Assessment and Strengthening of Partitions in Buildings

Synonyms

Infills; Infill walls; Masonry infills; Reinforced concrete infills

Introduction

Partitions in buildings are necessary architectural features that separate spaces in order to facilitate various functions that are required depending on the use of a building. These partitions in modern structures can be lightweight, for example, in case that they are used to subdivide office spaces, or can be heavy masonry partitions that are used mainly in reinforced concrete structures to subdivide the plan area of a building into various rooms. In the latter case, these are mostly built within the frame of the structure filling the gap, and they are therefore called infills or infill walls. Depending on the type of the material that is used to construct them, they can be called masonry infills or reinforced concrete (RC) infills.

Infill walls have attracted the attention of many researchers since the early 1950s, and much work has been undertaken to study their behavior and interaction with the surrounding frames. In addition, efforts have been made to utilize infill walls as a means of producing economic designs by reducing the sizes of the members of the bounding frames.

A large number of researchers have studied the behavior of infilled frames. It is evident from their studies that infill walls can provide both an economic and practical means for the lateral stability of framed structures and a viable alternative for retrofitting existing structures to resist seismic, wind, and blast loads. Despite this, there is reluctance from the engineering community to use this structural system widely and treat infill walls as structural elements.

Therefore, although the subject of infilled frames has been studied for more than 60 years, there are still no definitive answers either about their behavior and interaction with the bounding frame or about the estimation of their stiffness and strength. Some problems that make it difficult for practicing engineers to use infill walls as structural elements are the inherent nonlinearity, the high degree of variability, and the inherent degradability of infill walls. The fact that infill walls exhibit completely different in-plane and out-of-plane behavior makes the problem even more difficult to tackle. This reveals the difficulties associated with this problem and the need for more research to answer these long-standing questions.

Nevertheless, the presence of infill walls has shown that in most cases they have beneficial effects on the behavior of structures during earthquakes and have contributed in the prevention of collapse of many structures. It is therefore for this reason that efforts have been made to use engineered infills to retrofit existing seismic-deficient buildings, since they present an economically viable solution.

In this entry, the in-plane behavior and in-plane failure modes of infilled frames are presented. Then, macro- and micromodels for infill walls are presented followed by the results of experimental investigations for the replacement of masonry infill walls with reinforced concrete infills, as an economically viable method for retrofitting existing structures. Finally, methods for strengthening existing partition walls are presented.

Behavior of Partitions

The modeling of the behavior of infilled frames under lateral loading (and mainly earthquake-induced loads) is a complex issue because these structures exhibit a highly nonlinear response due to the interaction between the masonry infill panel and the surrounding frame. This results to several modes of failure, each of which has a different failure load, and hence a different ultimate capacity and overall behavior.

At moderate loading levels, the infill of a non-integral infilled frame separates from the surrounding frame and the infill acts as a diagonal strut (Fig. 1). As the racking load is increased, failure occurs eventually in either the frame or the infill. The usual mode of frame failure results from tension in the windward column or from shearing in the column or beams or plastic hinging in columns or beams; however, if the frame strength is sufficient enough to prevent its failure by one of these modes, the increasing racking load eventually produces failure of the infill. In the most common situations, the in-plane lateral load applied at one of the top corners is resisted by a truss formed by the loaded column and the infill along the diagonal connecting the loaded corner and the opposite bottom corner. The state of stress in the infill gives rise to a principal compressive stress along the diagonal and a principal tensile stress in the perpendicular direction. If the infill is made of concrete, successive failures, initially by cracking along the compression diagonal and then by crushing near one of the loaded corners or by crushing alone, will lead to collapse; if the infill is made of brick masonry, an alternative possibility of shearing failure along the mortar planes may arise (Fig. 1).

Assessment and Strengthening of Partitions in Buildings, Fig. 1

Failure modes of non-integral infilled frames: (a) Corner crushing (CC) and diagonal compression (DK) modes. (b) Sliding shear (SS), frame failure (FF), and diagonal cracking (DK) modes (Asteris et al. 2011 with permission from ASCE)

Based on both experimental and analytical results during the last five decades, different failure modes of masonry-infilled frames were proposed that can be classified into five distinct modes given below:

• 1. The corner crushing (CC) mode, which represents crushing of the infill in at least one of its loaded corners, as shown in Fig. 1a. This mode is usually associated with infilled frames consisting of a weak masonry infill panel surrounded by a frame with weak joints and strong members (Mehrabi and Shing 1997; El-Dakhakhni 2002; Ghosh and Amde 2002; El-Dakhakhni et al. 2003).

• 2. The diagonal compression (DC) mode, which represents crushing of the infill within its central region, as shown in Fig. 1a. This mode is associated with a relatively slender infill, where failure results from out-of-plane buckling of the infill.

• 3. The sliding shear (SS) mode, which represents horizontal sliding failure through bed joints of a masonry infill, as shown in Fig. 1b. This mode is associated with weak mortar joints in the infill and a strong frame.

• 4. The diagonal cracking (DK) mode, which is seen in the form of a crack across the compressed diagonal of the infill panel and often takes place with simultaneous initiation of the SS mode, as shown in Fig. 1b. This mode is associated with a weak frame or a frame with weak joints and strong members infilled with a rather strong infill (Mehrabi and Shing 1997; El-Dakhakhni 2002).

• 5. The frame failure (FF) mode, which is seen in the form of plastic hinges developing in the columns or the beam-column connections, as shown in Fig. 1b. This mode is associated with a weak frame or a frame with weak joints and strong members infilled with a rather strong infill.

Ghosh and Amde (2002), based on the finite element method and including interface elements at the frame-infill interface, confirmed the order of occurrence of the above five distinct failure modes. Of the five modes, only the CC and SS modes are of practical importance (Comité Euro-International du Béton CEB 1996), since most infills are not slender (El-Dakhakhni et al. 2003) and therefore the second mode (DC) is not favored. The fourth mode (DK) should not be considered as a failure mode, due to the post-cracking capacity of the infill to carry additional load. The fifth mode (FF) relates to the failure of the frame, and it is particularly important when examining existing structures, which in many cases exhibit frame weakness. It should be noted that these failure modes are only seen/applicable to the case of infill walls without openings on the diagonal of the infill panel.

Kappos and Ellul (2000), based on an analytical study of the seismic performance of masonry-infilled RC-framed structures, found that taking into account the infill in the analysis resulted in an increase in stiffness as much as 440 %. It is clear that, depending on the spectral characteristics of the design earthquake, the dynamic behavior of the two systems dealt with by the author (bare vs. infilled frame) can be dramatically different. They also presented a very useful global picture of the seismic performance of the studied infill frames by referring to the energy dissipated by each component of the structural system. It is clear that at the serviceability level, over 95 % of the energy dissipation is taking place in the infill walls (subsequent to their cracking), whereas at higher levels, the RC members start making a significant contribution. This is a clear verification of the fact that masonry infill walls act as a first line of defense in a structure subjected to earthquake load, while the RC-frame system is crucial for the performance of the structure to stronger excitations (beyond the design earthquake).

The quantification of the in-plane properties of non-integral infilled frames and the prediction of the failure modes is a rather cumbersome task even when the load is applied monotonically. The problem becomes even more difficult when a dynamic load is applied, since the hysteretic behavior of the materials and their deterioration with the cyclic loading should be taken into consideration. Chrysostomou and Asteris (2012) discuss these issues and present methods for the determination of the in-plane stiffness, strength, and deformation capacity of infills along with the results of a parametric study that compares these methods and checks them against experimental results whenever possible. Based on the above material, recommendations are made for the in-plane material properties, failure modes, strength and stiffness, as well as deformation characteristics, of infilled frames.

Modeling of Partitions

In order for engineers to be able to use infills as an engineering element, a reliable mathematical model needs to be developed that simulates the combined behavior of the infill partition and the bounding frame. As explained in the previous section, the behavior of infilled frames is rather complex and it depends on a large number of parameters. Therefore, for the mathematical model to accurately simulate the behavior of infilled frames, it should take into consideration all these parameters.

The models for infills can be subdivided into macro- and micromodels. In the former, simple models are developed that are simulating the overall global behavior of infilled frames, while in the latter, surface or solid finite elements are used along with interface ones that simulate both the local and global behavior of both the infills and the bounding frames and their interaction. In the following two sections macro- and micromodels are presented.

Macromodeling

Since the first attempts to model the response of the composite infilled frame structures, experimental and conceptual observations have indicated that a diagonal strut with appropriate geometrical and mechanical characteristics could possibly provide a solution to the problem. Early research on the in-plane behavior of infilled frame structures undertaken at the Building Research Station, Watford (later renamed Building Research Establishment and now simply BRE), in the 1950s served as an early insight into this behavior and confirmed its highly indeterminate nature in terms solely of the normal parameters of design. On the basis of these few tests, a purely empirical interaction formula was later tentatively suggested for use in the design of tall framed buildings.

Diagonal Strut Model

In the early 1960s, Polyakov suggested the possibility of considering the effect of the infilling in each panel as equivalent to diagonal bracing, and this suggestion was later adopted by Holmes, who replaced the infill by an equivalent pin-jointed diagonal strut made of the same material and having the same thickness as the infill panel and a width defined by

$$ \frac{w}{d}=\frac{1}{3} $$

(1)

where d is the diagonal length of the masonry panel. The “one-third” rule was suggested as being applicable irrespective of the relative stiffness of the frame and the infill. One year later, Stafford Smith, based on experimental data from a large series of tests using masonry-infilled steel frames, found that the ratio w/d varied from 0.10 to 0.25. On the second half of the 1960s, Stafford Smith and his associates using additional experimental data related the width of the equivalent diagonal strut to the infill-frame contact lengths using an analytical equation, which has been adapted from the equation of the length of contact of a free beam on an elastic foundation subjected to a concentrated load. They proposed the evaluation of the equivalent width λ _h as a function of the relative panel-to-frame-stiffness parameter, in terms of

$$ {\lambda}_h=h \sqrt[4]{\frac{E_w{t}_w \sin 2\theta }{4EI{h}_w}} $$

(2)

where E _w is the modulus of elasticity of the masonry panel, EI is the flexural rigidity of the columns, t _w is the thickness of the infill panel and equivalent strut, h is the column height between centerlines of beams, h _w is the height of infill panel, and θ is the angle, whose tangent is the infill height-to-length aspect ratio, being equal to

$$ \theta ={ \tan}^{-1}\left(\frac{h_w}{L_w}\right) $$

(3)

in which L _w is the length of infill panel (all the above parameters are explained in Fig. 2).

Assessment and Strengthening of Partitions in Buildings, Fig. 2

Definitions for the equivalent diagonal strut (Asteris et al. 2011 with permission from ASCE)

The use of this equation to seismic design is recommended for a lateral force level up to 50 % of the ultimate capacity.

Based on experimental and analytical data, Mainstone proposed an empirical equation for the calculation of the equivalent strut width, given by

$$ \frac{w}{d}=0.16{\lambda}_h^{-0.3} $$

(4)

Mainstone and Weeks and Mainstone, also based on experimental and analytical data, proposed a slightly modified empirical equation for the calculation of the equivalent strut width:

$$ \frac{w}{d}=0.175{\lambda}_h^{-0.4} $$

(5)

This formula was included in FEMA-274 for the analysis and rehabilitation of buildings as well as in FEMA-306, as it has been proven to be the most popular over the years. Although this equation is accepted by the majority of researchers dealing with the analysis of infilled frames, several variations were presented by various researchers trying to improve its applicability. A discussion and comparison of these proposals as well as references to the work of the various researchers mentioned above is given by Chrysostomou and Asteris (2012), and Asteris et al. (2011).

Multiple-Strut Models

In the last two decades, it became clear that one single-strut element is unable to model the complex behavior of the infilled frames. The bending moments and shear forces in the frame members cannot be adequately represented using a single diagonal strut connecting the two loaded corners. More complex macromodels were hence proposed, still typically based on a number of diagonal struts.

Thiruvengadam proposed the use of a multiple-strut model to simulate the effect of the infill panel. This particular model consists of a moment-resisting frame with a large number of pin-jointed diagonal and vertical struts. Initially, a perfect frame-infill bond condition is assumed, and the lateral stiffness of the infill, by its shear deformation, is modeled by a set of pin-ended diagonal struts running in both directions. These diagonals represent the shear and axial stiffness of the masonry infill. Similarly, the vertical stiffness contribution is accounted for by providing vertical struts. The objective of the aforementioned study was a realistic evaluation of the natural frequencies and modes of vibration, purposes for which the nonlinear phenomena do not play an important role. This model has been adopted by many researchers to investigate the effect of infill on the behavior of infilled frames and has been also included in FEMA-356, due to the great number of struts, as a method for modeling the special case of infilled frames with openings. Similarly, Hamburger and Chakradeo proposed a multiple-strut configuration that can account for the openings also, but the evaluation of the characteristics of the struts is rather complicated. They showed that for panels of typical configuration, the formation of these struts protects the beam-to-column connections, which have limited capacity in withstanding significant flexural demand, with plastic hinges forming instead within the mid-span region of the beam. They postulated that this resulted in a system of significant strength, stiffness, and ductility that behaves much like the modern eccentrically braced frame systems. Such behavior could, in part, be responsible for the observed good performance of these buildings in the 1906 San Francisco earthquake (Hamburger and Meyer 2006).

The main advantage of the multiple-strut models, in spite of the increase in complexity, is the ability to represent the actions in the frame more accurately. Syrmakezis and Vratsanou (1986) employed five parallel struts in each diagonal direction. It was stressed how different contact lengths play a significant effect on the bending moment distribution in the frame members.

Chrysostomou (1991, Chrysostomou et al. 2002) aimed at obtaining the response of infilled frames under earthquake loading by taking into account both stiffness and strength degradation of infills. Each infill panel was modeled by six compression-only inclined struts (Fig. 3). Three parallel struts were used in each diagonal direction and the off-diagonal ones were positioned at critical locations along the frame members. These locations are specified by a parameter α, which represents a fraction of the length or height of a panel and is associated with the position of the formation of a plastic hinge in a beam or a column. At any point during the analysis of the nonlinear response, only three of the six struts are active, and the struts are switched to the opposite direction whenever their compressive force reduced to zero.

Assessment and Strengthening of Partitions in Buildings, Fig. 3

Six-strut model for masonry infill panel in frame structures (Chrysostomou 1991)

In order to conduct nonlinear analysis, the force-displacement relationships corresponding to the equivalent strut model must be adequately defined. The modeling of hysteretic behavior increases not only the computational complexity but also the uncertainties of the problem.

In Chrysostomou’s model, the hysteretic behavior of the six struts is defined by a hysteretic model, which consists of two equations. The first equation defines the strength envelope of a structural element and the second defines its hysteretic behavior. The shape of the envelope and the hysteretic loops is controlled by six parameters, all of which have physical meaning and can be obtained from experimental data. The advantage of this strut configuration over the single diagonal strut is that it allows the modeling of the interaction between the infill and the surrounding frame and it takes into account both strength and stiffness degradation of the infill, which is vital for determining the response of infilled frames subjected to earthquake load.

Saneinejad and Hobbs developed a method based on the equivalent diagonal strut approach for the analysis and design of steel or concrete frames with concrete or masonry infill walls subjected to in-plane forces. The proposed analytical model assumes that the contribution of the infill panel to the response of the infilled frame can be modeled by replacing the panel by a system of two diagonal masonry compression struts. The method takes into account the elastoplastic behavior of infilled frames considering the limited ductility of infill materials. Various governing factors such as the infill aspect ratio, the shear stresses at the infill-frame interface, and relative beam and column strengths are accounted for in this development. This model has been adopted by Madan et al. (1997) for static monotonic loading, as well as quasi-static cyclic loading. This model for masonry infill panels was implemented in IDARC 2D Version 4.0 (Vales et al. 1996), a computer-based analytical tool for the inelastic analysis and damage evaluation of buildings and their components under combined dynamic, static, and quasi-static loading.

El-Dakhakhni (2000, 2002) and El-Dakhakhni et al. (2001) suggested replacing the infill wall by one diagonal and two off-diagonal struts, on making use of the orthotropic behavior of the masonry wall as well as on some experimental observations and analytical simplifications, in order to simplify the nonlinear modeling of these structures.

Crisafulli investigated the influence of different multiple-strut models on the structural response of infilled reinforced concrete frames, focusing on the stiffness of the structure and the actions induced in the surrounding frame. Numerical results, obtained from the single-, two-, and three-strut models, were compared with those corresponding to a refined finite element. The lateral stiffness of the structure was similar in all the cases considered, with lower values for two- and three-strut models. It must be noted that, for the multiple-strut models, the stiffness may significantly change depending on the separation between struts. The single-strut model underestimates the bending moment because the lateral forces are primarily resisted by a truss mechanism. On the other hand, the two-strut model leads to higher values than those corresponding to the finite element model. A better approximation is obtained from the three-strut model, although some differences arise at the ends of both columns. Although the single-strut model constitutes a sufficient tool for the prediction of the overall response and the triple-strut model is superior in precision, Crisafulli adopted the double-strut model approach, accurate enough and less complicated compared to the other models.

More recently, Crisafulli and Carr (2007) proposed a new macromodel in order to represent, in a rational but simple way, the effect of masonry infill panels. The model is implemented as a four-node panel element which is connected to the frame at the beam-column joints. Internally, the panel element accounts separately for the compression and shear behavior of the masonry panel using two parallel struts and a shear spring in each direction. This configuration allows an adequate consideration of the lateral stiffness of the panel and of the strength of masonry panel, particularly when a shear failure along mortar joints or diagonal tension failure is expected. Furthermore, the model is easy to apply in the analysis of large infilled frame structures. The main limitation of the model results from its simplicity, since the panel is connected to the beam-column joints of the frame, being thus not able to properly predict the bending moment and shear forces in the surrounding frame. A detailed presentation as well as references to the work on macromodels of the various researchers mentioned above is given by Asteris et al. (2011).

Micromodeling

The ability to model accurately the behavior of a structural system, including the infilled RC frame, depends on experimental work. A review of the available published experimental data describing the effect of infill walls on the overall structural response of RC frames finds significant scatter in the obtained results, due to the large number of uncertainties involved in the various investigations carried out to date. This data can describe qualitatively the effect of infill walls, but it is not yet possible to quantify this effect experimentally. Since the formulation and validation of mathematical (numerical or analytical) models is based on available experimental data, the validity of the yield predictions of current software packages is currently questionable.

Due to the difficulties, limitations, and high costs associated with the testing of infilled RC frames, resort is frequently made to the use of nonlinear finite element analysis (NLFEA). The use of the FE method can provide a more detailed description of the effect of infill walls on the response of infilled frames. At the same time, it allows the investigation to be extended to structural forms more complex than the simple structural elements that are usually studied experimentally (i.e., scaled models of one- or two-story infilled frames with column height to beam length equal to values from 0.75 to 1.25). To qualify for accurately capturing the behavior of a concrete structure, a FEA package must predict the structural response within an error of up to 20 %. Such a package should also employ appropriate constitutive models and adopt nonlinear numerical procedures capable of capturing the force redistribution and cracking processes of brittle materials.

To this end, constitutive relationships must be defined for the use of finite element models of infilled frames. The use of 3D solid, instead of linear, elements in constitutive models requires a considerably higher level of model sophistication. Models of concrete behavior are based either on regression analyses of experimental data (empirical models) or on continuum mechanics theories, which should also be verified against experimental data. Many such models have been proposed, but the application of FE packages in practical structural analysis has shown that the majority of constitutive relationships are case dependent, since the solutions obtained are realistic only for specific types of problems. The application of these packages to a different set of problems requires modification, sometimes significant, of the constitutive relationships. The situation is better for the reinforcement. However, complications arise with the introduction of bond-slip laws, which results in large discrepancies in predicted behavior.

Regarding the modeling of the infill panels, the great number of influencing factors, such as dimension and anisotropy of the bricks, joint width and arrangement of bed and head joints, material properties of both brick and mortar, and quality of workmanship, make the simulation of plain brick masonry extremely difficult. The level of complexity of the analytical model depends on whether masonry is considered as one-, two-, or three-phase material, where the three phases are comprised of the units, mortar, and unit-mortar interface.

Another level of complexity is added with the modeling of the infill-frame interface, which determines the boundary conditions of the infill and its interaction with the bounding frame. Springs and interface elements have been used by various researchers to represent the interaction between deformable structures such as these, where interfacial separation and sliding may arise.

Infill walls have also been represented by boundary elements. These are combined with FE, which can represent the bounding frame. This type of model reduces computational costs considerably, since the infill is represented by a single boundary element. However, the applicability of these models is limited to the elastic behavior of the infill; once cracks are initiated, boundary elements hold a disadvantage as they can be implemented only with extreme difficulty in nonlinear problems.

Mesh sensitivity poses a further challenge (Fig. 4). Very dense meshes are often thought to provide better results, but this is not always true when brittle materials are modeled. The size of the 3D brick FEs used for modeling concrete continua should be equivalent to the size of the cylinders used in testing. These are assumed to constitute a “material unit” for which average material properties are obtained. Hence, the volume of these specimens provides a general guideline for the size of the FE that should be used for the modeling of concrete structures. Furthermore, each Gauss point of a FE must correspond to a volume at least three times the size of the largest aggregate used in the concrete mix, in order to provide a realistic representation of concrete, rather than a description of its constituent materials.

Assessment and Strengthening of Partitions in Buildings, Fig. 4

Finite element mesh for a non-integral infilled frame (Reprinted from Asteris et al. 2013 with permission from Elsevier)

Regarding the crack formation, the smeared-crack approach, rather than the discrete-crack approach, is usually adopted for the modeling of the cracking. With this approach, one may avoid the complexities linked to remeshing, which is required by the discrete-crack approach.

The discrete element method (DEM) originated from the studies of fractured rock masses and is also used for the modeling of infill walls. Due to its capability of representing explicitly the motion of multiple, intersecting discontinuities, the DEM is particularly suitable for the analysis of discontinuity, such as masonry structures, where significant deformation occurs as relative motion between the blocks. The disadvantage of this element arises in the case of complicated deformation problems (e.g., beam-column component behavior in an infilled frame), where the number of triangular elements into which the area has to be discretized may become very large. The main difference between DEM and other continuum-based methods (e.g., FEM) is that the contact points in DEM are automatically updated and changed based on the “contact overlap” concept, leading to the detection of new contacts and complete detachment of blocks as the calculation process allows.

For the solution of the nonlinear dynamics problem, FEA packages use an iterative scheme, such as the Newton-Raphson method, and an implicit or explicit integration scheme. Some of these procedures have been adapted to take into account the brittle nature of the constitutive model of concrete with satisfactory results.

A large number of procedures/techniques have been used to apply micromodeling to infill frames, while taking into account all of the above parameters. Each of these models has its own merits and limitations, as well as level of complexity, and has been used to study different sets of parameters affecting the behavior of infilled frames.

The knowledge acquired over the years has made possible the development of commercial packages that provide a wide selection of elements, constitutive relationships, and solution schemes that researchers may use as tools for the study of such complex problems as the ones described above. Nevertheless, only recently have such packages been used for the study of infilled frames. A thorough discussion of the above concepts and models is presented by Asteris et al. (2013).

Reinforced Concrete Infills in Reinforced Concrete Frames

The construction of new walls is the most effective and economic method for retrofitting multistory reinforced concrete (RC) buildings, especially those with pilotis (soft ground story). Their structural and economic effectiveness increases when selected bays of an existing RC frame are fully infilled with integral RC walls replacing masonry ones. Such a method is appealing since the intervention is concentrated in only certain bays of the frame and reduces the disturbance to the inhabitants as compared to the case of using concrete jackets on the columns.

Most of the experimental research work performed in the last decades has focused on other frequently used types of retrofitting, in particular on fiber-reinforced polymers (FRP) and concrete jackets. Research on the use of RC infill walls has mainly targeted what is feasible: testing of one- to two-story specimens. However, data is lacking for taller full-scale specimens that reflect real-life applications, due to the practical difficulties associated with the high forces needed for the tests. Regarding code provisions, Eurocode 8 – Part 3 fully covers retrofitting with FRP or concrete jackets, while it does not address the retrofitting of RC frames with the addition of new walls created by infilling selected bays.

Experimental research on reinforced concrete frames converted into walls by infilling with RC has been carried out almost exclusively in Japan and Turkey. The experiments in Japan (Chrysostomou et al. 2013, 2014) were performed on a total of 27 1:3–1:4 scale single-story one-bay RC-infilled frames with RC infill walls with a thickness of 26–60 % (on the average 43 %) of the width of the frame members. The test results were compared in most cases with monolithically cast specimens of the same geometric characteristics (in which the frame and the infill wall were cast at the same time and integrally connected). The connection of the RC infill to the bounding frame was done by means of epoxy-grouted dowels (17 specimens) or through mechanical devices, such as shear keys and dowels without epoxy (6 specimens). In four other test campaigns, the thickness of a preexisting thin wall was increased by 100–150 % without any direct connection of the new wall with the bounding frame. The failure mode of all specimens was in shear (including sliding at the interface). It is interesting to note that for epoxy-grouted dowels, the force resistance of the infilled frame was on average 87 % of the integral one, while for the mechanical connections it was 80 % on average. For the increased thickness of an existing thin infill wall, the force resistance was on average 92 % of the monolithic specimen, while the displacement at failure was on average 13 % smaller than for the integral specimen. For the epoxy-grouted dowels and for the mechanical connection, the ultimate deformation was on average 55 % and 115 % larger than in the integral specimen, respectively. The results show that although a deformable connection gives a somewhat reduced strength with respect to the monolithic case, the ultimate deformation of the retrofitted structure is considerably increased.

Among the specimens tested in Turkey (Chrysostomou et al. 2013, 2014), some were single-story one-bay 1:2 and 1:3 scale, with RC infill thickness 25 % and 33 % of the width of the frame members. Others were two-story one-bay scaled at 1:3, with infill wall thickness 33 % and 40 % of the width of the members of the bounding frame. The RC infill was in most cases fully connected on the perimeter with dowels, in some cases there was a gap between the infill and the columns, while in some other cases there was no connection other than simple bearing. In one case, the rebars of the infill were welded to those of the members of the frame, instead of using dowels, and monolithic specimens (not exactly similar to the infilled ones) were included for comparison purposes. Finally, there was a two-story three-bay specimen scaled at 1:3, with the middle bay infilled with a wall with 63 % thickness of the width of the frame members. The connection was made with epoxy-grouted dowels and the failure mode was predominantly flexural. In all other cases, the single-story walls failed in shear, while the two-story walls failed in a combination of flexure and shear sliding at the base.

The test specimens used in the experiments described in the previous paragraphs correspond to walls with failure modes dominated by shear, with low aspect ratios not representative of multistory slender walls. In fact, the failure mode of multistory slender walls is controlled by bending and the design is governed by the formation of a plastic hinge at the base. In such a case, shear will not have a detrimental effect on displacement and energy dissipation capacity. In addition, it has been shown numerically that higher modes may increase considerably the shear forces at the upper floors of a wall after the formation of a plastic hinge at the base. This aspect has never been studied experimentally because their height and number of stories has not been large enough to allow higher mode inelastic response. Another common element of past tests is the smaller thickness of the RC infill wall relative to the width of the frame members. As a result, the weak link of the structural system is either the infill wall in diagonal compression or its connection with the surrounding frame.

In order to start bridging the gap of knowledge regarding infilling of existing RC frames with RC walls, the effectiveness of seismic retrofitting of multistory multi-bay RC-frame buildings by converting selected bays into new walls through infilling with RC was studied experimentally at the European Laboratory for Structural Assessment (ELSA) of the Joint Research Centre in Ispra (Italy). Similar tests on scaled 0.75:1 specimens were performed at the Structures Laboratory of the University of Patras. Detailed description of the former results is presented by Chrysostomou et al. (2013, 2014) and for the latter by Fardis et al. (2013).

The test specimen in ELSA was designed based on a four-story prototype building structure consisting of four three-bay frames spaced at 6 m, with RC infilling of the exterior frames only. The specimen was designed at full scale to represent the two exterior frames of the prototype structure, spaced at 6 m and linked by a 0.15 m thick RC slab.

The dimension of the specimen in the direction of testing was 8.5 m (two exterior bays of 3.0 m and a central bay of 2.5 m), with an inter-story height of 3.0 m and a total height, excluding the foundation, of approximately 12.0 m. The structure was designed for gravity loads only using the code provisions of the 1970s in Cyprus.

In order to facilitate the study of the effect of as many parameters as possible, the walls in the two frames, which had a thickness of 0.25 m equal to the width of the beams and columns of the bounding frame, were reinforced with different amounts of reinforcement, with the north frame being the stronger of the two. Figure 5 depicts the south and north frames, with the south one being on the right of the picture.

Assessment and Strengthening of Partitions in Buildings, Fig. 5

Elevation of the specimen in the lab with the south wall on the right of the picture (Chrysostomou et al. 2014 with kind permission from Springer Science and Business Media)

The design approached used by Fardis et al. (2013) was adopted and two parameters were examined: (a) the amount of web reinforcement in the walls and (b) the connection detail between the wall and the bounding frame. Regarding the connection with the bounding frame, two distinct connection details were used. In the first detail, the web bars are connected to the surrounding frame through lap splicing with same diameter starter bars epoxy grouted into the frame members. Short dowels are then used in order to transfer the shear at the interface between the wall and the frame members (see bottom beam of Fig. 6a).

Assessment and Strengthening of Partitions in Buildings, Fig. 6

(a) Dowels and starter bars. (b) Dowels, starter bars, and web reinforcement (Chrysostomou et al. 2014 with kind permission from Springer Science and Business Media)

In the second detail, longer dowels were used to act both as dowels and as anchorage of the web panel to the surrounding frame (see left column and top beam of Fig. 6a); to this end, the dowels are considered as lap spliced with the nearest – smaller diameter – web bars. The complete wall reinforcement (including web, starter bars, and dowels) is shown in Fig. 5b.

Since the lapping of the column reinforcement can only take compression forces, a lap-splice failure in tension would be highly detrimental to the whole experiment. Therefore, in order to safeguard against this type of failure and allow the experiment to be performed without any premature failure, it was decided to reinforce the bounding columns of the wall at the first floor with three-sided CFRP (carbon-fiber-reinforced polymer) for a height of 0.60 m from the base of the column (Fig. 7).

Assessment and Strengthening of Partitions in Buildings, Fig. 7

CFRP reinforcement of the column to safeguard against lap-splice failure (Chrysostomou et al. 2014 with kind permission from Springer Science and Business Media)

The testing campaign consisted of two pseudo-dynamic tests (one at 0.10 g and the other at 0.25 g) and a cyclic test. Some findings regarding the behavior of the structure are:

• 1. The structure managed to sustain an earthquake of 0.25 g without significant damage.

• 2. Some column lap splices failed with concrete spalling, but the structure continued to carry load.

• 3. The three-sided CFRPs protected the wall bounding columns at the firstt floor and prevented lap-splice failure.

• 4. The “weak” south frame behaved equally well as the “strong” north frame.

• 5. The slip displacements at the horizontal interfaces of the ground-floor walls were of the order of 0.8 mm, which is very close to the full engagement of the starter bars but not of the dowels.

• 6. The slip displacements between the wall and the bounding columns of the ground floor were of the order of 0.4 mm.

• 7. The behavior of the wall was mainly flexural; yielding took place at both the ground-floor and the first-floor walls.

• 8. The distribution of strains along the bounding columns of the walls shows that the ones for the ground floor are much larger than those of the first floor, while those of the second and third floors are negligible.

• 9. The two connection arrangements used behaved satisfactorily.

• 10. Higher mode effects appeared in the response of the structure.

• 11. Some vertical cracks appeared at the connection of the beams to both the exterior and the wall columns.

• 12. A horizontal crack appeared at the ground beam of the walls, which was the main cause for the loss of strength of the south frame.

It was demonstrated that this is a viable method for retrofitting and it can be used to strengthen existing ductility and strength-deficient structures. The recorded global and local behavior of the structure provides data for the development of numerical models, to facilitate the proposal of design guidelines for such a retrofitting method.

Strengthening of Partitions

In general, infill walls and in particular the masonry infill panels exhibit a small plastic region on the stress-strain curve due to significant decrease in stiffness, strength, and energy absorption capacity. In order to enhance the behavior of infill partitions and remove their inherent deficiencies, various researchers have proposed methods for improving the performance of these elements.

One of the methods is to use shotcreting on the faces of masonry infills which can increase the stiffness and the lateral load capacity of the infilled frame and reduce the lateral drift at the ultimate load. Calvi and Bolognini (2001) have performed full-scale tests in which they have placed a 10 mm plaster on both sides of a masonry infill wall covering either reinforcement (Ø5 mm or Ø6 mm) or wire meshes and have studied the behavior of the strengthened infilled frame. Based on their observations from the experiments, the introduction of some reinforcement in the mortar layers, with a geometrical percentage lower than 1 %, will almost double the acceleration levels for the occupational and damage limit states.

Another technique is the use of fiber-reinforced polymer (FRP) reinforcement to enhance the overall response of such systems. Several experimental studies have been performed which demonstrate that significant improvement of strength and energy absorption capacity can be achieved if adequate anchoring is provided.

Altin et al. (2008) investigated experimentally the behavior of strengthened masonry-infilled reinforced concrete (RC) frames using diagonal carbon fiber-reinforced polymer (CFRP) strips under cyclic loads. Test results indicated that, CFRP strips significantly increased the lateral strength and stiffness of perforated clay brick-infilled nonductile RC frames. Specimens receiving symmetrical strengthening showed higher lateral strength and stiffness, compared to the ones at which CFRP strips of the same width were applied to one of the interior or exterior surface of the infill wall. In the latter case, similar lateral strength and stiffness was obtained, irrespective of the side of placement of the strips.

Lunn and Rizkalla (2009) used glass fiber-reinforced polymer (GFRP) systems for increasing the out-of-plane resistance of infill masonry walls to loading. They have concluded that GFRP strengthening of infill masonry walls is effective in increasing the out-of-plane load-carrying capacity when proper anchorage of the FRP laminate is provided. They also note that the latter has a significant effect on the failure mode of the assemblage.

Papanicolaou et al. (2007) have used the textile-reinforced mortar (TRM), which is a new structural material, for testing its capability for increasing the load-carrying capacity and deformability of unreinforced masonry walls subjected to cyclic in-plane loading. TRMs comprise fabric meshes made of long woven, knitted, or even unwoven fiber rovings in at least two directions. The density of rovings in each direction can be controlled independently, thus affecting the mechanical characteristics of the textile and the degree of penetration of the mortar matrix through the mesh openings.

Based on the experimental results, the authors stated that, in terms of strength, TRM jackets are at least 65–70 % as effective as FRP jackets with identical fiber configurations. In terms of deformability, of crucial importance in seismic retrofitting of unreinforced masonry walls, TRM jacketing for shear walls is about 15–30 % more effective than FRP. They therefore conclude that TRM jacketing is an extremely promising solution for strengthening and seismic retrofitting of unreinforced masonry walls subjected to in-plane loading.

A more recent development is the use of engineered cementitious composites (ECC) which are a special class of fiber-reinforced cement-based composite materials, typically reinforced with polyvinyl alcohol fibers. Dehghani et al. (2013) have tested in diagonal compression a number of specimens with different ECC-strengthening configuration, and they evaluated their in-plane deformation and strength properties, including the post-peak softening behavior. They state that the proposed technique can effectively increase the shear capacity of masonry panels (1.5–2.8 times), improve their deformability, enhance their energy absorption capacity (35 times), and prevent the brittle failure mode.

Summary

The presence of infill walls has shown that in most cases they have beneficial effects on the behavior of structures during earthquakes and have contributed in the prevention of collapse of many structures. It is therefore for this reason that efforts have been made to use engineered infills to retrofit existing seismic-deficient buildings, since they present an economically viable solution. Several equations have been introduced to define the stiffness of infilled frames and their capacities, which are related to a number of failure modes. These were considered in macro- and micro mathematical models that have been proposed to simulate their behavior and their interaction with the bounding frame. The former are simpler (especially the single-strut models), while the latter require a higher level of modeling sophistication. The construction of new walls is the most effective and economic method for retrofitting multistory reinforced concrete (RC) buildings, especially those with pilotis (soft story). Their structural and economic effectiveness increases when selected bays of an existing RC frame are fully infilled with integral RC walls replacing masonry ones. Recent experimental investigations make possible the quantification of this methodology and the proposal of code equations for the design of such systems. Several methods exist for improving the properties of infilled panels avoiding brittle failure and increasing their load-carrying capacity as well as their energy absorption during an earthquake.

Cross-References

• Numerical Modeling of Masonry Infilled Reinforced Concrete Frame Buildings

• Retrofitting and Strengthening of Structures: Basic Principles of Structural Interventions

• Seismic Fragility Analysis

• Seismic Strengthening Strategies for Existing (Code-Deficient) Ordinary Structures

• Seismic Vulnerability Assessment: Reinforced Concrete Structures

• Strengthening Techniques: Code-Deficient R/C Buildings

• Strengthening Techniques: Code-Deficient Steel Buildings

• Time History Seismic Analysis