EXPERIMENTAL SIMULATION OF A MARTENSITIC STAINLESS-STEEL COATING SUBJECTED TO THERMAL FATIGUE

Studies of thermal fatigue damage have been an area of ever-growing in terest in the industries such as aerospace, nuclear, metallurgy and ot her industries where metallic structures have been subjected to thermo mechanical loadings. Thermal fatigue effects will only be evident, if the thermal cycled structure size is big enough. In this work, we have studied continuous cast cylinders which are components submitted to t hermal cycles. The repetition of temperature cycles induces many fract ures on material surfaces and could induce cylinder failure. In order to attempt to solve this important technological problem by protecting the surface, a coating was deposited on the cylinder by a semi-transf erred plasma are technique; this method provides a coating which is no t too diluted into the support. In the present work, a martensitic sta inless steel was deposited on a cylindrical low alloyed steel substrat e, simulating a true cylinder; the specimen used in this study has dim ensions of a 1/5 of those of a real cylinder. It was high frequency he ated with a pancake type inductor and cooled with a compressed air sys tem on its outer surface and an axial flow cooling water inside; the s pecimen was rotated at a speed of 1.1 R.P.M. An experimental temperatu re map was established using two infrared pyrometers for temperature s urface measurements and Thermocouples for infernal temperature measure ments; these were inserted into the cylinder. The objective of this tr idimensional simulation is to obtain the best reproduction of thermal loading of the casting cylinder, When the thermal stabilized cycle is reached (after 15 cycles), twenty temperature measurements were carrie d out in order to construct the experimental thermal map of the stabil ized cycle, Furthermore, different experimental maps were constructed from the initial state to the stabilized cycle. The temperature range was established to correspond to typical industrial values, and moreov er, the experimental thermal maps will be used to a future numerical s imulation of the thermomechanical behaviour of the cylinder. The highe st measured temperature was 500 degrees C close to the inductor; the l owest temperatures were measured near the axis at 40 degrees C and als o at the end of a generating line at 180 degrees C. Furthermore, resid ual stresses and deformations were measured in the coating depth befor e the thermal fatigue experiment and at different periods of the test. These measurements were performed using X Ray and neutron diffraction , before and after thermal fatigue cycles. Experiments performed on cy lindrical specimens are difficult and expensive; thermal fatigue tests serve only for the verification of the coating toughness and for the construction of thermal maps. The study of the microstructural evoluti on of the coating was conducted with the help of other smaller specime ns, which were submitted to thermal shocks;some microhardness measurem ents were carried out to estimate hardness variations of the coating. These specimens are small coated cylinders; their thermal cycles were identical with thermal fatigue cycles. Moreover, the temperature evolu tion of the specimens was adjusted in order to correspond to the therm al evolution of the thermal fatigue coating cylinder. However, only a numerical simulation of these thermal shock experiments will allow to evidence the balance between this simplify test and the thermal fatigu e test. After 150 fatigue thermal cycles, no cracks were observed on t he surface of the cylinder. After 100 additional cycles, with a new hi gher temperature of 700 degrees C, no damage was evident on the surfac e. Residual stresses decrease quickly with thermal cycles; decreases i n hardness were also correlated with lower yield points. Several isoth ermal mechanical tests were performed on the coating material (low cyc le fatigue and relaxation) and on the support to determine their therm omechanical behaviour in the temperature range 20-700 degrees C. The p arameters were identified by means of a systematic numerical procedure (a computer program of simulation and identification). The microstruc tural evolution, during mechanical tests, was examined and related to the evolution of internal variables of the viscoplastic model. It was found that, at every temperature, the material showed cyclic softening behaviour. Moreover at 450 degrees C, a high stress level was correla ted with the evolution of the material microstructure and the increase of the kinematic variables of the model. The viscoplastic model chose n in this study permits a good description of, not only the macroscopi c mechanical results, but also of the microstructural evolutions of th e material, An anisothermal constitutive model could be introduced int o a numerical procedure; the latter, with the aid of a finite element method, should permit the calculation of the thermomechanical effects of thermal fatigue on structures such as cylinders. It should also per mit a comparison between the two tests of thermal fatigue and thermal shocks.