The result of the fatigue analysis is strongly influenced by the quality of the material data. Unfortunately, during the development process, there is hardly any chance to know all material values, therefore there is the possibility in FEMFAT to generate a new material based on only few data.
This support, offered by FEMFAT, is based on certain materials laws, specific for each material class (e.g. aluminum casting alloys, gray cast iron,...). It is very important to activate the correct material class before generating the new material. Please check the Haigh diagram after using the material generator, because possible failures can there be seen very easily.
If you need help for generation of new materials, please do not hesitate to contact the FEMFAT support at firstname.lastname@example.org
Today’s commercial simulation tools provide the possibility of simulating the casting process and of predicting the associated inhomogeneous structure conditions within the component. This gives, amongst other things, a distribution of the secondary dendrite arm spacing (SDAS), the solidification time, the cooling rate or the microporosity. It is possible to import these distributions into FEMFAT and to consider their influence on the endurance stress limit. Here, the solidification time and the cooling rate can be converted to an SDAS by means of an exponential approach. The currently implemented dependence of the endurance stress limit on the SDAS is optimized for aluminum sand and gravity die casting. However, the influence of the SDAS on the endurance stress limit can, if known, also be given by the user as a value pair table for any material. A new material dataset with the number 220 was created for this purpose. Nevertheless, it is also possible to consider microporosities or other structural parameters. In FEMFAT, the distribution of any structural parameter can be imported via the same interface, instead of an SDAS distribution. The only requirement is that the values are node-referenced. The file format thus corresponds to (node-referenced) temperature data. Moreover, for an operational strength evaluation, the influence of the specified structural parameter on the material endurance stress limit must be known and be defined by means of a table (dataset 220) in the material file (*.ffd).
The influence of cast structure in FEMFAT is already being successfully implemented by BMW in their engine development process, e.g. for cylinder heads.
S/N curves for varying SDAS
Effects of SDAS on the analysed endurance safety factors*
*[P. Nefischer, F. Steinparzer, H. Kratochwill, G. Steinwänder, BMW Motoren GmbH;
"New approaches in damage analysis of cylinder heads" (New approaches in damage analysis of cylinder heads), 12th Aachen Colloquium for Vehicle and Engine Engineering.]
The technological size in FEMFAT considers the material endurance stress limit which decreases with increasing component dimensions. The imported shell thickness, which is averaged at the nodes in FEMFAT, is used as the technological size for FEM shell entities. For FEM solid entities, on the other hand, the user must explicitly specify the technological size as a numerical value in [mm] in the node properties. You can activate the technological sice influence factor in the menu "Influence Factor".
The sample thickness is also included in the analysis, where it should be noted that the sample thickness at five points is included in the material data, i.e. in the strength data for tension, compression, bending and shear, and again in the general S/N data. However, only the value from the general S/N data is used in the technological size influence!
Details from the FKM guideline
With the size influence activated in FEMFAT, an influence factor on the material alternating stress limit is calculated, in accordance with the FKM guideline [1, 2]. The "effective diameter" used in the equations there is identical to the "Technological size at 3D nodes" entered in FEMFAT for the node properties of the component.
In the following figure the influence of the technological size on the endurance stress limit is shown for various material classes. The corresponding FEMFAT material classes are given in brackets.
 FKM guideline "Numerical Strength Analyses for Machine
Components in Steel, Cast Iron and Aluminum Materials",
4th Edition, 2002, www.vdma-verlag.de.
 FKM guideline "Analytical Strength Assessment", 5th Edition
Enabling the isothermal temperature influence means that the local operating temperature is taken into account at all analyzed nodes (usually upper temperature).
FEMFAT provides various analysis methods for this:
“FKM guideline” method
The isothermal temperature only affects the dynamic strength values in the Haigh diagram (vertical scaling).
“FEMFAT 4.5” method
This option changes both the dynamic strength parameters in the Haigh diagram (similar to the FKM guideline) and the static strength parameters (horizontal scaling of the Haigh diagram). The same formulae from the FKM guideline are used for modifying the yield stress and ultimate tensile stress. The user is warned if the temperature exceeds the range defined in the FKM guideline, and the influence factor is extrapolated on the basis of the implemented formulae.
“User defined” method
In this method, the temperature dependent material data specified by the user is taken into account; this data can be defined in the menu “Material data -> Properties at higher temperatures”. The temperature dependent fully reversed endurance limit and the ultimate tensile strength must be defined as a minimum for this method.
“FEMFAT 4.6” method (default)
This option is based on method 4.5, but in addition the cyclic coefficient of hardening K’ of the cyclic stress-strain curve is reduced in a similar way as the ultimate tensile strength. This means that the isothermal temperature influence is also correctly taken into account in the mean stress rearrangement of linear elastic stresses according to Neuber (FEMFAT plast module).
Isothermal temperature influence on Rm, Re and K´
One can see from the curves that all temperature influences based on FKM produce conservative results.
If you are unable to find the required material in the material database, you can also generate the data for the calculation on a few familiar properties of the material. The user has 2 options.
The first is called the 'stress-controlled' method by FEMFAT. Here you only need to specify the material class (Fig. 1) and the ultimate tensile strength for the material. Based on the predefined ratios taken from the FKM guideline, the system will now automatically generate all the missing material characteristics needed for the calculation (Fig. 2).
To increase the reliability of our conclusions, it would naturally be better if one was able to also pre-define the ultimate yield stress, pulsating load strength and ultimate compressive strength/alternating stress limit from specimen test data. However, one should be careful when changing materials outside of the material generator since the FEMFAT calculation uses a number of different relations based on the material data, for example the 'k' factor used to measure the ductility of the material.
The second method for generating a material is called the 'strain-controlled' method. The user not only needs to enter the material class (as in fig. 1), but also the data for ultimate tensile strength, yield stress and E/N curve (unnotched, polished samples, fully reversed tension/ compression) (fig. 3).
Finally, the number of cycles at the fatigue limit is also needed. Next, for this number of cycles, the strain amplitude is converted into a stress amplitude (corresponding to the fully reversed endurance limit). The fatigue strength exponent b is used to determine the gradient of the stress S/N curve (k=-1/b). After the material characteristics have been determined, you should then check the following and change where necessary: Survival probability, specimen diameter, elongation of rupture, check Haigh diagram & S/N curve for plausibility.
Since FEMFAT version 5.0 it is also possible to estimate parameters of materials with pores inside. In many cases as e.g. in Aluminum die cast components, there are areas with pores and defects, whereas surface layers are pore free. In FEMFAT a boundary layer model is provided for correct fatigue assessment of such cast components. Nevertheless the necessary parameters for material with and without pores are often not known. A new dialog for defect definition is now provided (fig. 4). Based on a pore free material or a material, where the maximum defect size is known (which is relevant for fatigue failure), material parameters can be derived for any other defect sizes.
Fig. 4: Defect definition Dialog
FEMFAT Support is of course also happy to assist you with any further material data generation issues.
FEMFAT provides an option for defining temperature- dependent material parameters (static and dynamic) and thus to take them into consideration in the fatigue analysis. The simplest solution is to import an existing material from the FEMFAT material database, e.g. the cast metal AlSi12CuNiMg.
Right at the bottom of the "Material data" dialog there is an entry titled "Properties at high temperature". Here, value pairs consisting of the temperature and the corresponding material parameter can be entered.
Attention should be paid that the strength parameters correspond to the imported basic parameters at 20°C. For a user-defined assessment of the temperature influence at least the temperature-dependency of the ultimate tensile strength and the alternating stress limit must be given.
The remaining temperature-dependent material parameters can be estimated automatically by FEMFAT – although it is better to provide more temperature-dependent material properties, such as the pulsating stress limit, for example.
This extended material can now be saved to the material database once again using the "Write to materials database" command and be used for subsequent analyses. It is adopted automatically for the current FEMFAT session.
A nodal temperature distribution must now be defined as an "isothermal influence". Do not forget to activate the isothermal temperature influence in "Influence factors" and to change the method used to "user-defined".
Here, a simple example will be used to reveal the applicability of the FEMFAT plast method.
The same cyclic-plastic material data are used in FEMFAT and ABAQUS (parameters K´ and n´). The following analyses were performed on a fine-grain steel:
1. ABAQUS with a cyclic-plastic material curve and true stress-strain definition with non-linear geometry.
2. ABAQUS with linear material behavior and stress rearrangement using FEMFAT plast.
A notched, tensile specimen of steel with notch radius 0.7 mm, (d = 2.15) was analyzed for an alternating load of up to 40 kN (stress ratio R = -1). In Fig. 1 the stress history at the base of the notch can be seen for the Mises and max. principal normal stress. Initial plastification in the notch occurs at approx. 8 kN.
The Neuber rule applies to regions with limited local plastic zones. The individual regions are shown in Fig. 1. In region 1 (0-10 kN) local plastification restricted to the notch prevails.
Fig. 1: Notch stress development in the notch base with plastic material behavior in ABAQUS
In region 2 (10-20 kN) the entire cross-section begins to plastify. In region 3 (above 20 kN) the entire cross-section plastifies. In our case the Neuber rule can be applied up to around 20 kN, because the initial cross-section plasticizes only very little up to this limit. Due to the loss of stiffness there is a large increase in local stresses at a load of 40 kN.
Fig. 2 shows a comparison of the stress amplitudes between ABAQUS and FEMFAT plast. The principal normal stress is used for redistribution in BASIC. The plastic analysis in ABAQUS uses the Mises equivalent stress for this purpose. This results in stress deviations in FEMFAT, but also because the node-averaged stresses are used for stress rearrangementin FEMFAT and not those at the element integration point.
Fig. 2: Comparison of stress amplitudes from FEMFAT PLAST/ABAQUS
Although a multi-axial stress state already exists in the notch, PLAST provides useful results up to approximately 15 kN (~ 0.3% plastic strain) for the local stress in the notch. If greater strains or stress multiaxiality are anticipated deviations may increase and a non-linear material law should then generally be employed.
In the most simple case, a new data set is created with the material generator in FEMFAT by specifying the material class and at least the ultimate tensile strength.
Sometimes, however, even the first step in this simple situation proves to be impossible: this happens when the desired material class (such as powder metal) is not available.
For such cases you will find a table which shows the equivalent material classes we recommend.
After generating the material, the actual material class must still be specified in the general data (if available; see the 3rd column in the table).