Auto-tuning ensemble models for estimating shear resistance of headed studs in concrete

The shear resistance of headed studs is of paramount importance for the design of steel-concrete composite structures and an accurate predictive model is highly needed. Ensemble learning is expected to be a powerful solution while it relies on laborious selection of suitable hyper parameters. For efficiently predicting the resistance of headed studs, this work presented an auto-tuning ensemble learning-based strategy. It employed Sequential Model-Based Optimization method with Gaussian processes (GP) or Probabilistic random forests (PRF) as surrogate model to automatically explore the hyper-parameter configurations of ensemble learning algorithm-Light gradient boosting machine (LightGBM). To this end, the largest stud database to date of 1092 tests was established. The shear mechanisms of studs were analyzed and then integrated into the models via feature extraction and combination. The performance of GP-LightGBM and PRFLightGBM were assessed to outperform LightGBM and three standalone models, with PRFLightGBM being the most accurate. The superiority of PRF-LightGBM was further confirmed by comparison with the code equations in EC 4, AASHTO, GB 50017, and JSCE. Data-driven interpretation on PRF-LightGBM quantitatively revealed that the tensile capacity of stud shank and concrete performance are the most influential features on the shear resistance followed by the projected area of weld collar and longitudinal spacing of studs. Finally, an application StudATEML was created for efficient evaluation and practical design of headed stud connection.

Seismic behavior of lightly reinforced concrete squat shear walls

Christian Greifenhagen

This thesis addresses the seismic evaluation of existing buildings. In particular, it focuses on the seismic behavior of lightly reinforced shear walls that are not designed to withstand earthquake actions. A shear strength envelope for the assessment of deformation capacity of these non-ductile walls is presented. The approach is the result of experimental investigations and analytical modeling. Existing models for plastic hinges in beams are enhanced in order to determine drift capacity of lightly reinforced concrete shear walls. The static-cyclic behavior of non-ductile, reinforced concrete shear walls is investigated by testing four small-scale specimens of shear span ratio equal to 0.8. The design of the specimens includes reinforcement ratios, and axial force levels in existing shear wall buildings. Although the specimens were expected to fail in brittle shear, low to moderate ductile response is obtained. The deformation capacity, not the shear strength, is found to be restricted by shear failure. It is observed that inherent shear strength of concrete and the concrete compression zone are the principal contributors to the shear capacity of lightly reinforced shear walls. It is also observed that low reinforcement ratios and moderate levels of axial force can efficiently prevent brittle response in shear. The analytical model consists of a plastic hinge over the entire height of the low-rise shear wall. Proposals are made for the strain distribution inside the plastic hinge. Explicit relationships between drift and base shear are established and it is found that the model accurately predicts the envelope curve of static-cyclic loading. The shear strength envelope is formulated by using the analytical model. Criteria for the failure modes of diagonal tension, of concrete crushing, and of sliding enclose the shear strength envelope. In addition, inherent shear strength forms the lower bound of this envelope. The contributions of reinforcement and concrete to shear capacity are formulated in terms of initial strength and strength decay. Accurate prediction of both the ductility supply and the drift capacity obtained in static-cyclic tests is observed. Validation of the shear strength envelope on full-size walls prevalent in existing buildings shows potential for further application. The proposal contributes to more realistic evaluation of shear strength in selected situations where available methods are too conservative. Hence, it allows for both avoiding costly seismic strengthening in such situations and better allocation of resources where they are really needed.

EPFL2006

Structural Response of R-UHPFRC - RC Composite Members Subjected to Combined Bending and Shear

Talayeh Noshiravani

The addition of a thin overlay of Ultra-High Performance Fibre Reinforced Concrete (UHPFRC) to Reinforced Concrete (RC) members is an emerging technique to strengthen and protect existing structures and to design durable new structures. Combining UHPFRC with closely spaced, small-diameter steel rebars in Reinforced UHPFRC (R-UHPFRC) layers improves the UHPFRC's strain hardening behaviour. For reasons of practicality, R-UHPFRC layers are cast or glued (in the case of prefabricated elements) on top of RC members, thus changing the latter into R-UHPFRC - RC composite members. The high strength and deformation capacity of R-UHPFRC elements make them a suitable external flexural reinforcement for RC members over intermediate supports, e.g., bridge decks and slabs or beams in buildings. Over reinforcement of RC beams and slabs with tensile flexural reinforcement can result in their shear failure at either a lower resistance or deformation than the associated values for member failure in flexure. A comprehensive experimental program was conducted to study the flexure-shear behaviour of R-UHPFRC - RC composite beams. The program comprises two test series on cantilever beams and continuous beams. The test parameters include shear span-depth ratio (a/d), the amount of transverse reinforcement ( ρν), the amount of longitudinal reinforcement, and the strength and bond condition of the R-UHPFRC rebars. The experimental results reveal the different failure modes of R-UHPFRC - RC composite members and the contribution of the R-UHPFRC elements to the member resistance, ductility and capacity to redistribute the internal stress. It was shown that in R-UHPFRC - RC beams with ribbed rebars and a shear span to depth ratio greater than 2.5 the stresses are carried by beam action. Depending on the degree of longitudinal reinforcement, all but two of the beams with 3.0≤a/d≤3.4 and ρν≤0.17 had a flexure-shear failure; the rest failed in flexure. The flexure-shear failure of the composite beams was at an approximately equal rotation level as their RC reference beam but at a resistance 2.3 times that of the RC beam. This is due to (1) the debonding interface zone between the elements that allows the R-UHPFRC - RC beams to rotate more freely and (2) the out-of-plane resistance of the R-UHPFRC element that contributes to the shear resistance. The internal flow of forces and the structural response of composite members strongly depend on the bond condition between the R-UHPFRC and RC, the UHPFRC and its rebars, as well as the concrete and its rebars. Cracking of the concrete along the interface zone causes bond reduction, i.e., softening of the shear connection, between the two elements. In presence of high shear stresses and diagonal flexure-shear cracks, interface zone softening is observed between the elements prior to the maximum resistance, while UHPFRC is strain hardening. The cause of this softening behaviour is the prying action due to the relative rotational movement of the RC rigid bodies separated by the flexure-shear cracks. Static and kinematic solutions of the theory of plasticity for RC beams are extended to predict the collapse load of R-UHPFRC - RC composite beams at the ultimate limit state. A mechanical model for predicting the structural response of composite beams is proposed. In combination with truss models, the concept of an R-UHPFRC - RC plastic hinge is introduced to calculate the force-displacement response of composite beams. The failure criterion based on the collapse mechanisms (kinematic solutions) sets the limit of the force-displacement response. The model is corroborated by the experimental results. This model provides a tool for analysis of RC members reinforced with an added tensile R-UHPFRC element.

EPFL2012

Assessment of Shear Strength for Existing Bridges with Low Amounts of Shear Reinforcement

Miguel Fernández Ruiz, Aurelio Muttoni, Michael Markus Rupf

Design of concrete girder bridges has significantly evolved during the last decades. This has been particularly relevant with respect to shear design, motivated by changes in actions and design models. As a consequence, assessing the shear strength of existing bridges leads in many cases to unsatisfactory safety levels. Furthermore, many existing bridges do not comply with current code regulations with respect to minimum amounts of shear reinforcement. This situation can lead to expensive retrofitting and strengthening of a significant number of existing bridges. The assessment of the shear strength of these bridges has thus become a significant task for structural engineers. Design codes are not always appropriate for assessing the strength of existing bridges. They propose safe models providing sufficient accuracy for design of a wide number of structures. However, these models do not account for some particularities of prestresses bridges and neglect a number of shear-transfer actions that can be relevant for their strength. This is typically the case of shear carried by the inclination of the compression chord, the increase of the stress in the tendons or the effective strength of concrete in the web of cracked prestressed girders. Accounting for more realistic approaches for these parameters may significantly increase the estimated strength of a structure with respect to code provisions, avoiding in many cases unnecessary retrofitting or strengthening. In this paper, the results of a tests campaign carried out at Ecole Polytechnique Fédérale de Lausanne on 10 prestressed concrete girders (10 meters long, 0.78 m high) are outlined. The specimens were provided with very low amounts of shear reinforcement and some of them present defective stirrup anchorage to simulate realistic conditions of existing structures. The experimental results are discussed with reference to the stress field method where the role of the different shear-transfer actions is investigated and compared to current code provisions.

fib Symposium Tel-Aviv 20132013