Introduction to Fatigue Analysis Using Altair SimSolid
Altair SimSolid’s meshless solver is a revolution in structural analysis. Built to handle large assemblies with complex geometry and provide results at the speed of design, SimSolid has made simulation more accessible than ever before. With the addition of fatigue analysis to the already ground-breaking SimSolid solver, Altair now provides the world’s first entirely meshless structural simulation and durability tool.
SimSolid fatigue leverages this proprietary solver to produce stress inputs which can dramatically reduce the time spent in pre-processing and solving for structural analyses compared with traditional, mesh-based FE tools. That means getting to fatigue analysis faster. Designers, engineers and analysts can use SimSolid to rapidly evaluate design changes, run final stress validation, and test the durability of design variants all without having to create a mesh or leave the SimSolid interface.
As an introduction to SimSolid’s fatigue capabilities, let’s first discuss the basics of fatigue analysis and how SimSolid accomplishes this task.
Introduction to General Fatigue
Any load applied to a structure causes stress within its components. At a microscopic scale, applied stresses that are well below the yield stress of a material can exceed the material's strength, leading to microscopic failures within the material (crack initiation). When the same or similar loads are applied to a structure repeatedly, microscopic failures can grow and combine leading to macroscopic failures which can compromise the integrity of the component or structure (crack propagation). The initiation and propagation of cracks in a material due to repeated or cyclic loading is called fatigue. A structure’s resistance to fatigue can be called durability.
The goal of fatigue analysis is to identify the locations where this kind of crack initiation and propagation is likely to occur. This is done by first analyzing the structure under some load conditions to generate stress, then applying a cyclic load repetition pattern, and mapping the cyclic stresses or strains to established material fatigue data known as stress-life (SN) or strain-life (EN) curves. Therefore, the three main components of fatigue analysis are initial structural damage, cyclic loading, and SN/EN curve data.
Fig. 1 (a) Example of a fully reversed, cyclic load history. (b) Example of a stress-life (SN) curve showing scattered data.
SN and EN curves are usually plotted from controlled experiments with completely reversed loading. Since no two material samples have exactly the same microstructure, cracks can initiate and propagate differently in seemingly identical specimens causing some scatter in the final data. This is why fatigue analysis is considered a stochastic or probabilistic event: trying to capture where and when failure is most likely to occur by comparing analytical results with experimental data. Well-defined baseline stresses, material data, and solution settings are key to accurate fatigue analysis.
Since its initial release in late 2020, SimSolid’s fatigue capabilities have been continuously expanded to include many key functions. A list of some of these key fatigue features in SimSolid with brief descriptions can be found below.
| Stress-life Fatigue | Also called high-cycle or SN fatigue, requires a stress-life curve. Fatigue calculation method best suited for models experiencing stresses are well below the yield stress of the material and expected cycles to failure is above 10,000. |
| Strain-life Fatigue | Also called low-cycle or EN fatigue, requires a strain-life curve. Fatigue calculation method best suited for models experiencing stresses are well above the yield stress causing plastic deformation in the part and expected cycles to failure is below 10,000. |
| Uniaxial Fatigue | Stress tensor is converted to scalar quantity to calculate fatigue. Useful for simple, uniaxially loaded models. |
| Multiaxial Fatigue | Full stress tensor is used to calculate fatigue. Useful in most complex, real-world models. |
| Spot Weld Fatigue | Fatigue method specifically related to spot welds. Stresses within the weld and surrounding sheets are used with SN fatigue method to predict life. |
| Rainflow Counting | A method of processing variable load histories to isolate only the stress peaks and valleys and eliminate insignificant cycle data. |
| Mean Stress Correction | Methods used to normalize variable load histories to properly compare to SN or EN curves which are created under very uniform, reversed loading conditions. Many different methods depending on the load history and application. See SimSolid Help for descriptions of each. |
| Palmgren-Miner Damage Summation | A method used to sum up the damage done by each cycle and predict failure. States that failure is predicted when the number of reversed cycles at a certain stress level exceeds that which the SN or EN curve predicts. |
| Endurance Limit Factors | Factors that can be applied to the model which directly impact fatigue life such as surface condition/treatment or fatigue strength reduction factor. |
Fatigue in SimSolid
When running fatigue analysis in SimSolid, the meshless solver is used to create the baseline structural analysis then, once the cycle and solutions settings are ready, the inputs are passed to Altair’s world-class fatigue library (the same one used in OptiStruct fatigue and HyperLife) for fatigue calculations and the results are displayed within SimSolid for review.
As mentioned before, well-defined stress inputs, material data, and solution settings are key to accurate fatigue analysis. Let’s focus on the material models and solution settings SimSolid offers for fatigue analysis.
Fatigue Material Data
In SimSolid, fatigue properties must be defined for each material to run fatigue analysis. SN and EN curve parameters can be input manually or estimated from the ultimate tensile stress of the material. Creating fatigue properties manually requires a curve definition method and inputs for relevant parameters of that curve. The parameters used by SimSolid are available for all SN and EN curves and are defined in more detail in the SimSolid Help.
SN curves can have one or two slopes. EN can have only one slope. Estimating from UTS uses standard functions to estimate the SN or EN curve but is designed for steel materials only. To edit materials, go to Settings>Material Database>Edit current>Edit material. After fatigue curves have been defined, select Review to see the curve graphed in log-log format.
Solution Settings
Generally, fatigue curves are obtained from standard experiments with fully-reversed cyclic loading. However, the real fatigue loading could be not fully reversed, and the normal mean stresses have significant effect on fatigue performance of components. Mean stress correction (MSC) is used to account for the effect of non-zero mean stresses.
Depending on the material, stress state, environment, and strain amplitude, fatigue life will usually be dominated either by microcrack growth along shear planes or along tensile planes.
There is no consensus yet as to the best method to use for multiaxial fatigue life estimates. For stress-based mean stress correction method, Goodman and FKM models are available for tensile damage. Findley model is available for shear damage. For strain-based mean stress correction method, Morrow and Smith,Watson and Topper are available for tensile damage. Brown-Miller and Fatemi-Socie are available for shear damage.
If multiple models are defined, SimSolid selects the model which leads to maximum damage from all the available damage values.
Surface finish of the part(s) and scatter in the material data can be accounted for by changing endurance limit modifying factors. These factors directly impact the endurance limit of the material and in turn the fatigue life. Surface condition and treatment are extremely important factors influencing fatigue life because fatigue failures often initiate on the surface of parts. A chart showing the surface condition correction factors for steels can be seen below.
Options for surface condition include polished, ground, machined, hot-rolled, and forged. Options for surface treatment include nitrided, shot-peened, and cold-rolled.
There are other physical factors that could affect the fatigue strength of a structure, such as notch effect, size effect, and loading type. The fatigue strength reduction factor Kf is introduced to account for the combined effect of all such corrections. The fatigue endurance limit is divided by this value to affect the analysis.
Survival certainty is designed to deal with scatter in fatigue material data. As the certainty increases, more conservative parts of the scattered data are taken into account effectively lowering the mean life. This is used to estimate the worst possible mean life according to the standard error of the fatigue curve. If standard error is 0, fatigue life will not be impacted by changing survival certainty. Higher reliability of the result requires a larger certainty of survival.
See the video below for a demonstration of assembly level strain-life fatigue with SimSolid.