PSI - Issue 44

Salvatore Pagnotta et al. / Procedia Structural Integrity 44 (2023) 1909–1916 Author name / Structural Integrity Procedia 00 (2022) 000–000

1912

4

the test is shown in Fig. 4 (a). In general, there is a decreasing trend over time with a significant reduction from the initial preload. In detail, in the very early stages of the test, that is, at the moment when the slotted center plate starts to slide, there is a sudden loss of preload on all four bolts. This decrease is due to a settling of the bolt-disc spring load cell assembly due to the movement of the plates, which resulted in significant bolt loosening that was not compensated by the disc springs. As a result of this initial loss of preload, there is an asymptotic trend in the preload acting on the bolts to an average final value of about 15.5 kN, 35% lower than the initial value of 24 kN. As a result of the above analysis of the results, it can be stated that the change in sliding force that is recorded on the dissipative device during the test is mainly due to the change in preload acting on the bolts, the initial and final differences of the two quantities being similar in percentage terms. Starting from the final configuration of test 1, thus without disassembling or replacing the bolts, test 2 was carried out after applying the design preload of 24 kN to the bolts of the sliding friction connection. The force-displacement curve of test 2 is shown in Fig. 2 (b). Unlike what was seen before in test 1, there are no longer any significant changes in the sliding force in the initial phase. The initial sliding force is about 110 kN, while the final sliding force is 95 kN, a difference of about 15%, which is much less than previously recorded. Significant differences from the previous test can also be found for the trend of the friction coefficient over time (Fig. 3 (b)). In fact, it can be seen that the value remains approximately constant throughout the test, with small fluctuations and a slight decrease in the final part of the test itself. The initial value is 0.6, while the final value is 0.57, a difference of about 5%. This result highlights the ability of thermal sprayed aluminum to provide remarkable performance following a far greater number of cycles than required by EN15129. Once the outermost surface roughness flattens out and the surfaces tend to smooth out, the wear rate of thermal sprayed aluminum stabilizes as the number of cycles increases, providing a virtually unchanged friction coefficient during the test. Regarding the trend of preload on the bolts over time (Fig. 4 (b)), significant differences from Test 1 can be seen. In fact, the significant loss of preload recorded at the beginning of the test was not repeated in Test 2, although a loss on the order of 5% of the initial preload occurs in the initial stages when the slotted plate starts to slide. During the rest of the test, the average preload on the bolts tends to decrease asymptotically to a final average value of about 21.5 kN, which is almost 10% less than the initial value. Test 3 was carried out using plates made of brass as friction shims. Unlike aluminum, brass had very low surface roughness to the touch. The phenomenon of "stick and slip" is even more pronounced in the force-displacement curve shown in Fig. 6 (a), with significant fluctuations in the slip force during the application of displacement. In general, the slip force tends to increase as the number of cycles performed and the displacement achieved increases. The first cycle is characterized by a sliding force in the range of 35 kN, which reaches 62 kN at the end of the first five cycles with amplitude ± 7.5 mm. At the end of the test the sliding force reached values of about 85 kN, with a difference from the initial value of more than 140%. The explanation for the result shown in Fig. 5 (a) can be found by looking at the trend of the friction coefficient over time shown in Fig. 6 (a). The friction coefficient increases gradually, rising from 0.25 to a value of about 0.43 from the second half of the test onward. This increase is due to surface wear of the brass, which has led to roughening of the surfaces resulting in an increase in the friction coefficient. Regarding the preload acting on the bolts (Fig. 7 (a)), unlike what was observed previously for test 1, there is no substantial loss of preload at the beginning of the test. In addition, the average preload value of the bolts remained practically constant throughout the test, albeit with significant fluctuations in preload force. Based on this result, it is possible to say that the variations in sliding force are mostly due to the variations in friction coefficient between the contact plates. Also in the case of the brass plates, a second test was performed from the final configuration of test 3, again applying the design preload of 24 kN to the bolts of the sliding friction connection. The force-displacement curve of test 4 (Fig. 5 (b)) shows a significantly different trend from what was previously observed. In fact, the hysteresis cycles show a virtually constant sliding force over time, with initial value of 85 kN and final value of 78 kN (8% difference). The stability of the sliding force is a direct consequence of the stability of the friction coefficient, as can be seen in Fig. 6 (b). In fact, the value of friction coefficient expressed by the two contact plates, which is about 0.45, remains practically constant throughout the test. In analogy to what was seen previously, the trend of the preload acting on the bolts (Fig. 7 (b)), although characterized by considerable fluctuations in the values during sliding, shows an average value that is substantially constant and equal to 24 kN. From the results obtained, it is possible to state that, even in the case of brass plates, the results obtained during a second cyclic test are considerably better than those obtained with the first, since, during the first test, the surfaces of the plates change their roughness characteristics until they stabilize and remain constant during the second test.