https://doi.org/10.14311/APP.2022.33.0052 Acta Polytechnica CTU Proceedings 33:52–57, 2022 © 2022 The Author(s). Licensed under a CC-BY 4.0 licence Published by the Czech Technical University in Prague APPLICATION OF HIGH PERFORMANCE CONCRETE FOR RECONSTRUCTION Iva Broukalová∗, Alena Kohoutková, Anna Horáková, Hana Schreiberová Czech Technical University in Prague, Department of Concrete and Masonry Structures, Faculty of Civil Engineering, Thákurova 2077/7, 166 29 Praha 6 - Dejvice, Czech Republic ∗ corresponding author: iva.broukalova@fsv.cvut.cz Abstract. One of the sustainability aspects of the construction is a sufficient durability. In objectively justi- fied cases, the durability is enhanced by reconstruction. To ensure economic design of the reconstruc- tion, non-liner approaches are often applied. This paper deals with aănonlinear numerical simulation of disassembled reinforced concrete (RC) panels strengthened by an ultra-high performance concrete (UHPC) layer. Based on the prior experimental programme, simulations of four-point bending tests of the original and strengthened panels are created using software ATENA Science. Calibration of the material parameters is based on the destructive and non-destructive investigations performed in the experimental programme. The comparison of the experimental and numerical model loading curves indicates that additional mechanical testing will be needed in order to achieve an accurate numerical simulation. Although the bending test of the RC panel and the crack mouth opening displacement (CMOD) test performed on the applied UHPC beam were calibrated, the final model of the UHPC strengthened panel does not correspond completely with the experimental measurements. In this paper the possible reasons for this result are discussed. Keywords: Analysis, high-performance concrete, reconstruction. 1. State of the art In recent decades, environmental aspects have been attracting more and more attention in the field of building structures. One of the possible approaches to an environmentally friendlier structure is prolon- gation of its service life. By improving the no-longer complying parts of structures or further utilization of removed building components, the negative impacts of construction can be reduced greatly. However, an increased demand for the load-bearing capacity or a different loading type of the recycled component very often results in necessary strength- ening of the structure. In the case of concrete struc- tures, an application of a new additional concrete layer, usually from a type of high-performance fibre concrete, has been proposed and investigated as a po- tential strengthening method [1–13]. In the previous studies, the layer from high- performance (HPC) or ultra-high performance (UHPC) concrete has been generally applied in four strengthening configurations in thickness from 20 to 100ămm (most frequently 30 mm). A strengthening layer located in the compression zone was most frequently used in the case of panels [2, 5, 8, 12]. The main advantage of this configu- ration lies in its easier performing as the strength- ening layer is applied on the upper surface, thus no complex formwork is needed. Concrete beams have been, in the conducted experiments, strengthened in the compressive zone [1, 4, 10, 11], the tension zone [3, 7, 11], two-sided (left and right surfaces) [7, 13], and three-sided (aăcombination of all of the above) [3, 7, 9, 11, 13]. In order to determine the most effective strength- ening configuration, a number of studies compared the impact of the layer location on the load-bearing capacity of the element [7, 11]. The results indicated that the three-sided configuration reaches the highest values (an increase of up to 89% in the loadbearing capacity) following by the two-sided (around 47%) and the tensile configuration with as low as 16%. Not surprisingly, the conducted experiments also showed that the magnitude of the tensile strength of the strengthening layer has a decisive influence especially in the three-sided configuration, whereas it has only a little effect in the case of compressive zone [7]. Another important aspect of the HPC or UHPC strengthening method is the adhesion of the new layer with the original concrete component. In [7, 13] which have dealt with this topic, a transverse tensile strength test was performed on cylindrical or cubic specimens which were prepared from both the orig- inal basic concrete and UHPC. The results of these experiments were also further used in numerical fi- nite element (FE) modelling in order to describe the behaviour of the contact layer. To further improve the adhesion of the layers, re- searchers have been also investigating different sur- face treatments, such as certain adhesion agents 52 https://doi.org/10.14311/APP.2022.33.0052 https://creativecommons.org/licenses/by/4.0/ https://www.cvut.cz/en vol. 33/2022 Application of High Performance Concrete for Reconstruction Figure 1. A typical geometry of the non-strengthened panels based on destructive investigations. Figure 2. Typical geometry of the strengthened panels after the UHPC layer application [14]. (e.g. epoxy resin) or roughening (e.g. sandblast- ing, hydrodemolition, or wire-brushing). A study [7] showed, that specimens with an epoxy resin surface treatment reach higher values in the transverse tensile strength test but lower values in the shear strength test compared to specimens with a surface treated with sandblasting. Importantly, the further bend- ing test showed that the sandblasted surface leads to higher load-bearing capacity compared to the epoxy treatment. The finding indicates that the strength of the new-to-old concrete bonding depends primarily on its shear strength. This paper focuses on numerical modelling of the UHPC strengthening method. As high-performance concrete is rather economically and ecologically de- manding material, careful optimization of its place- ment and characteristics is crucial for the sake of maintaining the positive effects of the application of the method. In order to achieve an effectively strengthened concrete structure by an UHPC layer, it is necessary to describe characteristics of the materi- als and their interaction as accurately as possible. For this reason, a large number of studies have been con- ducting numerical non-linear analysis alongside me- chanical testing [7, 11–13]. By using an appropriate finite element software (e.g., ATENA, ABAQUS) and material models, the studies have been generally able to successfully describe the structure behaviour with only little (no more than 13.4%) deviations from the mechanical test results. However, several issues in the topic of FE mod- elling needs further clarification as they might influ- ence the model accuracy. Firstly, the compressive and tensile stress-strain relationship for the UHPC is a rather complex problem, as it is significantly dif- ferent from the normal strength concrete. Further, the characteristics of the bonding between the orig- inal substrate and new UHPC layer may distort the results, although the majority of previous studies con- sidered perfect bonding. In our paper, we focus on FE numerical analysis of four-point bending tests performed on prefabricated reinforced concrete panels strengthened by a UHPC layer. In this study, several calibration procedures of material parameters are presented, and the outcomes are compared with experimental measurements. Fur- ther, the specifics of designing recycled structures are discussed and further development is outlined. 2. Experimental program 2.1. Specimens Bending tests were performed on several prefabri- cated reinforced concrete (RC) panels, which have been located in outside conditions for several years. The panels were around 120 mm high, 490 mm wide, 53 I. Broukalová, A. Kohoutková, A. Horáková, H. Schreiberová Acta Polytechnica CTU Proceedings and 2750 mm long; however, the dimensions varied slightly in every specimen. Destructive tests showed that the panels were reinforced with steel reinforce- ment - two !12 mm bars at the bottom surface and two !12 mm bars at the upper surface. However, the performed destructive investigations revealed that the location of the bars was slightly different in every specimen. Further, due to the inadequate reinforce- ment at the bottom surface, four !10mm B500B bars were glued into formed grooves to increase the load- bearing capacity of the panels. Before casting of the UHPC layer, the upper sur- faces of the panels were treated by hand hydro-jetting to ensure adequate bonding of the original substrate with the new layer. After the treatment, the UHPC layers were cast in the compressive zone in thickness of 30 mm. This study focuses on two experimental series - reference panels with no strengthening layer (Figure 1) and panels with 30 mm UHPC layer (Fig- ure 2). 2.2. Material properties In order to determine the strength class of the original RC panels, a number of destructive and nondestruc- tive tests were performed. According to the measure- ments, the material was categorized as C30/37. The material characteristics of the prepared ultra- high strength fibre concrete (UHPC) were investi- gated by destructive mechanical testing. The com- pressive test conducted on cubic specimens deter- mined its mean compressive strength as 133 MPa. Further, a three-point bending test, with support span 500 mm, was conducted on pre-notched beam specimens (150 × 150 × 700 mm) to obtain curves of the dependence of the loading force on the crack mouth opening displacement (CMOD). 2.3. Testing methods To determine behaviour of the investigated strength- ened and non-strengthened panels, a series of four- point bending tests were conducted. The span of the supporting pins was 2500 mm, the span of the loading pins was 800 mm. The loading force and deformation were monitored and recorded during the loading. 3. Numerical modelling of the four-point bending tests All of the analysis was conducted using a finite el- ement analysis software ATENA 5.6.1 Science with GiD interface. The actual geometry of the test setup was modelled in all cases. In the case of the fourpoint bending tests, the symmetry of the setup was consid- ered, and only halves of the specimens were modelled. In all cases, the loading was introduced into the model as a pre-described deformation through steel distribu- tion plates. 3.1. Numerical modelling of the reference non-strengthened panels Firstly, the four-point bending test of the non- strengthened panels was modelled. As the informa- tion about its materials was limited, a careful calibra- tion of the numerical model was crucial for further investigation of the UHPC strengthening method. In order to determine whether it would be possible to omit the spatial effect, the RC panel was modelled as both a plane stress (2D) and a three-dimensional stress (3D) structure model type. The geometry was identical in both cases. However, due to the signifi- cantly lower computational demands of the 2D model, the mesh size could be set to 15 × 15 mm in the case of the 2D model, whereas the 3D model used 25 × 25 mm. In order to obtain the loading curves, the loading force was monitored using a monitor placed under the force and the deflection was monitored on the lower surface in the middle of the span. 3.1.1. Material models As mentioned above, the information about the RC original materials were limited. As a starting point, the material parameters of both concrete and steel bars were generated using the predefined materials in ATENA GiD software. Material Concrete EC2 which uses the CC3DNonLinCementitous2 material model was selected as a representation of concrete. As this material generates the parameters according to the Eurocode2, the category C30/37 was set. The steel bars, both the original and the additional, were defined as Reinforcement EC2 using CCRein- forcement set to the category B500B. 3.1.2. Parametric study and calibration of the FE models Firstly, analysis of the 2D and 3D model with gener- ated default material parameters, using the material models mentioned above, was run. Subsequently, the loading-deflection curves obtained from the analysis were compared in order to evaluate the differences between the 2D and 3D modelling. Figure 3. A comparison between the 2D and 3D loading curve - a reference panel with default material parameters. As can be seen in Figure 3, generally, the curves did not differ significantly. The crack formation and the 54 vol. 33/2022 Application of High Performance Concrete for Reconstruction Parameter Units Default settin Multiplication factor Concrete material model Young modulus MPa 32000 0.56 Tension strength MPa 2.9 1.00 Compressive strength MPa −38 0.79 Fracture energy MN/m 0.000073 1.14 Tension stiffening − no yes Critical compressive displacement m −0.0005 1.10 Original steel reinforcement model Characteristic yield strength MPa 500 0.84 Class − B C Table 1. Quality and Operations Assessment Schedules [7]. Figure 4. The experimental loading curves of the tested non-strengthened panels compared with the calibrated ATENA simulation. reinforcing bars yield occurred at almost the identical loading force (at approx. 15 kN and 63ăkN, respec- tively). However, before the yield point, the lower slope of the 2D curve indicated a slightly lower stiff- ness in the case of 2D analysis, especially after the crack formation. On the other hand, after the yield- ing, the slope of the 2D curve was higher compared to the 3D curve. Furthermore, the failure of the 2D model took place at significantly higher values of the deflection, resulting in a higher maximum load force. On the base of a parametric study (the analysis was run several times, while changing the individual values of the default material models) the numerical models were calibrated (Figure 4) by setting the ma- terial parameter values to required values, so that the loading curve obtained from the FE analysis would correspond as closely as possible with the experimen- tal measurements. In Table 1, the key default param- eters together with the necessary multiplying coeffi- cient for the changed values are listed. 3.2. Numerical modelling of the panels with a UHPC layer After the calibration of the model of the non- strengthened panel, analysis of the strengthened specimens could take place. Conveniently, unlike the original materials, the newly introduced UHPC could have been subjected to several destructive tests. Thus, the information about its compressive strength Figure 5. The experimental force-CMOD curve compared with the FE analysis results (with default material parameters and after the calibration proce- dure). and tensile behavior was much more certain. To correctly determine its tensile material param- eters, a three-point bending test was modelled in the ATENA software and calibrated using the conducted experiment. Subsequently, the strengthened panel was modelled using the parameters obtained from the numerical representation of the bending test. 3.2.1. The CMOD analysis of the applied UHPC In order to define the UHPC material parameters, aăUHPC beam subjected to the three-point bending test was modelled using ATENA Science in 2D. In order to obtain the force-CMOD curves, the load- ing force and horizontal displacements in the notch mouth were monitored. As a representation of the UHPC, CC3DNonLinCementitous2User was utilized as it allows the user to define its own fracture-plastic material model, and it is recommended for the fibre reinforced concrete (FRC) modelling. The cali- bration procedure was performed according to the software documentation. The tensile function, which have a crucial importance for the FRC materials, were carefully set so that the force-CMOD curves corresponded and closely as possible (Figure 5). Calibrated values of the material parameters can be seen in Table 2. 55 I. Broukalová, A. Kohoutková, A. Horáková, H. Schreiberová Acta Polytechnica CTU Proceedings Parameter Units Calibrated values Parameter Units Calibrated values Young modulus MPa 43000 Tension strength MPa 6.5 Compressive strength MPa −133 Table 2. Calibrated values of the key material pa- rameters based on the CMOD analysis. 3.2.2. Modelling of the four-point bending test of the strengthened panels Based on the four-point bending test of the strength- ened panels, the strengthened panels were modelled using ATENA Science as a plane stress structure model type. The geometry of the model remained identical, only the 30 mm UHPC layer was added. The connection between the old and new layer was considered perfectly solid, thus no mechanical char- acteristics were prescribed. The material parameters of the steel bars and orig- inal concrete were set according to the calibration of the non-strengthened panels. The UHPC parameters were taken over from the calibrated CMOD analy- sis. Results of the analysis run with the parameters based on the prior calibration procedures can be seen in Figure 6. Figure 6. The experimental loading curve of strengthened panels compared with the ATENA sim- ulation with material parameters based on the prior calibration procedures. As can be seen, the load-deflection curve from ob- tained from the FE analysis did not match the ex- perimental measurements completely. Similarly, the formation of the tensile cracks occurred around the same loading force and deflection. However, the stiff- ness of the cross-section after the cracking differed quite significantly, as the curve slope of the FE model is notably steeper. The FE model also showed ap- prox. 12 % higher maximum loading force than the experimental measurements. 4. Discussion The finite element analysis of strengthened prefab concrete panels presented in this paper pointed out several crucial factors which influence the design of the UHPC strengthening method. As previously stated in the introduction, the environmental benefits of recycling or renewal of no-longer necessary or un- satisfactory concrete structures are unquestionable. However, as is apparent from the experimental part of the study, the knowledge of the original materials and its geometrical arrangement in the structure is frequently incomplete. For this reason, the numerical analysis inputs are then based on several destructive or non-destructive tests, which provide only limited information about selected specimens. As the parametric study of the non-strengthened panels conducted in our paper showed, the structure behaviour was fundamentally affected by the place- ment and characteristics of the original steel rein- forcement. In our experimental program, the place- ment, and possibly the diameter, of the bars were slightly different in every specimen, and it was not possible to subject the bars to any mechanical mea- surements. Thus, although the FE model of the four- point bending test corresponded well with the experi- ment after the calibration, it was adjusted only to the chosen specimens. Therefore, the model could have been unsatisfactory for the further strengthened pan- els due to a different geometrical arrangement. More known inputs were available for the mod- elling of the strengthened panels. The CMOD anal- ysis of UHPC provided information about the mate- rial tensile properties and the compressive test deter- mined its compressive strength. However, no load- deflection curve was known from the pressure test, only the maximum loading force. Thus, the compres- sive function (the stress-strain relationship) could not have been determined based on any measurements. The loading curve of the strengthened panel showed, that the calibrated model of the original panel and UHPC with defined tensile function based on the CMOD analysis did not match with the experimental measurements. This unsatisfactory result could have been caused by several factors. Firstly, as stated above, the model fitted on the se- lected non-strengthened panels could have been un- suitable for the selected strengthened panels due to the different geometry or materials. Secondly, as mentioned in the introduction, the tensile properties of the strengthening layer in the compressive zone have only a limited impact on the load-bearing ca- pacity. Thus, the setting of the compressive function based on the load-deflection curve could have been crucial and more experimental measurement is there- fore needed. Further, no stresses due to shrinkage of the UHPC layer and mechanical properties of the connection between the original substrate and new material were considered. Although no visible deteri- oration of the layers was apparent during the bending test, the effect may not be negligible in the FE analy- sis and determination of the interface parameters will be necessary in future investigations. 56 vol. 33/2022 Application of High Performance Concrete for Reconstruction 5. Conclusions Application of high-performance cementitious com- posites for reconstruction of concrete structures is aăstep towards sustainable construction. The reuse of damaged or no-longer satisfactory structural ele- ments leads to saving both material and waste. How- ever, an optimized design of the required strengthen- ing is the key to ensure its advantageousness. When the non-linear behaviour of the materials is taken into account, it allows the design to be less conservative, thus more economic. In this paper, the performed non-linear numerical analysis pointed out several crucial factors which in- fluence the numerical simulation suitability. For a sufficient numerical model, the geometry of the recy- cled structures must be known as precisely as possi- ble. Further, when modelling only a general geometry of the structures, the slight differences in the individ- ual elements might affect the model accuracy greatly. However, most importantly, it is crucial to ensure in- puts for the numerical analysis by conducting a suf- ficient number of mechanical tests of the materials present in the modelled structure. Acknowledgements FV20472 Application of high performance cementi- tious composites at rehabilitation of concrete structures, SGS19/149/OHK1/3T/11 Durability of concrete struc- ture and assessment of its life cycle. References [1] M. Genedy, J. Kim, M. R. Taha. Innovative Strengthening Of Rc Beams Using CFRPUHPC Composite. Proceedings of the 7th International Conference on FRP Composites in Civil Engineering, Vancouver, 2014. [2] T. Boeters, R. Braam, H. Kolstein, et al. Concrete overlay of movable steel orthotropic bridge decks. Steel Construction 2(2):104-8, 2009. https://doi.org/10.1002/stco.200910014. [3] F. A. Farhat, D. Nicolaides, A. Kanellopoulos, et al. Behavior of RC Beams Retrofitted with CARDIFRC after Thermal Cycling. Journal of Materials in Civil Engineering 22(1):21-8, 2010. https: //doi.org/10.1061/(asce)mt.1943-5533.0000008. [4] M. Genedy. A New CFRP-UHPC System for Strengthening Reinforced Concrete T-Beams, Master Thesis, 2014. https://digitalrepository.unm.edu/cgi/viewcont ent.cgi?article=1094&context=ce_etds. [5] M. Bastien-Masse, E. Brühwiler. Ultra high performance fiber reinforced concrete for strengthening and protecting bridge deck slabs. Bridge Maintenance, Safety, Management and Life Extension, p. 2176-82, 2014. https://doi.org/10.1201/b17063-333. [6] H. Yin, W. Teo, K. Shirai. Experimental investigation on the behaviour of reinforced concrete slabs strengthened with ultra-high performance concrete. Construction and Building Materials 155:463-74, 2017. https: //doi.org/10.1016/j.conbuildmat.2017.08.077. [7] M. A. Al-Osta, M. N. Isa, M. H. Baluch, et al. Flexural behavior of reinforced concrete beams strengthened with ultra-high performance fiber reinforced concrete. Construction and Building Materials 134:279-96, 2017. https: //doi.org/10.1016/j.conbuildmat.2016.12.094. [8] A. P. Garner. Strengthening reinforced concrete slabs using a combination of CFRP and UHPC, Master Thesis, 2011. [9] M. S. Senin, S. Shahidan, M. Z. Md Maarof, et al. Strengthening and repair of RC beams with sugarcane bagasse fiber reinforced cement mortar. IOP Conference Series: Materials Science and Engineering 271, 2017. https://doi.org/10.1088/1757-899x/271/1/012064. [10] Experimental Investigation on Reinforced Ultra-High-Performance Fiber-Reinforced Concrete Composite Beams Subjected to Combined Bending and Shear. ACI Structural Journal 110(2), 2013. https://doi.org/10.14359/51684405. [11] A. P. Lampropoulos, S. A. Paschalis, O. T. Tsioulou, et al. Strengthening of reinforced concrete beams using ultra high performance fibre reinforced concrete (UHPFRC). Engineering Structures 106:370-84, 2016. https: //doi.org/10.1016/j.engstruct.2015.10.042. [12] J. Moon, M. M. Reda Taha, J. J. Kim. Flexural Strengthening of RC Slabs Using a Hybrid FRP-UHPC System Including Shear Connector. Advances in Materials Science and Engineering 2017:1-7, 2017. https://doi.org/10.1155/2017/4387545. [13] A. A. Bahraq, M. A. Al-Osta, S. Ahmad, et al. Experimental and Numerical Investigation of Shear Behavior of RC Beams Strengthened by Ultra-High Performance Concrete. International Journal of Concrete Structures and Materials 13(1), 2019. https://doi.org/10.1186/s40069-018-0330-z. [14] A. Čítek. Experimentální ověření betonových prvků zesílených pomocí UHPC, ČVUT v Praze, Fakulta stavební, 2019. 57 https://doi.org/10.1002/stco.200910014 https://doi.org/10.1061/(asce)mt.1943-5533.0000008 https://digitalrepository.unm.edu/cgi/viewcontent.cgi?article=1094&context=ce_etds https://doi.org/10.1201/b17063-333 https://doi.org/10.1016/j.conbuildmat.2017.08.077 https://doi.org/10.1016/j.conbuildmat.2016.12.094 https://doi.org/10.1088/1757-899x/271/1/012064 https://doi.org/10.14359/51684405 https://doi.org/10.1016/j.engstruct.2015.10.042 https://doi.org/10.1155/2017/4387545 https://doi.org/10.1186/s40069-018-0330-z