Behaviour of glass and sisal incorporated gypsum based composite panels under axial compression – experimental and numerical study

Gypsum based composite ( GBC ) Wall panels are being used in precast concrete industry as the panels are low cost, energy consumption, good habitability, and good ﬁ re resistance. This paper presents the studies related to gypsum based composite panels incorporated with glass and sisal ﬁ bers. The main objective of the research is to investigate the capacity of gypsum based composite wall panel subjected to axial compression with and without eccentricity. The panels were loaded in axial compression with and without eccentricity. Gypsum, white cement, ﬁ bres ( glass ﬁ bers or sisal ﬁ bers ) are used to prepare composite panels of size 1000 ( length ) × 500 ( width ) × 125 ( thickness ) mm. Total four panels ( two with glass ﬁ bers and two with sisal ﬁ bers ) were tested under axial compression ( two panels without eccentricity and two panels with eccentricity ) were tested under load control. From the tests, it was observed that the eccentric load application on the panels is very signi ﬁ cant compared to normal load application. Signi ﬁ cant decrease in the ultimate load is observed for the case of panels subjected to eccentric loading. Nonlinear ﬁ nite element analysis for the ﬁ bre incorporated ( E-glass and sisal ) gypsum


Introduction
Gypsum-based composites are widely used for variety of applications since ancient times in view of low cost, less energy consumption, light weight, good habitability and superior fire resistance [1,2].Gypsum can also recycled as per hydration and dehydration theory [3][4][5].Gypsum is primarily obtained from natural minerals and the byproducts of various chemical processes.Gypsum by-products include flue-gas desulphurization gypsum (FGDG), phosphogypsum (PG), fluorogypsum, and red gypsum [6].FGDG is generated in the flue gas desulphurization system in coal-fired power plants [7].To conserve and protect the environment, gypsum based products can be used efficiently.Gypsum composites are one of the sustainable building materials.The utilization of gypsum by-products is still less and there is wider scope for variety of applications [8][9][10].The limitation of gypsum is its brittleness, lower mechanical properties, less water resistance, low thermal and sound isolation properties and shrinkage during fire exposure, which limits the range of applications [11][12][13][14].To overcome the brittleness of gypsum based products, fibers are being incorporated in the gypsum composites to make fibre reinforced gypsum based composites.The properties of fibers include superior tensile strength, toughness, durability etc [15].Several studies were reported in the literature that the addition of fibers in quasibrittle matrices, significantly improves the behavior of the specimens/panels/structural components [16][17][18][19][20][21].
Recently, more studies were reported on the use of fibres in gypsum to enhance the performance of gypsum composites [22].In the present study, experimental and numerical simulations were carried out on glass/sisal incorporated gypsum based composite panels subjected to axial compression.The details are presented in the subsequent sections.The applications of composite materials and their processing techniques and optimization for machining of composite materials are investigated [23][24][25].Gypsum board is one of the most largely used drywall products in construction.It has positive impacts on the environment.It also represents a sustainable product that can be re used over time.Nowadays, plasterboard is an essential product in construction and renovation of buildings.

Materials
2.1.Gypsum and white cement β -Hemihydrate gypsum plaster (β -CaSO41/2H2O) is produced by calcining beneficiated phosphogypsum at 150°-160 °C.The calcination process is done for a period of 4 h in an electric oven with spatulation at regular intervals.On heating at 150 °C, gypsum lost at about 14.7% of water of crystallization in the form of steam, resulting in the generation of β -hemihydrate plaster which contains about 6.0% of the remaining water of crystallization.The plaster is then ground in a ball mill to a fineness passing IS sieve 150 micron.Table 1 presents the chemical composition of phosphogypsum.
White cement is made from raw materials with a low content of colouring elements such as Mn, Cr, Ti and Fe.Gypsum, active and inert mineral additives are added in grounding stage and also exhibit whiteness corresponding to cement of the given class.White cement is produced with standardized surface-active plasticizers or hydrophobic additives (in amounts of up to 0.5 wt%) not impairing the whiteness of the product.Table 2 presents the physical properties of hemihydrate gypsum and white cement.
Mechanical properties such as compressive strength (cube of size 70.7 mm), split tensile strength (cylinder of size 100 x 200 mm) and flexural strength (prism of size 100 ×100 x 500 mm) were evaluated and presented in table 3.

Glass and sisal fibre
Glass fibres are most widely used among all the synthetic fibres as they offer excellent strength and durability, thermal stability, resistance to impact, chemical, friction, and wear properties.Sisal is a natural fibre of Agavaceae (Agave) family that gives a stiff fibre, commonly used for making rope and twine.Sisal fibres hold numerous elongated fibre-cells which are about 6 to 30 μm in diameter.These fibres are characterized by mechanical means.The individual fibre-cells are interconnected through the middle lamella.Table 4 presents the key differences between the natural fibers and glass fibers.
Electronic grade glass fibres are recycled glass fibres which are obtained from the glass of television, computer, laptops, etc (E-glass fibers).E-glass fibres are composed of 40% recycled glasses, 54% of silicon oxide (SiO 2 ), 15% of aluminium oxide (Al 2 O 3 ), and 12% of calcium oxide (CaO).The diameter of the fibre varies from 8 to 10 μm.E-glass fibres has the density of 2.54 g cm −3 .The chemical composition of the sisal fibre contains approximately 54%-66% cellulose, 12%-17%hemicellulose, 7%-14% lignin, 1% pectin and 1%-7% ash.The sisal fibres are used as reinforcement in the developed composite.Table 5 presents the properties of the sisal fibres used in the present study.

Methodology -Wall panel casting and testing
The feedstock comprising of the crushed gypsum and the reinforcing elements are initially dehydrated by calcinations or by the heating process.The suspension of the plaster of paris and water are thoroughly mixed with varied proportions of the reinforcement fibres, viz., e-glass or sisal fibre.Additives are then added to this feedstock and mixed with water to form a slurry and fed continuously to produced long papers and then these layers of papers are then pressed under high pressure to form composite material board.Higher the reinforcement the strength of the composite material board provides high tensile and flexural properties.
The volume fraction of the Glass fibre Gypsum based composite material board are , β -Hemihydrate gypsum −80%, White Cement −18%, Glass Fibre-2%, and volume fraction for Sisal fiber gypsum based composite are β -Hemihydrate gypsum-80%, White Cement-19%, Sisal Fibre −1%.Four wall panels of size 1000 (length) × 500 (width) × 125 (thickness) mm were cast and tested under axial compression with and without eccentricity.The eccentricity is 60 mm.The main objective of the test program is to investigate the capacity of gypsum based composite wall panel subjected to axial compression with and without eccentricity.All the panels were tested under UTM of capacity 500kN in load control mode.During testing, deformations were monitored at selected locations.Crack pattern and failure modes were also recorded during testing.Eccentric load was applied at 60 mm from the centre line of the specimen.The three LVDTs were used to measure the axial displacements.The support condition is pinned support at both ends of a wall, as shown in test setup of figure 1.All tested panels are depicting the typical failure mode of wall panels with two pinned supports crushing and tensile breaking at the mid span.The typical failure mode of wall panel, the breaking points are closer to the pinned support about one-quarter of the height above the pinned support for specimens for the case of eccentric All wall panels are failed due to plaster crushing, irrespective of the load eccentricity.Due to load eccentricity, there is no significant out of plane bending is observed in the panels.The failure of the panel is mainly governed by the geometry, type of fibre, loading position and support conditions.The Euler buckling load for the wall with an effective length of 1 m and without eccentricity is calculated approximately as 260 kN.
The experimental values are 245.5 kN and 174 kN for the case of E-glass incorporated gypsum based composite panel and sisal incorporated gypsum based composite panel respectively.Wall panels are also experienced vertical and inclined cracks during the testing.The vertical cracks are due to axial compression and the inclines are due to the friction forces at the interface between the loading surface and the specimen.The first crack is occurred at about 110 kN on left side under the loading point for the case of E-glass fibre incorporated gypsum based composite wall panel and gradually progressed towards top (loading point).Further, with the increase of load, multiple cracks were developed and final failure occurred due to crushing of plaster at a load of

Finite element analysis
Nonlinear finite element analysis for the fibre incorporated (E-glass and sisal) gypsum based composite panels under axial compression has been carried out by using the general-purpose finite element software, ABAQUS.Specific model of the composite laminate used in finite element analysis software for compression simulation is shown in the figure 2.
The length, width and thickness of the composite panel are 1000 mm, 500 mm and 125 mm respectively.For the finite element analysis, three dimensional hexagonal eight noded reduced integration element with a size of 12.5 mm is used.Mesh size has been arrived at by performing sensitivity analysis.Tables 6 and 7 show the elastic and plastic properties of composite made up of E-glass fibre and sisal fibre.
Loading and boundary conditions have been simulated as close as to experimental situations.Nonlinear static analysis has been carried out for the specimens made up of E-glass and sisal fibers (G0, S0, G60 and S60).G0 indicates, the specimen with E-glass fibers is loaded in axial compression without eccentricity.G60 indicates, the specimen is loaded in axial compression with 60 mm eccentricity.Similarly, S0 and S60 are corresponding to sisal fibers.From the analysis, primarily, the load versus deflection has been captured for all the panels.Further, other responses such as stress and strain contours have been captured.
The stress-strain curve exhibits the linear relationship between stress and strain during compression of the composite material.At the initial stage of compression, composite materials tend to initiate their strain-stress behaviour in a linear way.When the straining of the composite material reaches a maximum level, the linear elastic behaviour stops and the composite material specimen break and the stress strain relationship of the material behaves in a nonlinear elastic way.In the stress strain curve, the FEA and real experiments were very close to each other.Usually a 2% error in the prediction of the values is acceptable and this is due to the fact that true stress will be always greater than the engineering stress.In a real experiment, the stressed locations are to be calculated separately.Whereas FEA takes the whole structure into account, hence the highest stressed locations will always be considered.In addition to stress results, FEA provides many other results such as deformation, reaction forces etc. Minor discrepancies of the percentage of error between real experiments and FEA are acceptable with the percentage of error is within the range of 0.4% to 13%.
Figures 3(a) and (b) interprets the load-displacement plots for G0 and S0 specimens respectively.It can be observed that the predicted behaviour of gypsum based composite panel incorporated with E-glass and sisal fibers is in good agreement with the corresponding experimental observations.On closer examination E-glass composite, it can be noted that both the curves are meeting at a yield load of 125 kN and the corresponding deflection is 0.33 mm.Later again, the curves are intersecting each other at 210 kN and the corresponding deflection is 0.76 mm.The predicted ultimate load is 228 kN and the experimental load is 245.5 kN.Similarly for sisal composite, it can be observed that both the curves are meeting at a yield load of 108 kN and the corresponding deflection is 0.33 mm.Later again, the curves are intersecting each other at about 145 kN and the corresponding deflection is 0.67 mm.The predicted ultimate load is 161 kN and the experimental load is 174 kN.E-glass incorporated gypsum based composite panel is found to be superior in terms of load carrying capacity.Since, the experiments were conducted under load control, only failure load has been compared.
Figures 4(a) and (b) are load-displacement plots for G60 and S60 specimens respectively.It can be observed that the predicted behaviour of gypsum based composite panel incorporated with E-glass and sisal fibers under axial compression load being applied at an eccentricity of 60 mm is in good agreement with the corresponding experimental observations.On closer examination E-glass composite in figure 4(a), it can be noted that both the curves are following the same trend up to 110 kN.The corresponding deflection is 0.4 mm.Later, the complete behaviour of the panel has been captured through finite element analysis.The predicted failure load is 186 kN and the corresponding deflection is 0.96 mm.From figure 4(b) for sisal composite, it can be observed that both  the curves are following the same trend up to 55kN.The corresponding deflection is 0.25 mm.Later, the complete behaviour of the panel has been captured through finite element analysis.The predicted failure load is 133kN and the corresponding deflection is 1 mm.Further, from figures 2 to 5, it can be clearly observed that the loading effect is significant on the behaviour of the specimen.The ultimate load is significantly lesser for the composite panels loaded axially with eccentricity of 60 mm.The bonding between the fibres and the matrix should sufficient to transmit the loads evenly.The strength and stiffness becomes most integral part in the simulation of the composite material under axial compression.Debonding event often causes sliding of the fibres and the matrix material.Figures 5(a)-(d) presents the normal longitudinal strain of G0, S0, G60 and S60 respectively.Strain along the width of the specimens for the case of G0 and S0 found to be similar behaviour.Hexagonal failure pattern is observed for both cases with zero eccentricity, except the change in strain values.Typically, the elliptical strain contour region at the centre of G0 and S0 is −0.0011 due to the contribution of fibre in the composite, since the strain is less than 0.002.Figures 5(c) and (d) are also exhibiting the same nature as observed for zero eccentricity case.In all the cases, the failure of composite specimen with sisal fibre is much earlier than the specimen with E-Glass fibre.
Figures 6 and 7(a)-(d) are exhibiting the strain corresponding to width and thickness of G0, S0, G60 and S60 respectively.The elliptical strain contour region at the centre of G0 and S0 are 0.00036 and 0.00041 respectively in the lateral directions.
Figures 8-10 exhibit the plastic strain along the length, width and thickness directions.All the strain profiles are found to be similar pattern except the failure direction.In all cases, the failure of sisal fibre is earlier than the  E-Glass fibre due to the less stiffness of the fibre.Unsymmetrical failure pattern is observed for 60 mm eccentricity irrespective of the materials.This could be due to the boundary effect.From the FE results, it can be understood that, the compressive and tensile nature could be due to normal longitudinal and lateral strain.
Figure 11 shows the compressive stress of the specimens.The specimens G0 and S0 are seem to be compressed uniformly (about 85% of total area) and the values are 3.4 and 2.4 MPa respectively.The variation   observed in the compressive stress could be due to the material stiffness.Similarly, for the case of G60 and S60, the area undergone is about 35% and the values are 2 and 1.2 MPa corresponding to G60 and S60.The reduction in the strength can be noted in figures 11(c) and (d), due to the asymmetry prone to early failure as stated previously with respect to the eccentricity.except nature of the stress.The deformation profile is shown in figure 14.Figures 14(a) and (b) corresponding to zero eccentricity, which is similar to each other even though material differs, confirms the proper load application.Further, figures 14(c) and (d) corresponding to 60 mm eccentricity exhibiting the similar behaviour.

Summary and concluding remarks
Experimental and numerical studies were carried out on gypsum based composite panels incorporated with E-glass and sisal fibers under axial loading applied with and without eccentricity.Four wall panels of size 1000 (length) × 500 (width) × 125 (thickness) mm were cast and tested under axial compression with and without eccentricity.The eccentricity is 60 mm.All the panels were tested under load control mode.During testing, deformations were monitored at selected locations.All tested panels are depicting the typical failure mode of wall panels with two pinned supports crushing and tensile breaking at the midspan.Due to load eccentricity, there is no significant out of plane bending is observed in the panels.Wall panels are also experienced vertical and inclined cracks during the testing.Numerical simulations were carried out for the gypsum based composite panels incorporated with E-glass and sisal fibers under axial loading by using general purpose finite element software ABAQUS.The axial loading was applied on the panel with and without eccentricity.Material nonlinearity has been considered in the analysis to predict the complete behaviour of the panel.The predicted load versus deformation from finite element analysis is found to be good agreement with the corresponding experimental observations.Since, the experiments were carried out under load control, the predicted deformations were compared only up to peak load.The developed finite element model could able to capture the complete behaviour of the composite panel.

Figure 4 .
Figure 4. (a) and (b) Load versus displacement for E-Glass composite and Sisal composite corresponding to load at 60 mm eccentricity.

Figure 10 .
Figure 10.Plastic strain contour across the thickness.

Figure 12
Figure 12 exhibits the tensile stress of the specimens.It can be observed that the tensile stress of the specimens is lesser than 10% of the compressive strength, resulted in by the significant influence of fibre in the composite.Von-Mises stress distribution shown in figure 13 also looks similar to compressive stress distribution

Table 2 .
Physical properties of the hemihydrate gypsum and white cement.

Table 3 .
Mechanical properties of the hemihydrate Gypsum.

Table 4 .
Comparison between natural and glass fibres.

Table 5 .
Mechanical properties of E-glass and sisal fibers.
245.5 kN.For the case of sisal fibre incorporated gypsum based composite wall panel, the first crack occurred at 75 kN near to underneath of loading point and the ultimate load is 174 kN.Wall panel failed after crushing of mortar.

Table 6 .
Elastic properties of composite.

Table 7 .
Plastic properties of composite.