welds is key to the durability of off-highway vehicles,” he said. “We are continually working to improve [simulations for] thicker sheet welds and for solid FE models.”
Hack from LMS noted three improvements in analysis his company is developing for the future. One is further improving the seam weld methodology to account for the special requirements of off-highway equipment, such as complex weld and base sheet geometries. “We are working directly with a large European off-highway manufacturer of excavators.”
LMS also continues to improve interfaces to the manufacturing process software. “This means getting the behavior, for example, of cast parts more accurate,” said Hack. But not only the fatigue methods are important but also the further improvement of load prediction, using the LMS multibody solution combined with the LMS hydraulics and mechatronic simulation.
“The capabilities of fatigue and durability software is rapidly catching up with [physical prototypes,]” said MSC's Bishop. “For example, automated spot and seam weld analysis in MSC Fatigue is now well developed and offers dramatic time saving over traditional approaches.”
Future improvements to look for from MSC also include new weld analysis as well as new simulations for shaker table tests. “Components are routinely tested on shaker tables with random and sine sweep vibration loadings. MSC Fatigue will soon be able to replicate these test conditions in the virtual world,” noted Bishop.
They will also improve integrating fatigue CAE with the rest of the MSC environment. “There is a need to offer an integrated solution. Today the world operates in silos. Therefore, a motion engineer or a multibody dynamics engineer sits in his own world and does an analysis. The fatigue guy is not leveraging those loading histories to get an accurate understanding of their own fatigue life predictions,” he said.