**5. Investigated influences**

332 Recent Trends in Processing and Degradation of Aluminium Alloys

Crack initiation in various Al-Si-alloys occurs preferentially at the interface between Almatrix and Si-phase. A meso-scale modelling of the microstructure under thermomechanical loading conditions shows stress concentration in the Al-Si-interface (Thalmair, 2009). The repeated recurring thermo-mechanical cycles cause micro-stresses and hence the

Fig. 3. Crack tip of a TMF-specimen (left) and a thermo-shocked cylinder head (right) of

TMF-test as thermo-shock test of a cylinder head shows very similar crack behaviour, where cracking of the eutectic phase and an inter-dendritic crack-propagation along the interfaces

Owing to the multiple effects in components exposed to TMF loading, the developed testing methods are rather varied. Meanwhile, besides LCF experiments and *bi-thermal tests* (Halford et al., 2004) also TMF experiments with combined temperature and strain control are also common. Because of the missing standardisation and the different requirements the experiments differ significantly with respect to heating and cooling temperature application, measurement and control as well as cycle form, strain measurement and consideration of thermal strains. Data from the literature typically allow for little comparability as the

The mechanical loading force application is mostly achieved by means of servo-hydraulic or electro-mechanic testing machines. For this purpose, test equipment manufacturers have started to offer powerful all-in-one systems. However, for special experiments one is still dependent on adaptions or self-made constructions (Riedler & Eichlseder,2004; Minichmayr et al., 2005). Furthermore, it is common to examine specimens with geometries close to the actual component, which only require temperature control. In this case the geometrical constraint is provided for either by the specimen shape or by the mechanical boundary

It is often tried to simulate the actual behaviour of the component by means of laboratory tests such as for example multiaxial TMF experiments (Otaga & Yamamoto, 2001) or superimposed HCF loading (Minichmayr et al., 2005). By measurement of real components

information concerning the testing procedures is usually insufficient.

conditions (Simon & Santacreu, 2002; Prillhofer et al., 2005).

preferred crack initiation.

AlSi8Cu3 (Thalmair, 2009)

is observed, see figure 3.

**4. Methodology** 

Firstly, it is important to clarify the governing damage mechanisms that occur in out-ofphase TMF cycles in cylinder heads. Therefore the tests on specimens were specifically designed to take the real circumstances in components as best possible into account – with the aim of using the derived models for lifetime estimations of TMF loaded components. Investigated influences are amongst others (see test matrix in Table 1) mean and local strains, cyclic and constant temperatures, dwell times, pre-aging and aging during service life, HCF-interaction, strain and temperature rates as well as the ratio of mechanical and thermal strain. Further single and multiple step creep tests have been carried out to take into account the stress relaxation phenomena. Additional tests in an argon atmosphere have finally enabled the isolation of the predominating damage mechanism in cylinder heads. All analyses are done in the manner of hysteresis loops, stress-cycle and plastic strain-cycle plots, lifetime diagrams and cyclic deformation behaviour diagrams.


Table 1. Test matrix
