PSI - Issue 78

Sabatino Di Benedetto et al. / Procedia Structural Integrity 78 (2026) 1697–1704

1703

Even in this case, the Model Code 1990 provides the closest prediction, with a deviation of 27%. This highlights that the bond resistance does not result from a uniform tangential stress distribution along the interface, but rather from a more complex interaction that must be accurately captured to match experimental results.

Table 3. Design bond shear stress (  Rd ) according to different codes  Rd Embedded length (mm) f ck (MPa) NTC2018=Eurocode 2 Eurocode 4 Model Code 1990 300 50 4.28 0.55 2.12 200 32 3.18 0.55 1.70

Table 4. Evaluation of the bond resistance according to different codes NTC2018=Eurocode 2 Eurocode 4 Model Code 1990 Numerical model in OpenSees Experimental test

F max (kN)

825 2.50

118 0.36

420 1.27

325 0.98

330

F max,codes /F max,experimental

-

5. Parametric analysis The validated numerical model developed in OpenSees was used to conduct a parametric analysis by varying the tube diameter and the concrete strength class (in the assumption now that the plinth is made of the same concrete class for the full height), while the thickness of the tube and the length of the embedded part were fixed equal to 8 mm and 500 mm, respectively. The corresponding results are presented in Table 5.

Table 5. Bond resistance (kN) developed by the cases examined in the parametric analysis Concrete class D (mm) C20/25 C25/30 C30/37 C35/45 C40/50 C45/55 C50/60 101.6 176.2 192.7 207.1 219.7 231.0 241.3 250.7 114.3 198.6 217.2 233.4 247.7 260.5 272.1 282.7 139.7 243.3 266.2 286.1 303.8 319.5 333.8 346.9 168.3 293.6 321.4 345.5 366.8 385.9 403.2 419.1 177.8 310.4 339.7 365.2 387.8 408.0 426.3 443.0 193.7 338.4 370.4 398.2 422.8 444.9 464.9 483.2 219.1 383.1 419.3 450.9 478.8 503.8 526.5 547.3 244.5 427.8 468.3 503.5 534.8 562.8 588.1 611.4 273 478.0 523.3 562.7 597.6 628.9 657.3 683.3 323.9 567.6 621.4 668.2 709.8 747.1 780.8 811.7 355.6 623.4 682.5 734.0 779.6 820.6 857.7 891.7 406.4 712.8 780.5 839.3 891.6 938.5 980.9 1019.9

Based on these outcomes, a predictive formulation (Eq. 4) was developed to estimate the bond resistance of composite micropiles embedded in concrete plinths. The equation, derived through regression analysis, proved to be accurate: when applied to the parametric cases, it generated a mean predicted-to-computed force ratio of 1.02 and a coefficient of variation of 3%. The equation shows that, if the bond stress resistance is taken as 0.24√ instead of 0.3√ (as currently suggested by Model Code 1990), the simplified formulation still provides a reliable estimate of the bond resistance. = ∙ 0.24√ (4) 6. Conclusions This paper investigates the bond resistance of axially loaded concrete-filled steel tubes embedded in reinforced concrete plinths, a topic of particular relevance as it reflects the load transfer mechanisms typically encountered in composite micropiles used in foundations. To address this, a review of the literature was conducted to identify existing modelling approaches for this type of interaction. Based on a real case of a footing with a composite micropile, a full-scale specimen was designed and tested through a pull-out experiment carried out at the University of Salerno. As expected, failure occurred at the interface between the steel tube and the concrete of the plinth. Using the experimental data, a numerical model of the specimen was developed in OpenSees. Special focus was given to accurately reproducing the bond stress–slip behaviour at the interface, using the nonlinear relationship