Back to search

Democritus University of Thrace (DUTH)

Seismic retrofitting of damaged reinforced concrete columns with composites and other techniques

Abstract

dc:description

Greece is one of the most earthquake-prone countries. Therefore, many existing reinforced concrete (RC) buildings are quite likely to have been subjected to prior seismic excitation and may have suffered slight to severe damage due to previous seismic loading. Moreover, the majority of buildings are constructed taking into consideration older, obsolete seismic codes. Therefore, they probably have inherent insufficiencies and inadequate steel reinforcement detailing, such as short lap splices, sparse stirrups etc. In addition, many existing RC structures in coastal areas may suffer from damage due to the effect of corrosion agents. Damage accumulation could eventually lead the structure to exceed the limiting threshold capacity or even collapse.Greek retrofit code (KANEPE) proposes reduction factors for the mechanical characteristics of damaged elements without repair or strengthening. The code refers mainly to damage due to previous earthquake excitations. The aim of the present doctoral dissertation is the prediction of the seismic behaviour of damaged RC columns externally confined with fibre reinforced polymers (FRP). To address the problem, 261 experiments on columns of square or rectangular cross section subjected to cyclic loading (constant axial load and repeated cycles of increasing horizontal displacement) are initially collected from the international literature. An experimental database is created. It gathers the material properties and the features of the columns, as well as the results of the experiments. The database also includes the predicted values for chord rotation at failure (θu) and shear strength (VR) calculated according to retrofit codes applied in Greece [ΚΑΝΕPΕ, ΕC8.3] and models proposed by other researchers [Biskinis & Fardis 2013, Grammatikou et al. 2018b]. The columns are then divided into ten categories based on the existence of seismic-related damage, such as prior seismic excitations, corrosion of steel bars or inadequate lap splices of steel bars, and FRP confinement. These categories include i) columns with damage due to prior seismic loading, retrofitted with FRP, ii) columns with damage due to prior seismic loading and lap-spliced longitudinal bars, retrofitted with FRP, iii) columns with damage due to prior seismic loading and corroded steel reinforcement, retrofitted with FRP, iv) columns with lap-spliced longitudinal bars, v) columns with lap-spliced longitudinal bars, retrofitted with FRP, vi) columns with corroded steel reinforcement, vii) columns with corroded steel reinforcement, retrofitted with FRP, viii) columns with lap-spliced longitudinal bars and corroded steel reinforcement, ix) columns without prior damage and x) columns without prior damage, strengthened with FRP. The comparison of predicted values based on the existing design models with the experimental ones suggests that the divergence between predicted and experimental values is significant. Further evaluation and processing of the database results in modifications of existing design models [ΚΑΝΕPΕ, ΕC8.3]. Focusing on FRP confined columns, the steel hardening is taken into account by multiplying the yield stress by 1.25 (1.25·fy). This approach reduces the AAE for columns without damage, confined with FRP, to 16.17% for VR (the corresponding values for KANEPE and EC8.3 are 24.60% and 17.36%) and 21.48% for θu (the corresponding values for KANEPE and EC8.3 are 24.70% and 30.34%). As for columns with lap splices that are loaded up to failure (20% drop of their maximum horizontal load capacity) before repair and strengthening, the elaboration of the experimental database suggests that chord rotation at failure and shear strength are better predicted considering the residual shear strength. The prediction error is significantly reduced, especially for θu (AAE is 15.22%, while the corresponding values for KANEPE and EC8.3 are 39.93% and 29.84%). A detailed example is presented. For columns without lap splices loaded after yielding and before failure and then repaired and strengthened, an improved prediction of θu is provided taking into account the remaining deformations.Above findings reinforce the view that there is a need for further research to acquire more in-depth knowledge of the seismic behaviour of damaged RC columns retrofitted with FRP and the parameters that affect it. Three-dimensional (3D) finite element (FE) analyses can address gaps in knowledge, as well as save time and resources, and improve design for demanding seismic-resistant applications. Therefore, twenty-one critical cases of existing experimental tests are selected, in order to examine the effect of prior damage due to seismic loading, the corrosion level, the lap splice length and the FRP confinement. The column specimens are suitably modelled with 3D finite elements. The geometry and reinforcement details of the columns are explicitly reproduced, considering that the column is fixed to the bottom. For concrete, steel and composites, advanced material models are suitably calibrated and used. The columns are analysed pseudo-dynamically. Initially, an analysis is performed for the column R-NL-30 with damage due to previous seismic loading and then repaired and strengthened with FRP. A suitable model for pseudo-dynamic analysis with 3D finite elements is used. A history of loading, unloading and reloading of columns under cyclic loading can be reproduced with great accuracy. For the remaining twenty columns, pseudo-dynamic analyses are performed by imposing increased horizontal displacement. The lateral force-to-deformation envelope curves that are derived from the numerical results are satisfactorily close to the corresponding experimental curves. Therefore, the numerical results can be further utilized to enhance the experimental database and provide improved design models, following a hybrid research approach. The developed 3D FE models are extended to RC frames with or without infill walls and advanced seismic joints, with or without lap splices at the base of the columns. Finally, as chord rotation at failure and shear strength are considered, modifications of the existing relationships for FRP confined columns with previous seismic damage, for columns with corroded, lap-spliced steel bar reinforcements and for FRP confined columns with lap-spliced bars are proposed based on the numerical results. Specifically, for columns with previous seismic damage, without lap splices, the secant stiffness of the column before and after repair and strengthening is estimated based on the numerical results. The reduced concrete strength of the repaired column is calculated. An improved prediction of θu is provided using for first time this reduced concrete strength in combination with the remaining deformation. For corroded columns, the reduction of yield stress is correlated with the corrosion degree based on the numerical results. Subsequently, the application of the reduced yield stress and the reduced rebar diameter to the equations of the codes, are compared. It is suggested that the insertion of the crucial parameter of the reduced diameters to the equations of EC8.3 provides more reliable results.Regarding the RC columns with lap splices, a pseudo-ductile behaviour (with bar slippage) is determined for columns with inadequate lap splice length instead of the ductile behaviour (with bar yielding) of columns with adequate lap splice length. Moreover, the variable stress response of the steel bars along the lap splice region is retrieved and a variable plastic hinge length is identified. As for RC columns with lap splices and FRP confinement, a relationship derived from the elaboration of the numerical results is proposed, which modifies the stress of the steel in the lap region. It takes into account the lap-splice length and the confinement provided by the stirrups and FRP. It predicts if the longitudinal bars can reach their yielding or not and thus the efficiency of retrofitting the column with composite materials, in order to meet the code requirements. The minimum lap-splice length, according to the proposed approach, differs from the one provided by the seismic regulations. It improves the prediction of chord rotation at failure of deficient columns providing an average absolute error equal to 23.97%. The corresponding values for KANEPE and EC8.3 are 38.86% and 46.11%, respectively. The current doctoral dissertation provides a comprehensive hybrid approach for in-depth investigation of critical cases, revealing the effects of unidentified parameters that lead to improved design relations, as well as to important recommendations for future research.

Degree

thesis:*
Grantor dc:publisher
Democritus University of Thrace (DUTH)
Year dc:date
2023

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Anagnostou, Evgenia
  • Αναγνώστου, Ευγενία

Subjects

dc:subject × 32

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/53853

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Anagnostou, Evgenia; Αναγνώστου, Ευγενία. Seismic retrofitting of damaged reinforced concrete columns with composites and other techniques. Democritus University of Thrace (DUTH), 2023. http://hdl.handle.net/10442/hedi/53853