https://doi.org/10.14311/APP.2022.33.0058 Acta Polytechnica CTU Proceedings 33:58–62, 2022 © 2022 The Author(s). Licensed under a CC-BY 4.0 licence Published by the Czech Technical University in Prague DURABILITY OF GFRP CONNECTORS UNDER SUSTAINED COMPRESSION LOAD FOR USE IN SANDWICH WALLS Stefan Carstens∗, Matthias Pahn Technische Universität Kaiserslautern, Department of Civil Engineering, Erwin-Schrödinger-Straße 52, 67663 Kaiserslautern, Germany ∗ corresponding author: stefan.carstens@bauing.uni-kl.de Abstract. Precast concrete sandwich panels are used to fulfil the rising thermal requirements. The sandwich walls consist of three layers, a facing, a thermal insulation layer, and a load bearing layer. The two outer layers are coupled by connectors made of glass-fibre reinforced polymer. A lack of knowledge about load-bearing behaviour prevents the removal of sustained compressive loads. In the context of this article, tests under sustained compressive load are presented. To represent closely the in service-conditions of sandwich walls, the examined connectors were subjected to a saturated alkaline concrete environment as well as to a specified stress level till failure occurs. Thus, the experimental setup combines alkaline resistance and creep rupture tests into one comprehensive testing. By using temperature effects as an accelerating factor, reasonable test durations were enabled. The obtained time to failure line was determined to extrapolate the characteristic values of the long-term strength for a service life up to 50 years. The test results are compared and evaluated with existing test results under a sustained tensile load. Keywords: Durability, GFRP connector, sandwich wall sustained compression load. 1. Introduction Precast concrete sandwich panels have been used as exterior wall systems for many years and have proven themselves in practice [1]. They are characterised by numerous design options (Fig. 1) as well as high resistance to weathering and ageing [2]. The pre- fabricated wall elements have a three-layer structure consisting of a facing layer, a factory-installed ther- mal insulation layer, and a load-bearing layer. In standard cases, the facing and load-bearing layers are made of normal concrete [3] and in individual cases of lightweight concrete [4] or high-performance concrete [5]. The facing layer serves as an architec- tural design element that is at least 70 mm thick [6] and fulfils the function of weather protection for the thermal insulation material. The load-bearing layer performs load-bearing functions and is at least 140 mm thick. The thermal insulation materials used are expanded polystyrene (EPS) according to DIN EN 13163 [7], extruded polystyrene (XPS) according to DIN EN 13164 [8], and insulation materials based on polyurethane (PUR) according to DIN EN 13165 [9]. Usual insulation thicknesses are between 60 to 240 mm. Discrete connectors made of corrosion-resistant materials are used to connect the facing and the load bearing layer [6]. These must allow deforma- tions of the facing layer due to thermal effects as free as possible from constraining forces and simul- taneously transfer wind and dead-loads across the thermal insulation [10]. To avoid thermal bridges, metallic connectors such as stainless-steel connectors are increasingly being replaced by connectors made Figure 1. Precast concrete sandwich panels for use as a façade [5]. out of fibre-reinforced plastic (FRP). The connectors usually consist of a thermosetting plastic matrix and unidirectional reinforcing glass fibres. The individual components perform different tasks within the composite [11]. The fibres used are thin (! 3 . . . 25 mm), unidirectional oriented continuous glass fibres, which are completely enclosed by a syn- thetic resin matrix. Due to their mechanical prop- erties, the inorganic fibres have a decisive influence on the stiffness and strength of the composite ma- terial [12]. Thermosetting polymers such as epoxy resins (EP), vinyl ester resins (VE), and vinyl ester urethane resins (VEU) are mainly used as materi- als for the synthetic resin matrix. The matrix fixes the fibres in the desired geometric position, transfers the stress between the fibres, prevents the fibres from buckling under a compressive load, and protects the fibres from external mechanical and chemical stress. The advantages of the individual components are combined by the interaction of the glass fibre and the polymer matrix. In addition to the properties of the components and their proportions, pultrusion 58 https://doi.org/10.14311/APP.2022.33.0058 https://creativecommons.org/licenses/by/4.0/ https://www.cvut.cz/en vol. 33/2022 Durability of GFRP Connectors Name Isolink ThermoPin TM-MC Anchor Matrix material VEU EP VE Fibre material wt. % 87.1 n/a 74.0 vol. % 74.3 n/a 53.0 Surface ripped wound flat Min. tensile strength N/mm2 1.000 1.500 700 Tensile Young’s modulus N/mm2 60.000 60.000 40.000 Section dimensions mm !c 12.0; !e 13.5 !e 7.3 h/w 9.8/5.7 Table 1. Characteristics of GFRP Connectors acc. to [13]. [12] and the specific interaction of matrix resin and reinforcing fibre are decisive for the properties of the composite material. Connectors made of glass-fibre reinforced plas- tic differ significantly in terms of material composi- tion and cross-sectional geometry. Table 1 gives an overview of the material and geometry characteristics of different GRP connectors. 30° Ø �1 2� Ø �1 3. 5� Figure 2. The geometry of the examined connector’s head made out of glass fibre reinforced polymers. In the further course of this paper, the Isolink [14] shown in Figure 2 will be considered as the connector for sandwich wall elements. The connector has a core diameter of 12 mm made out of glass fibre reinforced polymers. The total diameter of the connector, core, and ribs, is 13.5 mm. The ribs are made by indented trapezoidal thread with a profile depth of 0.6 to 0.75 mm and a pitch of 8 mm. The ends of the connector are bevelled at an angle of 30◦. In addition to temporary loads due to wind and temperature effects, the connectors must also be withstanding permanent loads due to the dead load of the facing layer. Loads acting vertically on the fac- ing layer are carried by horizontally arranged anchors. The dead load of the facing layer is transferred using load-bearing anchor systems consisting of horizontal and diagonal anchors. The support anchor system consisting of horizontal and diagonal anchors and the associated static sys- tem for calculating the distribution of forces in the support anchor system under the dead load of the facing layer are shown in Figure 3. As a result of the force-deflection, tensile forces arise in the diago- nal connector Nt,sust and compressive forces in the horizontal connector Nc,sust which have a permanent effect. 2. Load-bearing behaviour under short-term compression load While the load-bearing behaviour of the connectors under short-term [15] and long-term tension load [16] is known, there is a lack of knowledge about the load- bearing behaviour under compressive load, especially the stability load-bearing behaviour under consider- ation of exposures (humidity, temperature and alka- linity). In the context of this paper, therefore, the results of experimental investigations into the load- bearing behaviour under compressive load with large insulation thicknesses are presented. During the test, the connectors are subjected to short-term and long- term pressure loads. In evaluate the load-bearing behaviour under short-term compressive loading, compression tests were carried out at insulation thickness of 160 mm, 250 mm, 300 mm and 350 mm [17]. The specimen geometry is shown in Figure 4. To ensure a realistic anchorage of the connectors in the facing and load bearing layer, the connectors are embedded in nor- mal concrete. The thickness of each concrete layer is 120 mm, the other cross-sectional dimensions are 300 × 300 mm. The core layer consists only of the connector, no thermal insulation is used. The test specimen is loaded during the test pro- cedure using a servo-hydraulic test cylinder con- trolled by displacement. The transverse speed is 0.5 mm/min. Simultaneously, the force is measured us- ing a load cell and the deformation through inductive displacement transducers. The determined loads are shown in Figure 5. With an insulation thickness of 160 mm, the connectors fail locally in the clamping area combined on thrust and pressure. With other in- sulation thicknesses, an interlaminar shear crack oc- curs over the entire length of the connector. Furthermore, Figure 5 shows the recalculation of the stability tests according to equation 1. A double- sided clamping of the connector is assumed (Euler Case IV, K = 0.5). The second moment of inertia is using the core diameter !c = 12 mm without consid- ering a stabilizing effect of the ribs. Pcr = π2EI (KL)2 (1) 59 Stefan Carstens, Matthias Pahn Acta Polytechnica CTU Proceedings Figure 3. Support anchor system (left) and static system (right) to calculate the distribution of forces in the support anchor system due to the dead load of the facing layer VEd. Figure 4. Geometry of the test specimens for short- and long-term tests under compression load. It can be shown that the results of the experimental investigations for the insulation thicknesses 250 mm, 300 mm and 350 mm correspond very well with the recalculation. With an insulation thickness of 160 mm, the load-bearing capacity is overestimated by the recalculation. The reason for this is the change in the failure mode described above. 3. Experimental investigation under sustained compression load The mechanical and physical properties of the com- posite material can degrade over time due to physi- cal effects from e.g. creep and environmental influ- ences. Since the transfer of the dead load of the facing layer must be ensured over the entire service life of the façade, these influences must be taken into ac- count when determining permissible load-bearing ca- pacities. The connectors are particularly exposed to temperature and humidity. Furthermore, the alkalin- ity of the concrete can damage the matrix. Figure 5. Experimentally and numerically deter- mined the load-bearing capacity of the connector un- der compression load as a function of the insulation thickness. To evaluate the load capacity of the connectors un- der compressive load, creep tests to failure are carried out to determine the respective service life. 3.1. Experimental set-up and execution of experiments under sustained load The test setup for the creep tests is shown in Figure 6. The test specimens described above are placed in a tempered water bath. The water level reaches to just below the upper edge of the lower concrete abutment. The water is held at 60 řC to simulate increased sur- faced temperatures in the area of the façade. The increased temperature causes water vapour to form, which settles on the connector and loads to a moder- ate moisture load. The force is applied by a centrally arranged hy- draulic cylinder. The hydraulic pressure in the cylin- 60 vol. 33/2022 Durability of GFRP Connectors 1 23 4 Figure 6. Experimental set-up for compression tests under continuous load and exposure to humidity, tem- perature and alkalinity; 1 hydraulic jack and oil pres- sure sensor; 2 test specimen; 3 temperature regula- tion; 4 stainless-steel profiles for guiding the speci- men. der is applied by a manual hydraulic pump and mea- sured by an oil pressure sensor. The deformations of the specimen are measured simultaneously by an in- ductive displacement transducer. The concrete abut- ments of the test specimens are guided in parallel by a stainless-steel construction to prevent later deflec- tions. 3.2. Results of experimental investigations under sustained load The time-deformation diagram in Figure 7 shows the course of the deformation over the load duration. The three technical creep ranges are marked. The elastic deformation of the test specimen can be seen at the beginning. In the initial area, the test speci- men shows an increase in linear deformation (primary creep area). The test specimen is then subjected to constant creep (secondary creep). This is followed by an increasing creep deformation (tertiary creep) which is finally followed by the failure of the test spec- imen. Compared to the load capacity of 18.9 kN shown in Figure 5, a reduction of the load capacity to 16 kN can be determined for a service life of approximately 100 h. This reduction is probably due in particular to the influence of temperature and humidity. 3.3. Evaluation of the results of experimental investigations Further service lives are achieved by varying the loads. Figure 8 shows the previous service lives of running and finished tests and the associated loads. With decreasing load, an increase in service life is ex- pected. From the service lives of the individual tests Figure 7. Deformation of the test specimen over the test period at an insulation thickness of 350 mm. and loads, a permissible load for the intended service life of 50 years is then to be derived by extrapolation. The extrapolation should be performed according to [18]. The calculated load capacity for an insulation thickness is on the safe side as insulation thicknesses range from 60 to 350 mm. Figure 8. Load - time to failure diagram of creep tests at an insulation thickness of 350 mm. 4. Conclusions Multilayer concrete sandwich panels have been used as exterior wall systems for many years and have been proven themselves in practice. Connectors made of glass fibre reinforced polymers are increasingly being used to couple the concrete layers. To transfer the dead load of the facing layer through a support anchor system consisting of connectors, the connectors must permanently absorb compressive loads. In this paper, experimental investigations under short-term and long-term compression load are pre- sented. It can be shown that the tests with insulation 61 Stefan Carstens, Matthias Pahn Acta Polytechnica CTU Proceedings thicknesses between 250 and 350 mm can be recalcu- lated very well using the Euler equation. In the long- term tests, an increased influence of temperature and humidity on the failure load can be derived from the initial results. Finally, a concept for extrapolation is proposed for the determination of permissible load capacities. Acknowledgements The authors thank Schöck Bauteile GmbH for the finan- cial and technical support provided. References [1] M. Pahn. Beitrag zur Ermittlung von Schnitt- und Verformungsgrößen bei mehrschichtigen Stahlbetonwandtafeln mit Verbindungsmitteln aus glasfaserverstärktem Kunststoff Kaiserslautern Germany, Ph.D. thesis, Kaiserslautern, Deutschland, 2011. [2] R. Gastmeyer. Neuentwicklungen bei der Konstruktion mehrschichtiger Stahlbeton-Wandtafeln mit integrierter Wärmedämmung. Beton- und Stahlbetonbau 98(6):360- 8, 2003. https://doi.org/10.1002/best.200301750. [3] M. Pahn, J. Schnell. Einfluss der Verbundtragwirkung bei mehrschichtigen Stahlbetonwandtafeln mit innen liegender Wärmedämmung. Beton- und Stahlbetonbau 106(8):551-60, 2011. https://doi.org/10.1002/best.201100040. [4] F. Hanz. Beitrag zur Bemessung von dreischichtigen Wandtafeln mit Deckschichten aus haufwerksporigem Leichtbeton Kaiserslautern, Ph.D. thesis, Kaiserslautern (Deutschland), 2017. [5] F. Müller, C. Kohlmeyer, J. Schnell. Sandwichelemente mit Deckschichten aus Hochleistungsbeton und einem Kern aus extrudiertem Polystyrol. Beton- und Stahlbetonbau 110(12):799-810, 2015. https://doi.org/10.1002/best.201500038. [6] DIN EN 1992-1-1/NA 2013. Nationaler Anhang - National festgelegte Parameter - Eurocode 2: Bemessung und Konstruktion von Stahlbeton- und Spannbetontragwerken - Teil 1-1: Allgemeine Bemessungsregeln und Regeln für den Hochbau, 2013. [7] EN 13163 2016. Thermal insulation products for buildings - Factory made expanded polystyrene (EPS) products - Specification, 2016. [8] EN 13164 2015. Thermal insulation products for buildings - Factory made extruded polystyrene (XPS) products - Specification, 2015. [9] EN 13165 2016. Thermal insulation products for buildings - Factory made rigid polyurethane foam (PU) products - Specification, 2016. [10] P. Bindseil. Stahlbeton Fertigteile nach Eurocode 2 - Konstruktion, Berechnung, Ausführung, Werner Verlag, 2012. [11] A. Schumann, M. May, M. Curbach. Carbonstäbe im Bauwesen. Beton- und Stahlbetonbau 113(12):868-76, 2018. https://doi.org/10.1002/best.201800077. [12] H. Schürmann. Konstruieren mit Faser-Kunststoff-Verbunden, Springer Verlag, 2017. [13] DAfStb 615 (en). Commentary to EN 1992-4 - Design of Fastenings for use in Concrete, 2019. [14] Zulassung Z-21.8-1894. Schöck Isolink für mehrschichtige Betontafeln, 2018. https://www.dibt.de/de/service/ zulassungsdownload/detail/z-218-1894. [15] F. Micelli, A. Nanni. Mechanics of Composite Materials 39(4):293-304, 2003. https://doi.org/10.1023/a:1025638310194. [16] M. L. Keller, M. Pahn. Durability of GFRP Bars with Different Bar Diameters. IABSE Symposium, Guimaraes 2019: Towards a Resilient Built Environment Risk and Asset Management, p.603-10, 2019. https://doi.org/10.2749/guimaraes.2019.0603. [17] Bericht 06046Pa/528. Untersuchungen zur Eignung von ComBAR - Hohlwandankern aus glasfaserverstärktem Kunststoff als Verbindungselement von kerngedämmten Wänden mit aufstehender Vorsatzschale unreleased Bericht 10012Pav/14511 2015 Versuche zur Bestimmung des Widerstandes gegen Stabilitätsversagen von Schöck Thermoankern (unreleased), 2008. [18] A. Weber. Prüfkonzepte für Bewehrungsmaterialien mit zeitabhängigen Widerständen Bauingenieur, 2018. 62 https://doi.org/10.1002/best.200301750 https://doi.org/10.1002/best.201100040 https://doi.org/10.1002/best.201500038 https://doi.org/10.1002/best.201800077 https://www.dibt.de/de/service/zulassungsdownload/detail/z-218-1894 https://doi.org/10.1023/a:1025638310194 https://doi.org/10.2749/guimaraes.2019.0603