PSI - Issue 8

12 10 L. Landi et al. / Procedia Structural Integrity 8 (2018) 3–13 L. Landi et al. / Structural Integrity Procedia 00 (2017) 000 – 000 5.1 Simulations with = 0.6 − 0.5 On table 8, the results of impact simulation with DC01 RMS model and = . are presented. Table 8. Numerical experimental correlation of impacts with DC01 RMS model ε f = 0.6 Projectile mass [kg] v i [m/s] Mewes test v r [m/s] Numerical result 0.625 69.28 B -3,26 B 80 P 27,42 P 1.25 63.25 B -6,74 B 66.93 P 6,29 P 2.5 55.14 B n.p. B 56.92 P -6,87 B B= bulging P=complete penetration The 2.5 kg projectile with v i = 55.14 m/s simulation is not executed because the bulging as a result is obvious from v i =56.92 m/s numerical simulation results. On table 8 the results of impact simulation with DC03 RMS model and = . 5 is presented. Table 9. Numerical experimental correlation of impacts with DC01 RMS model ε f = 0.5 Projectile mass [kg] vi [m/s] Mewes test v r [m/s] Numerical result 0.625 69.28 B -3,25 B 80 P 49,38 P 1.25 63.25 B -6,64 B 66.93 P 19,28 P 2.5 55.14 B 18,83 P 56.92 P 25,45 P B= bulging P=complete penetration From the results shown on tables 8 and 9 we can conclude that the RMS model of DC01 is able to represent the behavior of the experimental tests of figure 4. A correct ε f = 0.5 − 0.6 is expected to achieve a result very close to experiments. It must be said that for heavier projectile the difference between critical and penetration limit is very small and the correlation between numerical and experimental tests have to be done using more reliable data such as a full regression of a proper curve of Recht Ipson as shown on Recht and Ipson (1963). Nowadays safety of machines to “impacting objects” of various kind is assured by international standards based on normative annexes of type B and C standards specifically developed for this purpose. Those standardized tests and procedures introduced on late 90s gives as a results a higher level of protection for safety guards but the interpretation of the experimental results is not always easy. By repeating the same tests in fact, for example by varying the supplier of the same standardized material, whose supply characteristics can be different, the results are not entirely reproducible due to great statistical dispersion of the tests. Moreover, as for the standardized test is made, one can not extend the results of tests carried for different “experimental fields” than those explicitly tested (for example, different masses, different shape and speed of the penetrator). The numerical and experimental research carried out by the authors in recent years have led to uncertainty values of the FE tests absolutely in line with uncertainty due to the nature of the experimental tests. All the numerical simulations and all experimental correlations have shown that, under limited and known variability of velocity mass 6. Conclusions