Issue 76

A. Huynh-Thai et alii, Fracture and Structural Integrity, 76 (2026) 99-116; DOI: 10.3221/IGF-ESIS.76.07

For Eqn. (18), the results fit across five cables, slopes b mean is nearly identical, indicating moderate baseline differences. With b>0 , tensile force decreases as X increases as shown in Tab. 8.

Figure 10: Median and 99% predictive band for loss factor ( η )- tensile force (N). The histogram shows the posterior distribution of μ a as shown in Fig. 11. The distribution is unimodal and approximately bell-shaped, centered near zero with a slightly heavier left tail. Its smooth, single-peaked shape, consistent with the prior trace plot, indicates good mixing and stationarity. The width encodes the uncertainty in the global mean of the average baseline tensile force across cables. As the model uses Eqn. (18), μ a corresponds to the baseline as 1/X → 0 (X →∞ ) ) ; a center near zero implies the global baseline is not strongly shifted in this parametrization. The absence of multiple modes suggests no competing posterior solutions for μ a .

Figure 11: The posterior distribution of mean μ a for for loss factor ( η )-tensile force ( N ). The Figs. 6, 8 and 10 present the posterior predictive distribution for the regression model where the shaded regions denote

112