https://doi.org/10.14311/APP.2022.33.0098 Acta Polytechnica CTU Proceedings 33:98–104, 2022 © 2022 The Author(s). Licensed under a CC-BY 4.0 licence Published by the Czech Technical University in Prague TRC SANDWICH SOLUTION FOR ENERGY RETROFITTING Isabella Giorgia Colombo, Matteo Colombo, Marco di Prisco∗ Politecnico di Milano, Department of Civil and Environmental Engineering, P.za L. da Vinci, 32 - 20133 Milano, Italy ∗ corresponding author: marco.diprisco@polimi.it Abstract. Concerning energy improvement of existing façades, a favourable system involves prefabricated multilayer panels, made of internal insulation core and outer textile reinforced concrete layers. It is a convincing alternative to external thermal insulation composite systems (ETICS) and ventilated façades, and it meets all the requirements for façade systems. The main advantage is the possibility to apply the panel using a crane, without any scaffolding. The paper considers two solutions: the former uses expanded polystyrene (EPS) as insulating material; the latter substitutes EPS with an innovative green insulation material made of inorganic diatomite. The paper aims at comparing the solutions in terms of mechanical properties of the components and behaviour of the composite sandwich at lab-scale level. Numerical models, previously calibrated, will be instrumental for the discussion. Keywords: Energy performance, sandwich panels, textile reinforced concrete. 1. Introduction In Europe, building sector uses a large amount of en- ergy (40% of the total) and produces relevant CO2 emissions (36%). For this reason, the focus of many European policies is to improve energy efficiency [1]. Each year, only approximately 1% of buildings are en- ergy renovated, in the framework of a building stock that is old and energy inefficient, with the half of res- idential buildings dated before 1970, year in which the first European thermal Standard was emanated [2]. Hence, improvement in energy efficiency of build- ings is a crucial topic. In multi-storey buildings, façades cover a large amount of the total outer envelope surface. Hence, solutions devoted to the energy retrofitting of existing façade result particularly effective in terms of build- ing energy saving. A convenient solution is repre- sented by multilayer precast sandwich panels made of an inner insulation core and thin external textile reinforced concrete (TRC) layers. TRC sandwich panels were proposed about 15 years ago by German researchers [3] and constituted a new idea in the framework of cement-based sandwich elements: TRC thickness (10-15 mm) is smaller than that of traditional R/C external claddings (>40 mm) [4], leading an important decrease in the weight of the panel. Other conveniences are durability and the good appearance, obtained using fine-grained mor- tar. Although connectors were introduced between the layers for a durable connection [5], advantage is taken from adhesive bond between TRC and the in- sulation layer to transfer the shear stresses. 2. TRC sandwich solutions The research concerns prefabricated sandwich panels constituted by thin external TRC layers and an in- ternal insulation core. Large size panels (3 × 1.5 m2) have been designed to reduce the manufacturing and installation costs. The installation is performed through a crane, without scaffolding, thus minimiz- ing installation timing and inconveniences for inhab- itants. Other advantages are: the integration of the insulation and finishing systems; the durability and possibility to obtain different colour and tex- ture as finishing; the protection of the insulation core through the external, waterproofed layer of TRC; the good resistance to concentrated loads. Only the inner TRC layer is connected to the existing R/C frame; the shear transfer between the two TRC layers is en- trusted to the insulation material (for fire safety, some supplementary connectors could be inserted). 2.1. Panel with expanded polystyrene core (EASEE project) The Consortium of the European project EASEE [6] developed a prefabricated TRC sandwich panel with expanded polystyrene (EPS) as insulation material. A thickness of 100 mm was considered for the core. All the details concerning the panel design, testing at different scale levels, durability, and modelling can be found in [7–12]. EASEE main output is repre- sented by a demonstration building [11] in Cinisello Balsamo, Italy, where a 70% decrease in the thermal transmittance of the wall was achieved [13]. One of the main criticisms of this panel is that EPS may melt during fire, leading to the loss of mechanical perfor- mances of the composite solution. 2.2. Panel with diatomite core (Smart P.I.QU.E.R.) Smart P.I.QU.E.R. project [14] was aimed at solving the criticisms of the previous project. Among the 98 https://doi.org/10.14311/APP.2022.33.0098 https://creativecommons.org/licenses/by/4.0/ https://www.cvut.cz/en vol. 33/2022 TRC Sandwich Solution for Energy Retrofitting Figure 1. Test results: TRC tensile behaviour (a) and core material compressive behaviour (b). goals, particular relevance is assumed by the devel- opment of a green insulation material, with suitable mechanical properties and good fire and insulating capabilities. At the same time, TRC layers need to be optimised, reducing the shrinkage and weight of the mortar, and enhancing the fabric fire behaviour and sustainability. 2.3. Materials 2.3.1. Textile reinforced concrete TRC developed in the EASEE project is made by high strength fine grain mortar reinforced with one alkali-resistant glass fabric. The warp is aligned in longitudinal direction. This fabric [7] (leno-weave production; SBR coating: water resin based on styrene butadiene rubber) is characterised by a nom- inal tensile strength in warp direction of 820 MPa (computed on the equivalent cross-sectional area of the glass). The cementitious matrix has a water to binder ratio equal to 0.225, a superplasticiser to ce- ment ratio of 9.3% and 1 mm maximum aggregate size. The average cubic compressive strength (fcc) of mortar, measured as specified in [15], is equal to 71.89 MPa on 12 specimens, with a coefficient of variation of 9.13%. Concerning Smart P.I.QU.E.R. project, several mortar mix-designs and alkali-resistant glass fabric reinforcements were investigated [14]. In this paper, a high-performance mortar that includes shrinkage reducing admixtures and is reinforced with one alkali- resistant glass fabric coated with a thermosetting epoxy coating is considered (see "HPC+SRA_AP1" in [14]). The epoxy coating is particularly suitable in case of fire exposure. The fabric is characterised by a nominal strength in the warp direction of 1041 MPa. The cementitious matrix used is characterised by a maximum aggregate size of 1 mm, a water to binder ratio of 0.20 and a super-plasticiser to cement ratio of 5.1%. The average cubic compressive strength (fcc) of mortar measured according to [15] is equal to 97.53 MPa on 32 specimens, with a coefficient of variation of 8.33%. In both cases, three 400×70×10mm3 TRC samples were tested in uniaxial tension using an electrome- chanical press INSTRON 5867 with a maximum load capacity of 30 kN, imposing a constant crosshead dis- placement rate of 0.02 mm/s. The same test appa- ratus and modalities described in [7] are used. The test results are shown in Figure 1a in terms of nominal stress (σ = P/A, with P=load and A=specimen cross section) vs. normalised displacement (δ/l, with δ = imposed displacement and l = free length) curves. 2.3.2. Core material In the framework of the EASEE project, expanded polystyrene was used as insulation material in sand- wich panels. The product EPS250, characterised by a nominal density of 35 kg/m3, was selected. As result- ing from experimental tests [10], it is characterised by a uniaxial compressive yield stress of 0.188 MPa, an estimated Young’s modulus in compression of 13.7 MPa, a uniaxial tensile yield stress of 0.392 MPa, a shear yield stress of 0.160 MPa and an estimated shear modulus of 5.04 MPa. As already anticipated, one of the main aims of the Smart P.I.QU.E.R. project was to develop a core material that could be more sustainable and more suitable in case of fire with respect to EPS. Hence, a novel lightweight sustainable insulation core (DHCF) was developed by Verdolotti and other authors [16] using diatomite (natural source) as matrix. Short fi- bres were added to guarantee a ductile behaviour of the insulation material both in compression and in bending, in order to guarantee a proper global be- haviour of sandwich panels. These fibres also con- tribute to the enhancement of thermal performances of the material, by reducing the thermal conductivity. Polypropylene (PP) fibres were selected among dif- ferent kind of reinforcement tested (jute, waste syn- thetic, mixed and hemp fibres, as such and chopped). 99 I. G. Colombo, M. Colombo, M. di Prisco Acta Polytechnica CTU Proceedings Figure 2. Four-point bending test set-up (Smart P.I.QU.E.R.). This choice was related not only to the proper ther- mal and mechanical performances of the foam in which PP fibres are included, but also because they are the only ones that allow the production at indus- trial level by means of an industrial plant. Consid- ering that the level of the plastic plateau is related to specimen density, a target density of 300 kg/m3 was fixed to guarantee comparable performances of the core material with those of expanded polystyrene. The core mix design is collected in Table 1. Diatomite (wt%) 19.57 Polysilicate (wt%) 65.24 Si (wt%) 0.046 Catalyst (wt%) 8.39 Vegetable (wt%) 5.82 PP fibres (wt%) 1 Table 1. DHCF_PP mix design. Compressive tests were performed on EPS speci- mens with a size of 100 × 100 × 150 mm3, and DHCF specimens with a dimension of 40 × 40 × 40 mm3. The tests were displacement controlled by imposing a constant displacement rate of the machine crosshead equal to 1e-3 mm/s and 2e-3 mm/s respectively. The compressive behaviour of both expanded polystyrene and diatomite-based insulation material is shown in Figure 1b where the nominal stress σ vs. nominal strain ε curves are plotted (σ = P/A, P = load, A = unloaded specimen cross-section; ε = δ/h, δ = top displacement, h = specimen height). 3. Experimental behaviour of lab-scale sandwich beams with diatomite-based core A wide experimental campaign on lab-scale sand- wich specimens were developed within the EASEE project. A four-point bending set-up was adopted for both deep (550 × 150 × 120 mm3) and slender (1200 × 300 × 120 mm3) sandwich beams, with EPS 100 mm thick. Detailed description of experimental procedures and results can be found in [8]. Hence, this Section is focused on the experimental tests per- formed at lab-scale level within Smart P.I.QU.E.R. on beams in bending. 3.1. Test set-up Four-point bending tests were performed on sand- wich beams using the same electromechanical press INSTRON 5867 already mentioned. Two nominally identical specimens with size of 600 × 150 × 80 mm3 and the warp of the glass fabric aligned in the beam longitudinal direction were tested. The specimen ge- ometry and the test set-up are shown in Figure 2. The insulation panel densities are collected in Table 2. Specimen ρ1st_measure ρ2nd_measure(kg/m3) (kg/m3) PIQUER1 330 (31 days) 303 (125 days) PIQUER2 348 (28 days) 297 (122 days) Table 2. Density of the insulation core. The insulation layer of full-scale panels designed in the Smart P.I.QU.E.R. project is obtained plac- ing side by side different core panels with size of 600×600×100mm3. For this reason, some cuts are in- troduced in the insulation layer also in the lab-scale beams (Figure 2). With the aim of reducing stress concentration, 50 mm wide steel plates are glued on the specimen over the cylindrical supports and un- der the loading-knives. The loading and the sup- porting cylinders are free to rotate on their axis and in the plane perpendicularly to the longitudinal axis of the specimen. During the initial phase of each test, specimens were instrumented with displacement transducers in order to measure: the specimen deflec- tion under the loading knives, the deflection of the upper TRC layer under the loading knives, the su- perior longitudinal displacement on the compressed side inside the constant bending moment region, and the crack opening displacement on the tense side. However, these aspects are not treated in this pa- per. Tests were displacement-controlled, considering the machine crosshead displacement (stroke) as feed- back parameter. An initial stroke rate of 2E-3 mm/s 100 vol. 33/2022 TRC Sandwich Solution for Energy Retrofitting Figure 3. Experimental results (Smart P.I.QU.E.R.): load vs. stroke curves (a) and pictures of specimen PIQUER 2 during the test (b) and at failure (c). was imposed; then, after crack formation, the rate is increased up to 5E-3 mm/s. 3.2. Experimental results The test results are shown in Figure 3a in terms of load (P) vs. stroke curves. Looking at the results, it is interesting to note that specimen PIQUER1 ex- perienced a premature shear failure of the core, thus leading to a localised failure and preventing the multi- cracking of TRC layers that guarantees the hard- ening behaviour of the sandwich solution. In fact, while performing a check before testing, core panel used to cast specimen PIQUER1 presented - to the touch - some softer regions at the edges. On the contrary, specimen PIQUER2 showed a ductile hard- ening global response, experiencing multi-cracking of TRC faces. A picture of this specimen during testing is shown in Figure 3b. The final failure is related to the debonding between the beam core and the lower TRC layer (Figure 3c). 4. Numerical modelling of the tests and comparison between the two sandwich solutions 4.1. Geometry and constraints Abaqus /CAE 6.14-5 was adopted for the modelling of the panel. Yz symmetry was exploited, and half of the beam was reproduced, thus minimising the nu- merical effort (Figure 4a). As shown in the figure, the panel layers (TRC and core), and the steel plates are modelled as solid and homogeneous. Perfect bond is assumed at core/TRC interfaces and between the panel and steel plates. The interaction between adjacent core panels is char- acterized by a hard contact in the normal direction and by a frictionless contact in the tangential direc- tion. Concerning constraints, displacement orthogonal to the symmetry plane are prevented, the lower steel plate is constrained in the vertical direction on the bottom face and a vertical displacement δ is im- posed on the loading line. The mesh is shown in the same figure: eight-node linear hexahedral elements (C3D8R - Continuum, 3-D, 8-node - reduced integra- tion) are used. 4.2. Constitutive laws In this Section, a description of the constitutive laws adopted for each material is proposed. The elastic phase of TRC is defined introducing a Poisson’s ratio of 0.2 and a Young’s modulus of 30 GPa (according to literature results [17] on mortar characterized by equal maximum aggregate size and compressive strength). The plastic behaviour of TRC is modelled through Abaqus Concrete Damage Plasticity model. No dam- age curve is inserted; hence, the model behaves as a plasticity model. In compression, an elastic- perfectly plastic behaviour is assumed, imposing a yield strength of 97 MPa according to the values shown in Section 2.3.1. The stress-irreversible strain relationship (Table 3) adopted in tension is extracted from the experimental results of PIQUER specimens shown in Figure 1a, considering strains computed basing on the measurements of displacement trans- ducers applied on TRC. TRC tensile behaviour is as- sumed homogeneous over the layer thickness; this as- sumption is reliable according to [10]. Point Yield stress Cracking strain T1 2.5 0 T2 5 0.00275 T3 20.81 0.017 Table 3. TRC plastic tensile constitutive law. Concerning the core material, the elastic phase is defined introducing a Young’s modulus of 20 MPa 101 I. G. Colombo, M. Colombo, M. di Prisco Acta Polytechnica CTU Proceedings Figure 4. FE model: geometry and constraints (a) and validation of core material model considering three-point bending (b). Figure 5. FE model results: numerical response compared with experimental results (a) and with EASEE bending behaviour (b); failure modes for specimen PIQUER2 showing core strut yielding at point C1 (c) and TRC multi- cracking at point T3_sup (d). 102 vol. 33/2022 TRC Sandwich Solution for Energy Retrofitting (according to experimental results in compression shown in Figure 1b) and a Poisson’s ratio of 0.2. The plastic behaviour of the core material is in- troduced using Abaqus Concrete Damage Plasticity model. Also in this case, no damage curves are in- troduced. An elastic-perfectly plastic behaviour is assumed both in tension and in compression: a com- pressive yield strength of 0.30 MPa is used according to experimental results shown in Figure 1b (average value for specimens of the same batch of core sample PIQUER2); a tensile yield strength of 0.053 MPa is adopted according to average experimental results of tensile tests performed on core specimens. With the aim of checking the validity of the mate- rial model adopted for the core, a three-point bending test performed on core prismatic specimen with size of 160×40×40mm3 was modelled. Experimental and numerical curves are compared in Figure 4b, demon- strating that the adopted material model is reliable. Steel is assumed elastic with a Young’s modulus of 210 GPa and a Poisson’s ratio of 0.1. 4.3. Numerical results The numerical response is shown in Figure 5a in terms of load (P) vs. stroke. A good correlation is found with respect to the behaviour of specimen PIQUER2 that showed a hardening behaviour. A perfect over- lapping is obtained in the initial phase of the test; the divergency of the two curves while the test is progress- ing could be related to a bond weakness at TRC-core interfaces, leading to the final failure due to debond- ing already underlined in Figure 3c. In order to investigate the failure mechanisms oc- curred for a specimen characterised by a good TRC- core bond, critical points of material constitutive laws are underlined on the numerical curve. In particu- lar: T points refer to the TRC layers (Table 3), with subscript "inf" related to lower TRC layer and sub- script "sup" related to the upper TRC layer; point C1 indicates the yielding of the strut in the core mate- rial. It is worth noting that the change in the slope of the global response corresponds to point C1 (Fig- ure 5c), when both TRC layers are already cracked (Figure 5d). In Figure 5b the numerical response of PIQUER specimens, in which the core thickness was enhanced to 100 mm, is compared with the experimental be- haviour of EASEE sandwich beams characterised by the same cross-section and size. 5. Conclusions The bending behaviour of sandwich solutions de- veloped within the EASEE and Smart P.I.QU.E.R. projects was compared at lab-scale level. Even though the response of the PIQUER sandwich beam showed lower maximum load and final displacement, a ductile hardening behaviour and a reasonable sus- tained load are guaranteed by the solution. It is worth noting that a good bond at TRC-core interfaces must be guaranteed to obtain a proper behaviour of the composite. Acknowledgements The research was developed within the Smart P.I.QU.E.R. project, funded by Regione Lombardia (Smart Living; de- cree n. 14782 - 24/11/2017; ID SiAge: 379284). Authors would like to thank the partners; in particular, iPCB- CNR and Isoltech involved in the development of the core material. References [1] EUR-Lex. Directive 2018/844/EU of the European Parliament and of the Council of 30 May 2018 amending Directive 2010/31/EU on the energy performance of buildings and Directive 2012/27/EU on energy efficiency, 2018. https://eur-lex.europa.eu/l egal-content/EN/TXT/?uri=celex:32018L0844. [2] European Commission. Data from EU Building Database, 2019. https: //ec.europa.eu/energy/en/eu-buildings-database. [3] J. Hegger, M. Horstmann. Light-weight TRC sandwich building envelopes. Excellence in Concrete Construction through Innovation 187-194, 2008. https://doi.org/10.1201/9780203883440.ch27. [4] A. Einea, D. C. Salmon, G. J. Fogarasi, et al. State-of-the-Art of Precast Concrete Sandwich Panels. PCI Journal 36(6):78-98, 1991. https://doi.org/10.15554/pcij.11011991.78.98. [5] A. Shams, M. Horstmann, J. Hegger. Experimental investigations on Textile-Reinforced Concrete (TRC) sandwich sections. Composite Structures 118:643-53, 2014. https: //doi.org/10.1016/j.compstruct.2014.07.056. [6] European Commission. European project EASEE (2012-2016) Envelope approach to improve sustainability and energy efficiency in existing multi-storey multi-owner residential buildings, grant no.: 285540, 2019. [7] I. G. Colombo, M. Colombo, M. di Prisco. Tensile behavior of textile reinforced concrete subjected to freezing-thawing cycles in un-cracked and cracked regimes. Cement and Concrete Research 73:169-83, 2015. https: //doi.org/10.1016/j.cemconres.2015.03.001. [8] I. G. Colombo, M. Colombo, M. di Prisco. Bending behaviour of Textile Reinforced Concrete sandwich beams. Construction and Building Materials 95:675-85, 2015. https: //doi.org/10.1016/j.conbuildmat.2015.07.169. [9] I. G. Colombo, M. Colombo, M. di Prisco. TRC precast façade sandwich panel for energy retrofitting of existing buildings. ACI Sp. Pub. SP 305, 2015. https://doi.org/10.14359/51688590. [10] I. G. Colombo, M. Colombo, M. di Prisco, et al. Analytical and numerical prediction of the bending behaviour of textile reinforced concrete sandwich beams. Journal of Building Engineering 17:183-95, 2018. https://doi.org/10.1016/j.jobe.2018.02.012. 103 https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:32018L0844 https://ec.europa.eu/energy/en/eu-buildings-database https://doi.org/10.1201/9780203883440.ch27 https://doi.org/10.15554/pcij.11011991.78.98 https://doi.org/10.1016/j.compstruct.2014.07.056 https://doi.org/10.1016/j.cemconres.2015.03.001 https://doi.org/10.1016/j.conbuildmat.2015.07.169 https://doi.org/10.14359/51688590 https://doi.org/10.1016/j.jobe.2018.02.012 I. G. Colombo, M. Colombo, M. di Prisco Acta Polytechnica CTU Proceedings [11] I. G. Colombo, M. Colombo, M. di Prisco, et al. TRC sandwich panel for energy retrofitting exposed to environmental loading. ACI Sp. Pub. SP 326, 2018. https://doi.org/10.14359/51711059. [12] I. G. Colombo, M. Colombo, M. di Prisco. Precast TRC sandwich panels for energy retrofitting of existing residential buildings: full-scale testing and modelling. Materials and Structures 52(5), 2019. https://doi.org/10.1617/s11527-019-1406-1. [13] G. Salvalai, M. M. Sesana, G. Iannaccone. Deep renovation of multi-storey multi-owner existing residential buildings: A pilot case study in Italy. Energy and Buildings 148:23-36, 2017. https://doi.org/10.1016/j.enbuild.2017.05.011. [14] I. G. Colombo, M. Colombo, M. di Prisco, et al. Lightweight TRC sandwich panels with sustainable diatomite-based core for energy retrofitting of existing buildings. Advances in Building Energy Research 15(2):231-52, 2019. https://doi.org/10.1080/17512549.2019.1697752. [15] EN 196-1:2005. Standard for mortar. Methods of testing cement Part 1: Determination of strength, 2015. [16] L. Verdolotti, B. Liguori, I. Capasso, et al. Synergistic effect of vegetable protein and silicon addition on geopolymeric foams properties. Journal of Materials Science 50(6):2459-66, 2014. https://doi.org/10.1007/s10853-014-8801-3. [17] W. Brameshuber. Report of RILEM technical committee 201-TRC RILEM Publications, 2016. 104 https://doi.org/10.14359/51711059 https://doi.org/10.1617/s11527-019-1406-1 https://doi.org/10.1016/j.enbuild.2017.05.011 https://doi.org/10.1080/17512549.2019.1697752 https://doi.org/10.1007/s10853-014-8801-3