Influence of test parameters on the thermal-mechanical fatigue behavior of a superalloy
- PDF / 1,895,187 Bytes
- 11 Pages / 594 x 774 pts Page_size
- 93 Downloads / 195 Views
I.
INTRODUCTION
T H E R M A L fatigue (TF) is one of the primary life limiting factors for blades and vanes in jet engines? 11Thermal fatigue has been the object of considerable effort in turbine components. However, damage modeling under TF cycling is still at an early stage t2,3,4] as compared to creep and isothermal fatigue at high temperature. One of the reasons is the complexity of such a loading. Thermal fatigue is far from being a pure fatigue phenomenon and may involve creep ~3,41 or interactions with environment. tS-sl Longtime operation and repetition of thermal transients may result in microstructure changes which can alter the stress-strain behavior I7,9,~~ and damage mechanisms. However, one of the major obstacles to the development of TF damage modeling has been the difficulty of simulating TF in the laboratory. In the past, simple components were tested using Glenny's tapered discs, or wedge specimens and fluidized beds, t",~2~ or flame heating. [~3] These tests provided an inexpensive simulation of service conditions. However, the assessment of a damage model needs sophisticated heat transfer analysis and an inelastic thermal stress computation. Es,141The prediction of TF life depends upon the validity of the damage model and upon the computation of the stress-strain behavior. Thermal-mechanical fatigue (TMF) tests were introduced to simulate the behavior of critical parts of real components, ttS,~61The TMF test uses an actual specimen and not a component. There is no temperature gradient across tile specimen section, and the stress is induced by a mechanical strain which is applied to the specimen to simulate the constrained free thermal expansion of a component part. Since all test parameters are experimentally known (measured or imposed), this test can acJ.L. MALPERTU, formerly with the Centre des Matrriaux, is now with the Etablissement Technique d' Angers, 49 Montreuil-Juigne, France. L. RI~MY is with the Centre des Matrriaux, Ecole des Mines de Paris, UA CNRS 866, BP 87, 91003, Evry Cedex, France. Manuscript submitted July 14, 1988. METALLURGICAL TRANSACTIONS A
tually simulate the behavior of a volume element under anisothermal conditions. It can thus be used to assess the validity of both stress-strain behavior and damage models to be used for real components. Pioneer authors in the field tlS'~6~ used two basic mechanical strain v s temperature cycles: the "in-phase" cycle where the maximum mechanical strain occurs at maximum temperature and the "out-of-phase" cycle where the maximum mechanical strain occurs at minimum temperature. All the tests were run under fully reversed strain cycling. This gave rise to some controversy about phasing effects and about the comparison between TMF and low cycle fatigue (LCF) at high temperature, t~7-2~ However, simple in-phase and out-of-phase cycles seldom occur in the thermal stress analysis of real components, such as blades and vanes, t2~l However, in-phase and out-of-phase cycles are useful for validation of models. Smooth specimen testing is widely
Data Loading...
