PSI - Issue 72

Péter Ungár et al. / Procedia Structural Integrity 72 (2025) 265–269

268

Fig. 2. Failure probability comparison with measurements using optimized values for mini CT.

While the Beremin parameters are material and temperature dependent, they shouldn’t vary wildly on similar temperatures. As we can see from the above detailed optimized values, they are similar. For this reason, an effort was made to optimize these simultaneously. Fig. 3 shows the result of this, where a parameter combination of m = 20.9 and σ u = 2650 MPa gave a good agreement with the measurements for both temperatures.

Fig. 3. Failure probability comparison with measurements, using simultaneous optimisation for mini CT.

Optimized values were obtained for both 73W and JRQ materials, with questionable or non-existent agreement between the numerical calculations and measurements. While the optimization algorithm worked mathematically, mainly with linear regression, the discrepancies couldn’t be explained in the scope of the project requiring further analysis. These may result from not pure brittle cleavage, flow curves or the optimization strategy itself. Fig. 4 shows an example of converged solutions for miniCT for the JRQ material at temperatures -80°C and -75°C . Even though a parameter combination was obtained at -80°C the agreement was not acceptable. A valid pair of m = 27.62 and σ u = 2160 MPa was determined for JRQ material at -75 °C . Besides the reasons for discrepancies, a future discussion topic is the minimum required measurement points for optimization. For three temperature – material combinations only three-three data points were available, and converged solutions were obtained (15Kh2MFAA miniCT at -160°C: m = 45.13 and σ u = 2408MPa; normal CT at -100°C : m = 30.73 and σ u = 2569 MPa; -80°C: m = 32.9 and σ u = 2583 MPa) . Despite this, the physical viability of m = 45.13 at -160°C is questionable. In the case of 1T-CT the solution at -100°C was obtained with the maximum likelihood method, while for -80°C linear regression converged.