• 한국소성가공학회:학술대회논문집
• /
• 한국소성가공학회 2005년도 춘계학술대회 논문집
• /
• pp.287-290
• /
• 2005

Once hot forgings for automotive parts such as wheel bearing flange to which cyclic asymmetric bending stress is continuously applied are produced, it is necessary to control their microstructure to obtain superior mechanical properties. It is however hard to control the microstructure uniformly because the strength is reduced as coarsening of ferrite grains. To investigate the microstructural alteration according to process variables during hot working, the variation of the ferrite grain size was studied by utilizing of the computer aided servo-hydraulic Gleeble tester which is hot deformation behavior reproduction equipment. In addition, the effect of the ferrite grain size of raw material on the austenite grain behavior of hot forgings was also examined. The rolling contact fatigue resistance of the induction hardened SAE 1055 steel was compared with the occasion of the same condition of SAE52100 bearing steel. As a result, it was confirmed that the ferrite grain sizes of the forgings depend on the heating temperature and cooling start temperature during hot forging and cooling processes. The induction hardened SAE1055 steel showed a superior rolling contact fatigue resistance to the induction hardened SAE52100 steel. The reason is that SAE1055 steel is freer from the material defect such as segregation than the comparative steel.

• Journal of Advanced Marine Engineering and Technology
• /
• 제19권2호
• /
• pp.36-46
• /
• 1995

In this paper, fatigue crack propagation behaviors were investigated experimentally for the materials, carbon steel forgings (SF45A, SF50A, SF60A) which are used in the marine propeller shaft. The results obtained are as follows: The number of cycles required to grow crack length 1.30mm from microcrack initiation was about 60% of the total fatigue life. Fatigue crack propagation rate was expressed by the equation d(2a)/dN_B 2a/$N_f$ and the result was agreed well with the experimented data. And the equation d(2a)/dN=$C{\sigma}_a^m(2a)^n$ was evaluated also. Obtained material property m and n are 3~5 and 1-1.5 respectably, and the result was reasonably agreed to the data obtained from experiments.