PSI - Issue 37

Junpeng Li et al. / Procedia Structural Integrity 37 (2022) 582–589 Junpeng Li/ Structural Integrity Procedia 00 (2021) 000 – 000

587

6

range of 30 ~ 60°, the median value is 45°, the angle of the crack surface and contact surface θ is in the range of 10.8~28.4°, the median value is 20°. When the number of test load cycles N is 5×10 4 , the contact surface has appeared longer oblique cracks and spalling off blocks, the surface crack length distribution in 0.10~0.35mm. When the number of test load cycles N is 8×10 4 , 1.2×10 5 and 1.6×10 5 , the surface crack length distribution ranges are 0.10~0.30mm, 0.10~0.35mm and 0.10~0.30mm respectively, the surface crack length distribution is basically the same, but the frequency distribution of crack length in different length intervals is different, when the number of load cycles is small, the cracks were basically distributed in the shorter length interval, and with the increase of the number of load cycles, the frequency of longer surface cracks showed a fluctuating development trend of increasing-decreasing-increasing. When N is 2.0×10 5 , about 35% of the crack length is greater than 0.30mm, at this time the surface of the specimen becomes rough and there are obvious wear dents; when N is 3.0×10 5 , about 14% of the crack length is greater than 0.30mm, at this time the surface of the specimen becomes rough and fat edges appear on both sides of the contact

surface, and serious plastic deformation occurs (2) Analysis of wear by test and simulation The wear growth of the twin-roller test and the simulation results were shown in Fig. 6.

Fig.6. Comparison of wear growth by the test and the simulation

It can be seen from the Fig. 6 that the wear depth and their trend were basically the same by the test and simulation, when the load cycles below 2.0 × 10 5 . After the load cycle was more than 2.0 × 10 5 , some difference between the results by the two methods occurred with the increase of the load cycles, which means that the shape of the contact surface of two specimens in test rig changed and increased significantly with the loading cycles accumulation. That would cause the wear increasing. (3) Crack initiation position The crack initiated at the sub-surface of the roller contact surface in the test, which was 0.034 mm deep to the contact surface, as shown in Fig. 7. The predicted crack initiated at the sub-surface of the roller contact surface, which was 0.031mm deep to the contact surface. The predicted crack initiation depth is basically consistent with the test results. The relative error between the two was about 9%.