You are on page 1of 5

Short Paper Proc. of Int. Conf.

on Advances in Civil Engineering 2011

Design and optimization of a GFRP panel for building construction


Guglielmo Carra1
1

Department of Building Environment Science and Technology Polytechnic University of Milan, Milan, Italy Email: guglielmo.carra@mail.polimi.it resin impregnation system (open bath or direct injection) where resin is exposed to the fiber at the polymerization temperature. The resin-filled fiber is then pulled through a die and cured to produce a solid with constant cross section and desired mechanical properties. At the end of the line, the continuous product is cut to length. The presence of industrial processes and awareness of physical and structural performance that GFRPs provide, make these materials a viable alternative to traditional ones (steel, aluminum, concrete). Among the numerous existing applications it is interesting the use of GFRP for the construction of all-FRP buildings. In practice, however, the use of these systems is limited to buildings with a height of 2-3 floors. Several reasons are related to this dimensional constraint. The low resistance to fire is one of the more binding. The fire safety standards in Europe and United States are very restrictive and the fire behavior of GFRP is still little known and generally perform poorly. When the temperature of an item reaches or exceeds the transition glass temperature (Tg) of the matrix, the bonding between fiber and matrix loses its strength and the element collapses rapidly [2]. The geometry of the fiber-reinforced components, due to their young age and the few years of experimenting on the material, is often not optimized for the specific task and decreases the structural performance of the material. In particular the problem of buckling for compressed elements is a critical issue.Moreover, to think in terms of all-FRP buildings involves a complex management of the global performance of the building. A deep and careful analysis of properties that are extremely different from those of traditional materials is requested.

AbstractNowadays building design can not prescind from high requirements both for structural and physic-energetic performances. It is well known that the behaviour of lightweight structures is even better than those built with traditional materials (concrete and steel), with respect to static and dynamic loads, due to the reduced impact of horizontal and vertical forces on the structure. FRPs, despite being a fragile material, have a good structural behavior because of their low weight and, similarly to wood, for the high level of ductility that is achieved through the use of mechanical connections among elements.Present research involves the design and optimization of a GFRP load bearing panel that maximizes the structural response with respect to concentrated and distributed loads, even of great intensity. In particular, the panel should maximize the response to compression, that is problematic because of buckling in presence of axial forces. The structural analysis is carried on using FEA software and the panel is modeled as a pultruded uniaxial - element.The study takes into account not only the structural issues and the panel is designed as an integrated element, for facades and floors, that responds to highest Italian standards for thermal performance. Panels can be assembled together, to create modular buildings. Moreover, the weight of the panel is optimized to reach significant advantages on the construction site and reduce the need for workforce. Starting from the design stage, is given emphasis to the lifecycle of the panel and particular attention is paid to the possibilities of reuse. Index Terms All-FRP buildings, GFRP panel, building construction systems, envelope systems, building performance.

I. INTRODUCTION During the past twenty years several building components made by Fiber Reinforced Polymers (FRPs) as basic material have been developed [1]. The FRPs consist of the union of a polymer matrix (resin) and fibers. To these two basic components are added fillers and additives to improve specific features of the material: increase the transition glass temperature (T g), increase the resistance to moisture penetration, etc. To manufacture building products, fiber glass is most widely used (GFRP), while more rare is the adoption of carbon fiber, Kevlar, or natural fibers. Between the glass fiber, Eglass is the most diffuse, thanks to the reduced cost that is combined with the opportunity to produce components, with a high level of homogeneity, through the pultrusion process.Pultrusion involves physically pulling material through a manufacturing line. Fiber threads are pulled into a 2011 ACEE DOI: 02.ACE.2011.02.25 31

Figure 1. Example of a pultruded GFRP construction system: the system is Composolite, produced by Strongwell. On the right a building made with the Composolite system.

Short Paper Proc. of Int. Conf. on Advances in Civil Engineering 2011 Following a brief list of the main properties of GFRP materials is presented: -Low weight, about 2000 kg/m3; -High specific resistance; -High ratio: live load/weight; -Low heat transfer; -Translucent to the passage of light; -High resistance to environmental agents; -Water and vapor proof. II. SCOPE AND OBJECTIVES Scope of the present research is design a GFRP panel - in a double derivation, one for walls and one for floors - be used as primary structure and high performance envelope for building construction. The panel, composed in a simple building system and designed with an optimized transversal section, is studied to carry the load of a tall building, comparable with those made by traditional materials. A theoretical analysis, through the use of simulation software, of the potential associated with the use of GFRP panels as a building envelope and a comparison with other typical building systems is proposed. The objective is quantify the maximum height of a building entirely made up of GFRP panels, such as in analysis, before incurring in structural failure mechanisms. It will be also defined the physical performance of the panel from the thermal point of view. A comparison with current Italian law is proposed, as a computational tool to support the software simulations. III. PANELS ANALYSIS: STRUCTURAL RESPONSE TO BUCKLING AD BENDING Panels structural analysis is conducted using the finite element software, Abaqus CAE. For the vertical panel a buckling analysis is carried out with the aim to define the values of the critical load for which the panel buckles. A global analysis - the whole panel - and a local analysis - the webs and flanges of the panel - was conducted. The horizontal panel has been tested to bending, after the application of a distributed load that simulates the real conditions in use. The permanent load corresponds to the one for a residential building. A comparison between numerical results from the FE analysis and the formulas from the Italian CNR-DT 205:2007 Guide for design and construction of pultruded FRP Structures Made by elements [3] is proposed. The aim is to identify not only a direct link between the formulas and the results from the analysis, but also provide a tool to control the accuracy of the results obtained through analysis. It is worth noting, at this stage, that the National Research Council (CNR) guideline refers in particular to long composite elements, with a maximum ratio between transversal dimension and inflection length, of approximately 1/20. For the present case of study this ratio is reduced to 1/6 and both panels should be considered as short beams. In this type of itemslocal buckling failure is going to occur before the global buckling failure. 2011 ACEE DOI: 02.ACE.2011.02.25 32
Figure 2. Geometric dimensions for the transversal sections of the two selected panels. On the left the transversal section for the vertical panel and, on the right, the transversal section for the horizontal panel. TABLE I. STRUCTURAL PROPERTIES FOR THE MATERIAL USED IN THE FE ANALYSIS. THIS VALUES ARE REFERRED TO A PERFECT MATERIAL

For the simulation a set of theoretical values were used to characterize the construction material. GFRP is considered made by E-glass fibers and epoxy matrix. The values, for the orthotropic material, are calculated as referred to a hypothetical case with no imperfections (Table 1). To increase the reliability of the analysis, an additional set of simulations using a modulus of elasticity of 23 GPa is performed. This value is suggested by the CNR-DT 205:2007 and represents the result of an average between the actual values obtained through a monitoring of Italian manufacturers of pultruded materials. A. GEOMETRIC DIMENSIONS OF THE PANELS The structural analysis is conducted comparing different cases for each derivation, through the use of different transversal sections: four cases for the vertical panel and two cases for the horizontal panel. At the end only one case for each derivation, the one that better optimizes the performance from a structural point of view, has been selected. Fig. 2 reports the two cases selected.

The vertical panel has a height of 3 meters while the vertical ones has a length of 4 meters. Following the results for the two cases studies through the finite element model analysis and the calculation according to the instructions of the guides CNR are presented. B. VERTICAL PANEL: GLOBAL BUCKLING Pultruded structural members have open or closed thinwalled cross sections. For long composite columns, overall (Euler) buckling is more likely to occur before any other instability failure. For short columns, local buckling occurs first leading either to large deflections (crippling). Because of the large elongation to failure allowed by both the fibers and the resin, the composite material remains linearly elastic for large deflections andstrains unlike conventional materials that yield (steel) or crack (concrete) for moderate strains. Therefore, buckling is the governing failure for this type of cross-section and the critical buckling load is directly related to the load carrying capacity of the member [4], [5], [6].

Short Paper Proc. of Int. Conf. on Advances in Civil Engineering 2011 Using the CNR the critical stress can be calculated as follows: In Fig. 4 are shown the results of the analysis for the cases studied with a comparison between the values obtained using expression (1) and those obtained from FE analysis. Looking to Fig. 4, seems clear that the FEA provides higher values, about 20%, compared to those obtained using the CNR formula. The comparison between different panels makes it clear that case Dver is the one that maximizes critical stress, and thus the load that must be applied to generate the buckling of the element. This case, in fact, has a thicker section and a double cavity, which increase the moment of inertia of the section and the amount of material that reacts to stress. Particular attention should be paid to the gray column referred to the Youngs modulus of 23 GPa, which is the most restrictive condition and the one closer to the real case. A red line on the graph shows a comparison between the values of critical stress for a panel at the base of a building entirely constructed by GFRP panels, of the type in study. Different heights and coverage ratios are identified and compared. The coverage ratio is a fraction that defines the number of panels that are used along the base perimeter of a square building.Ratio 1/1 indicates a building completely covered with panels and no openings; ratio 1/6 indicates a building in which alternate, for each panel installed, five voids. C. VERTICAL PANEL: LOCAL BUCKLING Local buckling for the constitutive elements of the panel [7], according to the formulas of the CNR-DT, can be calculated, similarly to what happens to a I-beam. The Italian guide suggests the use of two different formulas, depending on whether the phenomenon of local instability occurs in the web or in the flange of the panel. Between the two expressions the lowest value must be taken into account. for web:

where f is a corrective factor that, for the specific case in analysis, is equal to 1,45.

Figure 3. These images came from the FE software Abaqus CAE, they represent the four cases simulated for the vertical panel. Panels section vary in order to find the best shape that optimizes the critical load to bending.

In Fig. 3 are listed the four cases for the vertical panel that are analyzed with the aim to maximize the critical load to buckling.

for flange:

where tw and tf are the thickness of web and flange, and bw and bf is their length. In the cases analyzed vertical instability always occurs in the flange, due to the reduced thickness at the open end of the section.
Figure 4. Comparison between global buckling values for the four cases analyzed. Two different values for the critical stress are identified due to the variation of the Elasticity Modulus. With the dotted line is the value obtained using the CNR-DT 205:2008 and, with the red line, is indicated the confront between the maximum stress achieved in a panel posed at the base of a building entirely made by GFRP panels.

2011 ACEE DOI: 02.ACE.2011.02. 25

33

Short Paper Proc. of Int. Conf. on Advances in Civil Engineering 2011

Figure 6. The two images are from Abaqus CAE. The one on the left is about the displacement of the panel, under a uniform distributed load of 300 kN/m2. The one on the right represents the stress distribution inside the panel for the same load applied. Figure 5. Comparison between local buckling values for the four cases analyzed. Two different values for the critical stress are identified due to the variation of the Elasticity Modulus. With the dotted line is the value obtained using the CNR-DT 205:2008 and, with the red line, is indicated the confront between the maximum stress achieved in a panel posed at the base of a building entirely made by GFRP panels.

IV. PANELS ANALYSIS: THERMAL PERFORMANCES AS BUILDING


ENVELOPE

In Fig. 5 the values for local buckling analysis of the four panels are collected. In this case it is identified a reduction for the values of the critical stress of about 70%. Looking to the comparison of the local buckling stress with the height of a building constructed using the composite panels is evident that the local buckling represents the first structural constraint to the height of buildings realized with this kind of panels. A building, characterized by a coverage ratio of 1/6, will not be, conservatively, taller than 15 storeys, because the sum of structural loads would trigger a local failure mechanism. The comparison between the FEM analysis and calculation, through legislation, reveals a lower margin of error, about 10%, compared to the case of global buckling. In this case the comparison with the CNR seems to be more appropriate. From the FEM analysis it is possible to detect that the values of the stress inside the panel before buckling occurs, both the local and global, are always below the maximum allowable stress for the material, both in tension and compression. C. HORIZONTAL PANEL: BENDING A NALYSIS Bending analysis is performed for the horizontal panel. The panel is loaded with a uniform distributed load of 300 kN/m2. This load configuration represents, by excess, the one of a residential building under a permanent load, following the Italian regulation NTC 2008 [8]. For the horizontal panel, unlike the vertical one, two distinct cases are compared. Results from the FE analysis reveal a displacement that is lower than the maximum by law and, specifically, panel = 1,1 cm < law = 1/250. Values for the maximum stress inside the panel are always lower than the maximum allowable stress for the material, both in tension and compression.

The panel is designed with the intent to optimize not only the mechanical and structural performance, but also to improve physical performance. With specific reference to the vertical panel, inserting in the cavities of the panel proper insulation materials allows to achieve high thermal performance. For case Dver (Fig. 2) the two cavities are filled with malamminic foam resin - in the thinner cavity - and polystyrene - in the thicker cavity. The verification of the thermal properties of the panel is made applying the calculation methods of UNI EN ISO 13786:2008 [9]. Then the thermal performance of the GFRP [10] wall panel are compared to those of other building envelope systems made by conventional materials, such as concrete and bricks. Results are shown below along with the formulas used to calculate the thermal static and dynamic parameters. Thermal Trasmittance: U = 1/R (4) Thermal Resistance R = 1/ i + n n/dn + 1/ e (5) Thermal phase displacement: It is evident that the panels thermal performance are comparable or superior to those of a traditional building system. In the proposed comparison both concrete and brick walls are insulated with proper materials. The weight of the GFRP panel, on the other hand, is from five to seven times lighter.
TABLE II. MAIN THERMAL PARAMETHERS FOR THE GFRP PANEL AND CONFRONT WITH THE ONES OF OTHER TYPICAL CONSTRUCTION SYSTEMS.

2011 ACEE DOI: 02.ACE.2011.02. 25

34

Short Paper Proc. of Int. Conf. on Advances in Civil Engineering 2011 and it occurs for values lower than 70% with respect to global buckling. - A partial comparison between the formulas proposed by the CNR-DT 205:2008 has been identified for the study of local instability. For global instability, an error of 20%, induced to conclude that the formula (1) can not be used as a reliable forecasting tool to assess global buckling. -The bending behavior of the horizontal panel is optimal, and comply with the limits of the law. Panels displacement is 1.1 cm for a length of 4 meters and a uniformly distributed load of 300 kN/m2. After the simulation, case Bver, seems suitable as high-performance floor for residential buildings. - The study of thermal performance, both dynamic and static, reveals a good behavior of the panel to the transfer of heat, this is partly due to the low thermal transmittance of the GFRP, but especially to the use of proper insulating materials within the cavities of the panel. The weight of the panel is between 5 and 7 times lower than traditional building systems and its thickness is reduced of 50%. - The design of the all-GFRP building system, represents only an initial step but already reveals its potential. The connections between vertical and horizontal components can be ensured by metal elements (bolts or mechanical joints), enabling rapid assembly phase. The possibility of disassemble the system at its end of life is another advantage, especially thinking in terms of recycling of building components and their reuse in new building systems. REFERENCES
[1] T. Keller, Recent all-composite and hybrid fibre-reinforced polymer bridges and buildings, Progress in Structural Engineering and Materials, vol. 3, pp. 132-140, 2001. [2] Tracy C. D., Fire endurance of multicellular panels in an FRP building system, Ecole Polytechnique Fdrale de Lausanne EPFL, Phd Thesis, 2004. [3] CNR-DT 205:2007 - Guide for the Design and Construction of Structures made of FRP Pultruded Elements. [4] E. Barbero and J. Tomblin, Euler Buckling of Thin-Walled Composite Columns, Thin-Walled Structures, vol. 17, pp. 237 258, 1993. [5] G. A. Kardomateas, Global buckling of wide sandwich panels with orthotropic phases: an elasticity solution, in Sandwich Structures 7: Advancing with Sandwich Structures and Materials, Springer, 2005 [6] M. A. Moussa and N. Uddin, Global buckling of composite structural insulated wall panels, Materials and Design, vol. 32, pp. 766-772, 2011. [7] J. Tomblin and E. Barbero, Local Buckling Experiments on FRP Columns, Walled Structures, vol. 18, pp. 97 - 116, 1994. [8] D. M. 14 Gennaio 2008 - Norme Tecniche per le costruzioni. Technical Law (in Italian). [9] Uni En Iso 13786:2008. Thermal performance of building components - Dynamic ther-mal characteristics - Calculation methods. [10] B. Mutnuri, Thermal Conductivity Characterization of Composite Materials, West Virginia University, Master Thesis, 2006.

Figure 7. Assembly scheme of the building system, obtained by combining the horizontal and vertical panel with the use of metal joints.

V. PROPOSED BUILDING SYSTEM The proposed building system consists of the union between the vertical and horizontal panels through the use of metal connectors. Two different types of connections are studied, the first is between coplanar elements and it is guaranteed by steel bolts; the second ensures - out of plane - connections and it is handled by special metal connectors to provide an adequate strength and proper connection between wall and floor elements. This connectors are made by steel and composed by a central body with trapezoidal cross-section that perfectly fits into the cavities of the vertical panel. A shelf is welded - in horizontal position - with respect to the main body of the connector and serve as support for the horizontal panel. Several aspects, related to the study of the present building system, need further analysis. In particular, the behavior of the connections in presence of concentrated shear stress should be studied. A detailed analysis, through the FE analysis, is under investigation. CONCLUSIONS Several interesting conclusions can be drawn, both at the scale of the single building component - the panel - and at the scale of the building system. - Structural analysis carried out through a comparison between the Italian regulation and FE analysis, reveals that the vertical panel is capable of supporting the weight of a 15 storeys building. The reachable height is inversely proportional to the number of panels that are used along the base perimeter of the building. With the ratio of 1/6, one panel every five void modules, the height of 15 floors must be considered as a limit. If the coverage ratio is reduced, the height of the building is expected to increase. Anyway, after the simulation, is proved that the GFRP panels allow to reach heights comparable to those of traditional buildings made by brick or concrete. - The local buckling is the most probable failure mechanism 2011 ACEE DOI: 02.ACE.2011.02. 25 35

You might also like