13th International Conference on Fracture June 16-21, 2013, Beijing, China The dissipative and thermo-mechanical sources are coupled in Eq.(1)(thermo-mechanical coupling effects), although only the dissipative sources d1 are of interest in this work. The loading signal was used to compute and remove the thermo-mechanical sources Sth and associated temperature variations from Eq.(1). As a result, dissipative sources d1 were obtained. Eq.(1) was also integrated over the observed area Ωsp andΩdumto lower the measurement noise level. In this work, the mean dissipative sources were defined as ¯ d1 to underline the spatial averaging. The spatial average ¯ d1 is supposed to be representative of the dissipative sources field over the area Ωsp. This hypothesis has been validated by computing the fields of dissipative sources d1 using the method described in [9]. Eventually, the dissipated energy per cycle Ei d1 was computed by integrating ¯ d1 over each cycle: Ei d1 =∫ ti+1/f ti ¯ d1dt (2) Where ti is the starting time of cycle i. As the dissipated energy measurement is achieved in the elastic hysteretic domain, the dissipated energy should be constant during the test (15s). This characteristic was used a posteriori to check whether the test was performed in the elastic domain. The mean of the dissipated energy per cycle Em d1 was thus computed over a time tm=12s to lower the noise level (Fig.4): Em d1 = 1 tm∫ t0+tm/f t0 Ei d1 dt (3) 5 7 9 111315171920 0 1 2 3 4 Time(s) Ei d 1 (J.m−3.cycle−1) error bars ×104 t 0 Em d 1 t m =12s Figure 4: Definition of the mean dissipated energy per cycle Em d1 (316L specimen). The dissipated energy measurements have proved to be a reliable method[12]. With this method, the detection threshold of Em d1 is as small as ±222J·m−3·cycle−1 (2σEm d1 ) for a 316L stainless steel specimen. 5
RkJQdWJsaXNoZXIy MjM0NDE=