A rehabilitation study of sandwich GRC facade panels

Transcription

1 Construction and Building Materials xxx (2005) xxx xxx Construction and Building MATERIALS A rehabilitation study of sandwich GRC facade panels João R. Correia *, João Ferreira 1, Fernando A. Branco 2 Department of Civil Engineering and Architecture, Instituto Superior Técnico (IST), Technical University of Lisbon, Av. Rovisco Pais, Lisbon, Portugal Received 25 November 2004; received in revised form 13 January 2005; accepted 31 January 2005 Abstract Glass fibre-reinforced concrete (GRC) has been used for more than 30 years in the construction industry, especially in facade panels. In this paper, a rehabilitation case study is presented, of a building where cracking and excessive deformations were detected in the sandwich GRC facade panels due to thermal effects. The structural behaviour of the panels was simulated with a finite element method (FEM) model, validated by experimental tests performed on the facade, and then used to assess the efficiency of a repair technique, based on external insulation, which showed a good efficiency. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: GRC; Facade; Sandwich panels; Thermal effects; Defects; FEM; Rehabilitation; Repair 1. Introduction Glass fibre-reinforced concrete (GRC) is a composite material that consists of a cementitious matrix in which short length glass fibres are dispersed [1]. The glass fibres confer the brittle matrix a more ductile behaviour, a much greater impact resistance and some tensile strength. These properties allow for the manufacture of thin walled elements, with important advantages when compared to alternative solutions such as reinforced concrete or steel plates. This material has been used for more than 30 years in the construction industry in Europe, especially in facade panels, which are not usually designed as structural elements. In early the GRC development, one of the problems of most concern was the durability of the glass fibres, * Corresponding author. Tel.: ; fax: addresses: jcorreia@civil.ist.utl.pt (J.R. Correia), joaof@civil.ist.utl.pt (J. Ferreira), fbranco@civil.ist.utl.pt (F.A. Branco). 1 Tel.: ; fax: Tel.: ; fax: which became fragile with time, due to the alkalinity of the cement mortar [1], leading to a reduced strength of the material. Since then, significant progresses have been made, and presently the problem is practically solved with the new types of alkali-resistant glass fibres and with mortar additives which prevent the processes that lead to the embrittlement of the GRC [2 4]. In this paper, a rehabilitation case study is presented, of a 20-year-old GRC building facade, where the ageing phenomenon was not taken into consideration in the manufacturing phase. At the time, no alkali-resistant fibres nor adequate additives were employed. Important defects were detected, namely extensive cracking and panel deformation. The facade is basically built with sandwich panels, of two outer GRC thin plates with insulation elements (expanded polystyrene and mineral wool) in between. The finite element method (FEM) was used to assess the mechanical behaviour of the facade elements, taking into consideration the self weight and the temperature gradients. The numerical model was previously validated with results of experimental tests performed on the facade. The numerical analysis reproduced the de /$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi: /j.conbuildmat

2 2 J.R. Correia et al. / Construction and Building Materials xxx (2005) xxx xxx fects detected in the facade inspection, leading to the identification of the respective causes. To reduce the existing defects and prevent their evolution, several technical solutions were considered for the rehabilitation of the facade panels and simulated with FEM models. The results of the FEM simulations showed that the most effective repair measure was the use of an additional external insulating system. Mineral Wool GRC EPS GRC Brick wall 2. Description of the facade geometry and pathologies 2.1. Facade geometry The facade elements are 2.80 m wide and 3.25 m high sandwich panels (Fig. 1), with two external 10 mm thick GRC sheets, outside two 25 mm thick expanded polystyrene (EPS) sheets and a 30 mm thick mineral wool sheet in the middle (Fig. 2). The two external GRC sheets are connected by two vertical and two horizontal 10 mm thick interconnecting GRC laminas to provide global geometrical stability. The panels are supported at the bottom on the floor slabs with an identical 300 mm wide horizontal sandwich element. The facade panels have two vertical sandwich elements which are aligned with the vertical extremities of the window central opening and provide global stability of the facade. The facade panels are connected to the building pavement slabs (on the top and on the bottom) by means of bolted steel angles (Fig. 1). The facade panels are 0.40 m distant from an interior 0.07 m thick brick wall (Fig. 2) next to the interior of the Fig. 2. Facade section (dimensions in mm). building where the ambient temperature has a constant value of 23 C due to the air conditioning system Facade pathologies The main defect detected in the facade panels consisted of extensive cracking with the following pattern: horizontal or vertical cracks near the window corners, coinciding with the GRC laminas (Fig. 3); vertical and horizontal cracks mainly in the upper zones of the panels; these cracksõ openings are less than 1 mm and completely cross the GRC sheet thickness (Fig. 4). In the interconnecting vertical laminas of the panels some cracking was also observed. Slab D4 D5 3. Experimental testing D3 Window opening Regarding the observed defects, two different effects were initially considered as potential causes, besides the degradation of the panels and their self weight, namely the thermal effects due to solar radiation and the inner water pressure (considering that the panelsõ D1 D2 Slab Fig. 1. Geometry of the facade panels section, frontal view and plan (dimensions in mm). Fig. 3. Cracking pattern in the facade panels.

3 J.R. Correia et al. / Construction and Building Materials xxx (2005) xxx xxx 3 extensive cracking could allow for significant rain water penetration). To assess the importance of these effects, three different experimental in situ tests were carried out, namely a water test, a heating test and a cooling test. The in situ GRC tension and compression resistances were also determined with experimental tests, on specimens extracted from an existing panel, in order to perform a structural evaluation, considering the stresses analytically determined Water test Fig. 4. Cracks crossing the panel thickness. The water test was performed to check the sensitivity of the panel to absorb rain water and to evaluate its displacements under inner water pressure. The test consisted of submitting the surface of one of the most severely cracked panels to water pouring applied with a sprinkler, and measuring the horizontal displacements (with an electrical displacement transducer with 0.01 mm precision) perpendicularly to the facade plane. During the test, the weather was characterised by an alternate sunny/cloudy condition. The water pouring led to facade displacements towards the interior, opposite to those expected by the inner water pressure. The displacement measurements proceeded after the water pouring and it could be noted that the facade displaced towards the interior during the cloudy condition and towards the exterior under sunny condition. It was clearly concluded that the measured displacement was due to temperature decrease of the panel surface and not associated to the water pouring as expected. Fig. 5 shows the displacement measurements during and after water pouring. When the water pouring stopped, a drain hole was opened through the thickness of the facade panel (Fig. 4). This hole showed that there was no water accumulation in the interior of the panel, meaning that very little water had entered through the cracks, and so this effect was discarded as it was not really important Heating test Based on the previous experimental results a heating test was implemented, consisting of measuring the facade panelsõ deformation under a significant solar radiation exposure. During the test, the displacements of five points of a panel on the buildingõs west facade were measured (points D1 D5, Fig. 1), perpendicularly to the facade plane, as well as the temperature of the outer and inner panel surfaces and the ambient air temperature. Fig. 6 shows the panelõs outer surface temperature variation in two different points, respectively, near displacement measuring points D1 and D4, together with the outer air temperature. The inner air temperature had a constant value of 23 C, due to the air conditioning system. The results in Fig. 6 show that until 12:00 h, when the facade panel was not exposed to direct solar radiation, its outer surface temperature was identical to air temperature. After that period, when the panel was subjected to direct radiation, its temperature increased to significantly higher values. When the direct radiation stopped, Fig. 5. Displacement measurements during the water test (positive towards the interior).

4 4 J.R. Correia et al. / Construction and Building Materials xxx (2005) xxx xxx T empera ture (ºC ) h10 10h30 11h15 11h20 12h00 12h41 13h45 14h40 15h07 Time T,amb T,D4 T,D1 15h40 16h07 16h45 16h55 18h00 18h18 18h50 Fig. 6. Temperature measurements during the heating test. at about 16:00 h, the surface temperature decreased again. The maximum panel outer surface temperature was 58 C for an air temperature of 33 C and an inner surface temperature of 28 C. This situation corresponds to a maximum differential temperature of 30 C between the two GRC sheets. Fig. 7 shows the displacement measurements during the test. The displacements variation approximately followed the surface temperature variation in each measuring point. The higher displacements occurred in points D2 and D5, in the centre of the bottom and top zones of the panel Cooling test To understand the panels crack behaviour, a cooling test was also performed, consisting of measuring the opening of a crack (Fig. 8) when the outer surface temperature decreased from 35 to 24 C. Under this temperature variation, the crack opening increased of 0.35 mm. This result showed that, when the sandwich panel deforms towards the exterior during the outer surface heating, in plane compression stresses are generated and the opposite occurs when the outer surface is cooled. This suggests that the cracks at the outer Displacement (mm) h10 10h30 11h15 11h20 12h00 12h41 13h45 14h40 Time 15h07 15h40 16h07 16h45 16h55 D1 D2 D3 D4 D5 18h00 18h18 18h50 Fig. 7. Displacement measurements during the heating test (positive towards the exterior of the facade). Fig. 8. Crack opening measure during the cooling test. GRC sheet were caused by a differential temperature consisting of outer surface cooling for a constant internal air temperature, which happens especially during winter Tension and compression tests Considering the difficulties in performing direct tension tests in GRC specimens, the tensile resistance was determined by means of 3 point flexural tests on 6 small specimens, extracted from one existing facade panel. An average tension resistance of 9.8 MPa was determined. The resistance to direct tension is about 40% of the resistance obtained in flexural tests [4]. This led to an estimated value of 3.9 MPa for the average direct tension resistance of the tested GRC. For design purpose, and considering the variability of the results, the characteristic direct tension resistance was computed as 2.5 MPa. Compression tests were also performed in 4 GRC specimens obtained from the same facade panel. An average compression strength of 78.0 MPa was obtained from the samples. 4. FEM model of the facade panels 4.1. Model characteristics The experimental tests showed that the main cause of the extensive cracking observed in the facade panels is the effect of the temperature. This is a typical problem of imposed strains that could be avoided, if the panelsõ connections (GRC sheets interconnections and panel to structure fixings) were less stiff. An FEM model of the facade panels was then developed using the SAP 2000 FEM program [5], first in order to simulate the pathologies and check the assumed

5 J.R. Correia et al. / Construction and Building Materials xxx (2005) xxx xxx 5 causes and later to evaluate the efficiency of the solutions considered for the rehabilitation of the facade. The GRC facade sheets and laminas were modelled using adapted DKQ (Discrete Kirchhoff Quadrilateral) and DKT (Discrete Kirchhoff Triangular) shell elements, which consist of 4-node rectangular/triangular elements with bi-linear interpolation functions [5 8], following the geometry previously described (Fig. 9). The actual supporting conditions of the facade panels were simulated with node constraints, consisting of simple supports (free rotation) in the bottom sandwich element that is actually supported by the floor slab and in the nodes corresponding to the points with fixing steel elements. The value considered for the elastic modulus was 25 GPa, slightly higher than that of a young GRC [4], to take into account the strengthening effect caused by ageing (the material is 20 years old) and to provide a safety factor for the rehabilitation solution proposed. The PoissonÕs ratio was considered as 0.2 which is the current value for this mechanical property. The self weight of the panels was taken as 20 kn/m 3, based on measured values Simulation of defects The cracking of the GRC sheets was found to be due to the thermal differentials. This situation was simulated with the FEM model exposed to the experimentally evaluated maximum thermal differential of about 30 C. The winter peak temperature differential, condition for the outer sheet cracking, was not measured. In order to determine a maximum thermal differential in this situation, a classic thermal conductivity model was used, considering a steady-state regime, taking into account the whole constitution of the wall (a GRC sandwich panel, an air box and a ceramic brick pane Fig. 10). For this thermal analysis, the following thermal properties of the materials were considered: thermal conductivities of 1.75 W/(m C) for the GRC and W/ (m C) for the insulation material; thermal resistance of 0.17 m 2 C/W for the air box and a thermal conductance of 5.26 W/(m 2 C) for the brick wall. A minimum exterior air temperature of 0 C(h e ) was considered, while in the interior of the brick wall a value of 23 C(h i ) was adopted. In this situation, the average temperature for the external GRC sheet is 0.4 and 18.6 C for the internal one. The effects of these two extreme thermal situations were then evaluated with the FEM model, where the above temperature differentials, in each GRC sheet, were considered. For the summer situation, the following maximum tension stresses were obtained (Fig. 11): horizontal stresses of 2.8 MPa in the top and bottom zones of the inner GRC sheet, above and under the window; vertical stresses of 2.6 MPa in the lateral zones of the inner GRC sheet, beside the window; stresses of about 6.0 MPa in the horizontal and vertical laminas that connect the inner and outer GRC sheets. These values, higher than the tension resistance of 2.5 MPa experimentally determined, explain why the inner GRC sheet presents extensive cracking as well as the connecting laminas, leading to the separation of the GRC sheets. For the winter situation, the following maximum tension stresses were obtained (Fig. 12): horizontal stresses of 4.0 MPa in the top zone of the outer GRC sheet, above the window; vertical stresses of 4.0 MPa in the lat- GRC Insulation GRC Air box Brick wall θ2 θ3 θ4 θsi θi θe θse θ1 Fig. 9. 3D view of FEM meshes of the facade panels. Fig. 10. Thermal conductivity model winter situation (dimensions in mm).

6 6 J.R. Correia et al. / Construction and Building Materials xxx (2005) xxx xxx Fig. 11. Horizontal (left) and vertical (right) stresses in the facade panel (view from the interior) for the summer situation (values in MPa, tension positive). Fig. 12. Horizontal (left) and vertical (right) stresses in the facade panel (view from the exterior) for the winter situation (values in MPa, tension positive). eral zones of the outer GRC sheet, aside the window; in plane stresses of about 4.5 MPa in the vertical sandwich supporting elements. For the winter situation, these tension stress values are also higher than the tensile strength of 2.5 MPa, confirming that the cause of the extensive cracking pattern observed in the outer GRC sheets was the winter thermal gradient. In all the situations analysed, the maximum compression stresses were not critical, as the compression strength, experimentally determined, was higher. 5. Analysis of the rehabilitation solution In order to avoid the effect of the thermal gradients on the facade panels different rehabilitation solutions were considered. The solution that proved to be more efficient consisted of, after performing the crack injection, covering the outer GRC sheets with 40 mm thick insulation EPS panels (ETICS), mechanically fixed by means of stainless steel bolts. These panels are provided with a glass fibre grid and are to be painted with an elastomeric coating. The effect of these elements is to insu-

7 J.R. Correia et al. / Construction and Building Materials xxx (2005) xxx xxx 7 late the outer GRC sheet, reducing the temperature gradient between the two sheets of the sandwich panel (see Fig. 13). Fig. 13. ETICS insulation solution [9]. To numerically evaluate the efficiency of this solution, the temperatures of the GRC sheets were obtained using the classical thermal conductivity model, considering the effect of the EPS panel (Fig. 14). For the summer peak situation, a maximum differential of 20 C was obtained instead of 30 C without additional insulation. For the winter peak situation, a maximum differential of 13 C was obtained instead of 18 C without additional insulation (Fig. 15). Based on the formerly described FEM models, using these new thermal differentials and the additional self weight (of only kn/m 2 ), the stresses developed in the insulated panels were obtained. For the summer situation, the maximum tension stresses were less than 2.0 MPa at the inner sheet of the panel. For the winter situation, maximum tension stresses less than 2.5 MPa were obtained in the outer sheets of the panel, except for singular zones where stresses of 3.0 MPa occurred. For both summer and winter situations, tension stresses up to 4.8 MPa were obtained in the GRC horizontal and vertical laminas connecting the two GRC sheets. For that reason additional non-corrodible connecting elements (stainless steel or glass fibre composite bars) were also considered in the rehabilitation solution, to ensure the monolithic functioning of the facade panels. ETICS GRC Insulation GRC Air box Brick wall θ2 θ3 θ4 θsi θi θ0 θ1 θe θse Fig. 14. Thermal conductivity model for the rehabilitation solution (dimensions in mm). Fig. 15. Thermal conductivity model for the rehabilitation solution (winter situation with and without ETICS).

8 8 J.R. Correia et al. / Construction and Building Materials xxx (2005) xxx xxx 6. Conclusions This paper presents the study of a rehabilitation of a buildingõs GRC facade where extensive cracking and significant deformation were observed due to thermal effects. FEM models were used to simulate the structural behaviour of GRC panels, of which mechanical properties were experimentally determined. These models allowed for the identification of the actions that caused the observed pathologies, namely temperature gradients associated with excessively rigid panels connections (GRC sheets interconnections and panel to structure fixings). The FEM models were then used to evaluate the efficiency of a rehabilitation solution, with external insulation panels, which reduced the maximum thermal differential in the GRC elements and were easy to apply. According to the study, the proposed solution proved to be effective, reducing the maximum tension stresses to values that in general are compatible with the GRC experimentally determined tensile strength. References [1] Bentur A, Mindess S. Fibre reinforced cementitious composites. Amsterdam: Elsevier Applied Science; [2] Majumdar A, Ryder J. Glassfibre reinforcement for cement products. Glass Technol 1968;9: [3] Cem-FIL GRC Technical Data. Cem-FIL International Ltd, Vetrotex, UK; [4] Knowles E. Recommended practice for glass fibre reinforced concrete panels. Committee on Glass Fibre Reinforced Concrete Panels, PCI, USA; [5] SAP 2000 UserÕs Manual (version 7.40). [6] Zienkiewicz OC, Taylor RL. The finite element method volume 2. Solid and fluid mechanics. Dynamics and non-linearity. London: McGraw-Hill; [7] Cook RD, Malkus DS, Plesha ME. Concepts and applications of the finite element analysis. New York: Wiley; [8] Batoz JL, Tahar MB. Evaluation of a new quadrilateral thin plate bending element. Int J Numer Meth 1982;18: [9] Branco FA, Ferreira J, Correia JR. Safety evaluation and definition of inspection and maintenance measures for the facade panels of the CGD building. Technical Report ICIST EP No. 47/03 (in portuguese). Lisbon: ICIST; 2003.