**5.4 Influence of mean strain**

Impressed mean strains in the manner of pulsating LCF tests (strain ratio=0) show a visible decrease in lifetime for higher strain levels and only a slight lifetime decreasing effect at lower strain levels. At latest from the half of the number of cycles to failure *Nf/2* on, the cyclic stress deformation curves follow a common progression. Some slight lifetime-decreasing effects ascertainable at pulsating executed test series mostly result from the first few cycles, where the higher tensile stresses and plastic strains (see figure 5 (right)) cause higher damage rates. The comparison of alternating and pulsating executed TMF test series shows the same tendencies as at the LCF results. At latest from *Nf/2* on, the cyclic stress deformation

Comparison of Energy-Based and Damage-Related

**5.6 Influence of phase shifts / strain compensation** 

were tested. Besides the ideal OP-TMF situation (

**5.7 Influence of strain rates / argon atmosphere** 

(*KTM*=-1.0) were tested.

lifetime, as figure 10 shows.

**5.8 Influence of HCF interaction** 

(Minichmayr et al., 2005).

**5.9 Influence of creep** 

Fatigue Life Models for Aluminium Components Under TMF Loading 337

In order to investigate the influence of stiffness and phase shifts, different TMF conditions

conditions of the thermal strain (*KTM*=1.5 and 2.0), a 75% compliance (*KTM*=0.75, which is near to the real circumstances in cylinder heads) and an ideal in-phase-TMF situation

When the local strains are taken into account at rigid clamped specimens, all OP-TMF results can be drawn together in a common strain vs. cycles to failure diagram. Because of creep damage at higher temperatures the IP-TMF lifetime is shorter than the OP-TMF

The systematic investigations show that the influence of strain rate on the deformation behaviour is negligible within practical range of temperature rates, but the time and temperature dependent aging behaviour is very important. Unlike the deformation behaviour, the strain rate shows an important influence on the lifetime behaviour due to the additional creep damage involved. These differences and the differences at IP/OP-TMF in the number of cycles to failure allow the separation and investigation of different damage mechanisms. Additional tests in argon atmosphere were executed for this aim, which show

Superimposed HCF-loading has an important influence on the fatigue life of TMF-loaded components. Experiments with superimposed HCF strain amplitudes from 0.01% to 0.1% show a significant decrease in fatigue life depending on the amplitude of the HCF-loads, wherein a small influence of HCF-frequency is found. Metallographic investigations show crack propagation due to HCF-loading and TMF-loading, wherein a combination of HCFamplitude and the shift in mean-stress causes crack propagation. Good correlation between

HCF-loadings in a combustion engine occur especially during the heating of the component with maximum ignition pressure. Therefore additional experiments were conducted with superimposed HCF-loadings only during heating period and dwell time. In this case most of the HCF-cycles appear within the compression region of the hysteresis loop. Therefore the effect is obviously reduced. With regard to the typical ignition pressure in a diesel engine, the influence is small compared to the tests without superimposition; see figure 7

Because creep effects have to be considered in thermo-mechanical loaded components to take into account stress relaxation phenomena and creep damage, single and multiple step creep tests were carried out. Due to aging effects the decreasing strain rate of the primary creep stage of the single step tests directly merges into the stage of tertiary creep with material softening and therefore increasing creep strain rates. Multiple step tests show, that neither strain nor time are suitable to describe the creep behaviour for that case. Therefore different tests with varying pre-exposure times at test temperature were conducted. The pre-aging time was chosen according to the total time in the multiple step tests. It can be

striations (~crack propagation) and strain amplitude for different loadings is found.

a lifetime increasing effect, because of oxidation damage being minimized.

ε*t,mech*=ε

*th*, *KTM*=1.0), two overcompensated

curves follow a common progression and only some slight lifetime-decreasing effects are ascertainable at pulsating executed TMF test series.

### **5.5 Influence of dwell-time**

The typically start-stop-operation of a motor vehicle as well as the alternating fired and nonfired operation causes dwell times. To study this effect out-of-phase TMF tests with four different dwell times at the particular maximum temperatures were conducted (8, 24, 144 and 864 s).

Whereas with the alloy AlCuBiPb, a higher dwell time always causes a lifetime decreasing effect. This is not the case with the alloy AlSi7MgCu0.5, as figure 6 (left) shows, where the strain values are scaled between the minimum and maximum in this range. The TMF strainlife curve for the dwell time of *tD3*=144 s is quite steep for higher strain values (and therefore temperatures). A point of intersection of the curves for the lower dwell times (8 s and 24 s) is visible at about 1000 cycles. This phenomenon is explained with pronounced softening effects in the first few cycles that occur due to the high aging tendency at the high dwell time and temperature level. The capacious over-aging at this level can mainly be seen in the highly plastic parts, what results in an upward movement of the total strain-life curve. For that reason the high over-aging at the dwell time of 144 s shows that a high mechanical strain amplitude can be endured for a longer time compared to the smaller dwell times. If the maximum temperatures are low, this effect turns around at the dimensioning level for cylinder heads (about 5000 cycles) and a lifetime decreasing effect is visible with a higher dwell time.

Two extreme LCF tests were conducted at the mean value for the TMF maximum temperatures. A LCF test series at a constant higher temperature of 250°C and one tested at room temperature, but pre-aged at 250°C for 500 hours. Figure 6 (right) shows that these two LCF strain-life curves span the TMF range for all dwell times of the materials investigated, if the mechanical strain (and not the thermal strain) is considered. The comparison of the LCF hysteresis loops at 250°C with comparable mechanically strained TMF hysteresis loops for all dwell times also shows a good accordance. Moreover the cyclic deformation behaviour according to Ramberg-Osgood shows the best accordance of the LCF-250°C curve with the TMF curves. This investigation shows that the macroscopic behaviour is comparable if the aging status is similar (Riedler et al., 2004).

Fig. 6. Influence of TMF dwell-time and LCF-pre-aging and constant elevated temperature on the lifetime (on the left) and on the cyclic deformation behaviour (on the right)
