PSI - Issue 38

Alexander Erbe et al. / Procedia Structural Integrity 38 (2022) 192–201 Author name / Structural Integrity Procedia 00 (2021) 000 – 000

193

2

1. Introduction The parametrization and validation of fatigue life models is commonly based on experimental results from standardized uniaxial testing. Components in service, however, are usually subjected to an arbitrary multiaxial loading. For the assessment of such multiaxial load case in engineering practice, equivalent strain approaches are used to convert a three-dimensional, multiaxial load case into a comparable uniaxial situation. One of the most commonly used approaches for metal materials is the “maximum distortion criterion” or “von Mises yield criterion”. According to this, the equivalent strain ϵ ‡“ can be expressed in dependence of Poisson’s ratio ν and –Š‡ principle strai • ϵ Ǧ for instance: = √2(1 1 + ) [( − ) 2 + ( − ) 2 + ( − ) 2 + 6 [( 2 ) 2 + ( 2 ) 2 + ( 2 ) 2 ]] 1⁄2 (1) For general elastic-plastic conditions the equivalent strains can be evaluated by simply adding the elastic interpretation of equation (1) with ν ൌ ͲǤ͵ and the plastic interpretation with ν ൌ ͲǤͷǤ It is known that the use of the von Mises criterion will often result in under-estimations of fatigue life, which means that the economic potential of the material is not fully utilized [KUL15]. But also for established approaches, the available validation data base to assess the validity and precision is often very limited. This is especially true in case of typical gas turbine materials such as high alloyed cast steels and nickel-based materials. The few systematic data collections for proportional multiaxial fatigue life available are often based on test campaigns using cylindrical hollow specimens or biaxially loaded cruciform specimens (e.g. [ITO94]). For varying axis-ratios, it was shown that the von Mises criterion cannot represent the whole range of possible variations [SAK13]. With the aim to generate a validation database under service-relevant loading conditions, an extensive experimental campaign was initiated at TU Darmstadt to investigate the influences of axis ratio, temperature, hold times, secondary loading, thermo-mechanical loading and non-proportional loading conditions on the biaxial fatigue life of a cast steel. In this paper, a first excerpt of the results from this study concerning the influences of axis ratio and test temperature are presented. 2. Experimental campaign 2.1. Experimental setup and material To investigate the influence of different load-axis ratios on the lifetime, the cruciform geometry given in Figure 1 was used. With a width of 110 mm, a center area of 15 mm in diameter and 1.8 mm thickness, the specimen leads to a homogeneous stress distribution within the center maintaining plane-stress state conditions. It has been widely used in prior investigations [CUI13], [SIM09], [LSY12]. The experimental campaign was carried out with a servo hydraulic planar-biaxial test-rig (F max = 250 kN) with two perpendicular loading axes controlled by a INSTRON 8800 control unit. For testing at elevated and high temperatures, the test-rig is equipped with an induction heating device, the coil of which is attached to the back side of the test-rig and close to the specimen. In order to measure or control the strain in the center of the specimen, a suitable biaxial mechanical extensometer for high temperature applications is attached on the front side. A type S thermocouple welded onto the specimen allows to control the test-temperature. Prior to the actual test, a uniform temperature distribution within the specimen is set up and calibrated using a thermal imaging camera based on IR-measurement technique. For this, the specimen is prepared with a thin firm of a special thermal paint HE6 which provides a stable and calibrated coefficient of emission without interacting with the specimen surface itself.