Experimental and theoretical investigation of the structural behavior of reinforced glulam wooden members by NSM steel bars and shear reinforcement CFRP sheet

1 Introduction

Timber is an attractive and appealing building material. Some of the economic advantages of wood are its availability, cost, and ease of fabrication. Today, structural timber elements can be constructed with a wide range of sizes and shapes. However, unlike other engineering materials, timbers are of organic origin and anisotropic; their strength and mechanical characteristics vary considerably. Thus, extensive experimental research is required to achieve the demand for timber building [1,2,3].

Concrete and steel are artificial materials developed for specific activities; thus, they are less likely to contain defects similar to those caused by a natural result in timber. Glulam, or glued layered timber, is a form of engineering wood used in construction for well over a decade. Engineering wood can be constructed from wood panels, typically 19–50 mm thickness and 1.5–5 m length. The panels are indexed by strength and bonded together with high-strength adhesive by pressure, with wood grains aligned to the span length. Fracture typically occurs at the weak point, frequently where knots or cracks exist. Natural imperfections in the timber are dispersed more consistently throughout the section of glulam components, resulting in higher glulam element strength. As a result, glulam products are stiffer and more rigid than typical lumber products, as shown in Figure 1. As can be observed, glulam has a greater average strength (f _m); in addition, the characteristic strength (f _k) increases and the variance decreases [4,5].

Figure 1

Strength of standard timber vs glulam [2].

Although glulam is widely used and internationally known, it still requires strengthening owing to the natural weakness of its main wood component. External and internal bonding techniques are primarily utilized in timber reinforcements. Steel and FRP sheets are used in the tension area of a structural element when externally bonded reinforcements (EBR) are achieved. Bonded fractures within the concrete and externally reinforced steel or FRP elements conduct structural strengthening with the EBR technique when the plates or reinforcement bars are exposed to damage by vehicular traffic or fire [6,7,8]. For internal bonding systems, reinforcing bars are often fastened into grooves or slots of timber, which refers to near-surface mounted (NSM). Bond and load capacity are improved with NSM over EBR due to anchoring the bars into surrounding host members and a better bond connection between the host core elements and the adherence; also, the host wood protects the reinforcement materials [9,10,11].

Recently, the NSM method for reinforcing timber construction has been widely used, and studies have confirmed its application [12]. Based on investigations into the use of NSM strengthening, it involves enhancements in stiffness and bending strength of timber members [13,14,15]. Studies of long-term bending strength have also demonstrated the NSM technique’s effectiveness [16]. Some of these previous studies, with their contribution and main outcomes, are illustrated in detail subsequently.

Gentile et al. studied wooden solid beams reinforced with GFRP bars using the NSM technique. They indicated that the enhancement of the flexural beam strength in a range of 18–46% with the strengthening ratios were 0.27–0.82% in comparison with unreinforced members [14].

Johnsson et al. investigated glulam timber members reinforced by carbon fiber-reinforced polymer (CFRP) rods between the glulam lamellas. They indicated that NSM reinforcing technique enhanced the ultimate moment capacity by about 49–63% [5].

Raftery and Harte developed a finite element modeling of the nonlinear response of glulam members reinforced by GFRP bars, and this model was accurate in comparison with previous experimental results [15].

Raftery and Kelly explored the bending behavior of glulam timber members stiffened and repaired by basalt FRP bars in tension face only. They used the NSM technique with two arrangement locations of reinforcement; one located in a groove at the bottom face of the member and the other placed into a groove at the sides of the member. The average enhancement of ultimate capacity was 23%, and the structural stiffness was improved in the range of 8.4–10.3% [16].

Yang et al. explored the flexural behavior of glulam members reinforced using steel and FRP of bars and plates. They evaluated a significant improvement in flexural capacity and structural stiffness up to 56.3 and 27.5% on average, respectively. In addition, they introduced a theoretical model based on experimental results with acceptable accuracy [17].

Bakalarz et al. examined the bending characteristics of reinforced glued lumber beams by CFRP strips fixed into grooves by epoxy resin at the bottom face. They indicated that the reinforcements enhanced the flexural capacity by up to 32% on average in comparison with the control beam. Also, the modulus of elasticity increased by about 11% in bending and stiffness coefficients [18,19].

Yeboah and Gkantou used the NSM technology with reinforcements of glass and basalt FRP rods to increase the bending performance of glulam timber members. They used two reinforcement configurations: one reinforcement was located at the bottom beam face and the other was on both faces. They observed that the main failure in both reinforced and unreinforced members was the brittle tension mechanism. Furthermore, the NSM reinforcement increased the flexural stiffness (22–33%), while (33–69%) increasing the ultimate strength of the glulam timber members [4].

One of the significant challenges in the construction of timber buildings is determining the ultimate capacity. The current flexural strength calculation method is derived from elastic beam theory, which is insufficient in several aspects concerning the experimentation results of glued timber members, which are indicated as a high safety factor in the elastic method. It is essential to adopt a basic understanding theory for the structural performance of glulam members with the increase in building material cost, which is leading to the more cost-effective use of wood.

Therefore, the basic objective of the present study was to establish a proposed theoretical model of structural analysis and evaluate the efficiency of using the NSM procedure as a strengthening method to timber structures for the enhancement of the overall mechanical performance. Also, an experimental program was achieved for glulam cases under consideration utilizing various cross-sectional reinforcing schemes and ratios.

2 Experiment works

In the experimental part of this study, 11 specimens were examined in 2 groups to compare them with the control beam. Five reinforced glulam (RG) beams of the first group were reinforced with different schemes at tension and compression zones using the same bar size. While the other five specimens of the second group were reinforced with shear reinforcement using fully wrapping strips made of CFRP sheet in addition to the same flexural reinforcement schemes as the first group.

2.2 Specimen fabrication

Each of the 11 glulam timber members tested in this study was fabricated as shown in Figure 2. The steps of the fabrication process can be described as follows:

1. Rendering timber panels from factory production size to dimensions 85 mm × 44 mm × 1,500 mm.

2. Glued each of four timber panels to produce one glulam beam by adhesive under pressure for 72 h.

3. Rendering glulam beams to dimensions 85 mm × 175 mm × 1,500 mm.

4. Cutting the longitudinal groove with sectional dimensions of 15 mm × 15 mm at the corresponding location for the specimen.

5. Cleaning longitudinal grooves and reinforcement bars with high air pressure before placing the steel bar in the groove after epoxy resin insertion; in the specimen of reinforcement in tension and compression zone, the bars of tension face were inserted, and after 24 h bars of compression face were inserted.

6. Samples, the bars were first placed on the tensile zone, and 24 h later, the bar on the compression zone was inserted.

7. Curing time of the RG beam was 7 days prior testing according to the data sheet of the supplier of the epoxy resin of Sika-dur^®-330.

8. Wrapping CFRP sheets around glulam beams with epoxy resin if needed concerning specimen cases.

Figure 2

Fabrication processes of glulam member.

2.3 Specimen configurations

A total of 11 glued-laminated beam samples (Figures 3 and 4, and Table 4), one unreinforced (control), group 1 consisted of five RG beams with steel bars, and group 2 consisted of five beams reinforced by steel bars with wrapping by CFRP sheet. All beams were tested according to BS EN 408:2010 with a bending configuration test (i.e., four points load) under static loading. Each sample was 1.5 m long and rectangular cross-section of 85 mm × 175 mm. The NSM technique was used to reinforce glulam beams at corresponding zones (i.e., tension and/or compression zone; Figure 4). Each glulam beam has a span of 1.35 m, which was divided equally to three-thirds of 450 mm in a typical four-point loading system. Therefore, the ratio of shear span to total depth was equal to 2.5. The available wood in the local construction market governed the selection of this dimension. The top steel bars were used in some cases of the tested timber beams to reinforce the timber beam against compressive stresses that develop in the upper side of the beam due to the bending moment because there is no large difference between the compressive and tensile strengths of the timber. In addition, Figure 5 shows a wrapping CFRP sheet reinforcement for the beams of group 2 only. The wrapping reinforcement was 40 × 600 CFRP sheet @150 mm, which was constant for all beams and distributed at the shear force region only along the beam.

Figure 3

Specimen details.

Figure 4

Specimens configuration.

Table 4

Specimen identification

Group	Symbol	Rein. bar location		Reinforcement ratio (%)		Wrapping CFRP sheet
		Tension zone	Compression zone	Tension	Compression
Control	B1 (con.)	—	—	—	—	—
Group 1	B2 (T1)	1	—	0.528	—	—
	B3 (T1 + C1)	1	1	0.528	0.528	—
	B4 (T2)	2	—	1.055	—	—
	B5 (T2 + C1)	2	1	1.055	0.528	—
	B6 (T2 + C2)	2	2	1.055	1.055	—
Group 2	B7 (RT1)	1	—	0.528	—	40 × 600@150
	B8 (RT1 + C1)	1	1	0.528	0.528	40 × 600@150
	B9 (RT2)	2	—	1.055	—	40 × 600@150
	B10 (RT2 + C1)	2	1	1.055	0.528	40 × 600@150
	B11 (RT2 + C2)	2	2	1.055	1.055	40 × 600@150

Figure 5

Wrapping reinforcement for beams of group 2.

In addition, Figure 5 shows a wrapping CFRP sheet reinforcement for the beams of group 2 only. The wrapping reinforcement was 40 × 600 CFRP sheet @150 mm, which was constant for all beams and distributed at the shear force region only along the beam.

The longitudinal reinforcement ratio of beams in groups 1 and 2 was regularly increased using the same steel bar diameter (i.e., Ø10 mm) to investigate the effect on flexural behavior of glulam timer beams at different reinforcement ratios at tension and compression zones. Also, wrapping reinforcement was constant and used for beams of group 2 only with the same longitudinal reinforcement of beam in group 1 to evaluate the effect of shear strengthening on glulam beam behavior.

2.4 Bending test and instrumentation

Glulam timber beams were imposed for testing in the basic laboratory of the College of Engineering at the University of Kufa. At a loading rate of 2.5 kN/min until failure, all specimen members were tested using a compression hydraulic machine of 2,000 kN. Also, at the bottom face of the beam center, the digital dial gauge (100 mm) was installed as shown in Figure 6.

Figure 6

Glulam timber beam setup for the test.

3 Results and discussion

Experimental results of glulam beam specimens were presented in terms of ultimate load capacity, failure mode, and deflected loaded beam response. Table 5 presents the results of all glulam beams. As a result, the strengthening NSM technique adopted in this study improved the structural performance in many aspects.

3.1 Load–deflection curves

From the bending test, Figure 7 displays the curves of load vs the maximum central point deflection of glulam beams of group 1 in comparison with the reference beam (i.e., unreinforced glulam beam) behavior. In addition, curves of group 2 specimens and the controlled beam are presented in Figure 8. It seems, the using of reinforcing bars at the tensile region of glulam beams enhanced the flexural capacity and decreased deflection values when compared with reference beam behavior as well as the ultimate load capacity were increased with considerable value in a range of (16–49)% due to the improvement in flexural stiffness which is a factor of moment capacity as well as the amount of reinforcement ratio. On the other hand, the wrapping reinforcement of the CFRP sheet of beams in group 2 did not provide a significant increment, and the curves of beams with and without reinforcement wrapping were relatively close to each other, and the percentage of ultimate load that increases between any pair of beams in the first and second groups (i.e., B2-T1 and B7R-T1 one five set of the coupled beam) not more than 7% which is less than what researchers expected due to inefficient wrapping reinforcement using strips of CFRP sheet which is sustained high strain to fail. Furthermore, Figure 9 demonstrates the load vs deflection response of reference members with coupled beams from groups 1 and 2.

Figure 7

Load–deflection response of group 1 beams.

Figure 8

Load–deflection response of group 2 beams.

Figure 9

Load–deflection response of coupled beams of groups 1 and 2. (a) Coupled beams B2-T1 and B7R-T1. (b) Coupled beams B3-T1 and B8R-T1. (c) Coupled beams B4-T1 and B9R-T1. (d) Coupled beams B5-T1 and B10R-T1. (e) Coupled beams B6-T1 and B11R-T1.

Introducing fully wrapped strips of CFRP sheet 40 mm × 600 mm at 150 mm c/c as shear reinforcement for RG members gave a small significant improvement in terms of flexural performance, and the range of increasing ultimate load capacity was 2–7% in comparison with the same reinforced specimen but without CFRP wrapped strips.

3.2 Failure modes

Failure mode description is one of the important results of experimental works; many factors affected various failure modes based on load arrangements, material characteristics, section properties, and knots, and other inherent timber imperfections were integrated to identify the failure mode. Mainly, three models of failure were recognized in this study, which are described as follows:

1. Tension failure was presented in an unreinforced glulam member (i.e., reference beam) in which the tensile stress at extreme fiber in the tension zone exceeded the tensile strength limit of the timber, then vertical cracks were introduced and brittle tension failure occurred as shown in Figure 10a.

2. Shear flow failure was admitted in the RG members in which the tensile and compressive stresses were withstood with the reinforcement, at which the failing shifted to shear flow failure mode as it exceeds the shear limit strength of the glulam timber. This failure was sudden and initiated by cracks propagated along fibril direction of timber as shown in Figure 10b.

3. Compression failure was presented in the RG beam in the compression zone, the strengthening of the tension reinforcement shifted the failing toward the compressive area. It is a ductile failure initiated by plastification of the extreme top fiber of the compression zone, then goes down till failure.

Figure 10

Failure modes recognized. (a) Tension failure. (b) Shear flow failure. (c) Compression failure.

The recognizing failure modes were associated with most of the specimens of groups 1 and 2 except the control glulam beam failed in pure tension failure mode. Generally, the RG member is initially loaded in tension and compression zones until the stress reached the strength limits, at this stage of loading if the reinforcement is introduced in the tension zone, the plastic stress in the outermost fiber of the compression zone reached then tends toward neutral axis location, at this stage of loading and if the reinforcement is introduced in a compressive zone the horizontal cracks of shear flow failure mode would be occurred depending on the shear strength limit. Furthermore, the presence of the knot and other timber imperfections could change the failure path as in specimens B6 and B8 in which the tension failure occurred due to the presence of the knot just above the zone of tension reinforcement as shown in Figure 11. In addition, the location of horizontal cracks of shear flow associated with compression or tension failure occurred in the top, middle, or bottom third of the cross-section depending on the amount of reinforcement ratio in tension and compression zones.

Figure 11

Associated failure modes. (a) Compression and shear flow failure B5. (b) Tension knot and shear flow failure B6. (c) Tension knot and shear flow failure B8. (d) Shear flow failure and cutting CRFP B11.

4 Theoretical model

As there is no established method for calculating the bending moment capacity of an RG timber beam, the present study introduced a specific model for the analysis of glued laminated timber beams reinforced by ordinary steel bars using the NSM technique. In this model, the principles of strain compatibility and forces equilibrium were considered, which were adopted by many researchers when dealing with RG beams [4,17].

Basic strain–stress curves and laws of constitutive for timber and steel are used as shown in Figure 12. Also, basic assumptions for pure flexural behavior were adopted, plain section, the perfect bond between bars and sides of timber grooves, and disregard knots and other wood defects.

Figure 12

Constitutive laws of material [4]. (a) Stress–strain relation of timber. (b) Stress–strain relation of steel.

The nonlinear model was designed to evaluate the beam behavior including curvature, stiffness, ultimate load capacity, and deflection of the member until failure. With an incremental strain, the model simulated the overall behavior. At first, a small strain is assumed in the outer compressive edge, then the associated strain/stress state is evaluated across the cross-section, and these values are compared with the corresponding yield limits, considering the conditions of strain compatibility and forces equilibrium with the influence of timber plasticity and reinforcement yield, the load, moment, curvature, and deflection of stage load would be calculated. A small increase in strain is added if the stress/strain state does not cause any parts to fail. This process is repeated until it fails to perform. The theoretical RG beam model implemented two stages of loading depending on whether the compressive zone is plastified or not. The linear elastic stage and plastic stage of cross-section analysis of the RG beam with reinforcement in tension and compressive zones are presented in Figure 13. By introducing an initial small value of strain in the outermost fiber in compressive zone ε_ci, the neutral axis is achieved according to the load stage using strain compatibility and force equilibrium equations (1–19).

1. Within the linear elastic stage:

   strain compatibility includes

   (1) ε gt = h − d e d e ε gc ,

   (2) ε rc = d e − d c d e ε gc ,

   (3) ε rt = h − d e − d t d e ε gc ,

   where ε gt denotes the strain of glulam at tension area, ε gc denotes the strain of glulam at compressive area, ε rc denotes the strain of compressive reinforcement, ε rt denotes the strain of tensile reinforcement, h denotes the total section depth, d e denotes the depth of neutral axis from top face, d c denotes the depth compressive reinforcement, and d t denotes the depth tensile reinforcement.

   Equilibrium forces include

   ∑ F t = ∑ F c ,

   (4) F rt + F gt = F rc + F gc ,

   (5) σ rt A rt + 1 2 σ gt b ( h − d e ) = σ rc A rc + 1 2 σ gc b d e ,

   where F t denotes the resultant of tension forces, F c denotes the resultant of compression forces, F rt denotes the force of tension reinforcement, F rc denotes the force of compression reinforcement, F gt denotes the force of glulam in tension, F gc denotes the force of glulam in compression, σ rt denotes the stress of tensile reinforcement, σ rc denotes the stress of compression reinforcement, A rt denotes the area of tensile reinforcement, A rc denotes the area of compression reinforcement, σ gt denotes the stress of glulam in tension area, σ gc denotes the stress of glulam in compression area, and b denotes the section width.

   The ultimate moment capacity of RG member can be evaluated from equation (6) as follows:

   (6) M u = F gc . 2 3 d e + F rc . ( d e − d c ) + F gt . 2 3 ( h − d e ) + F rt . ( h − d e − d t ) ,

   where M u denotes the ultimate moment.

   Finally, the corresponding deflection of mid-point of RG beam can be calculated from equation (7) as follows:

   (7) Δ u = P u . ( 3 L 2 − 3 a 2 ) 48 EI eff ,

   (8) P u = M u a ,

   (9) EI eff = M u ∅ ,

   (10) ∅ = ε gc d e ,

   where Δ u denotes the central deflection, P u denotes the ultimate load, L denotes the span of the beam, a denotes the distance from support to load point, EI eff denotes the effective flexural beam rigidity, and ∅ denotes curvature.

2. Within the plastic stage:

Figure 13

Cross-section analysis of RG beam. (a) Elastic stage. (b) Plastic stage.

strain compatibility includes

where d p denotes the depth of neutral axis from top face plastification depth.

where c 1 denotes the plastification depth and ε gc_e denotes the strain elastic limit of glulam in tension.

Equilibrium forces include

where F gc 1 denotes the force of glulam in compression of plastic portion, F gc 2 denotes the force of glulam in compression of elastic portion, F rc denotes the force of reinforcement in compressive area, F rt denotes the force of reinforcement in tensile area, F gt denotes the force of glulam in tension zone, and σ g c p denotes the plastic stress limit of glulam in compression area.

Finally, the ultimate moment capacity of RG member can be evaluated from equation (17)

where M u denotes the ultimate moment, E r denotes the modulus of elasticity of reinforcement, E g denotes the modulus of elasticity of glulam, σ rt_y denotes the yield stress of reinforcement in tension, and σ rc_y denotes the yield stress of reinforcement in compression.

The corresponding deflection of the midpoint of RG beam can be calculated from equations ((7) and (8)), while the flexural rigidity of the plastic stage was evaluated from equation (18) as follows:

where ∅ p denotes the curvature for plastic phase.

The tension failure model of the RG beam may have occurred when the tensile stress at the outermost filament in the tensile region reached the ultimate stress limitation of timber in tension σ gt_u and the tensile crack penetrated the quarter of RG beam depth h . Whilst the compression failure mode of the RG beam can occur when the strain of the outermost fiber in the compressive zone reached a value of three times the plastic strain limit of timber in compression where the plastification penetrated in a compressive zone by a plastic depth c 1 .

4.1 Theoretical model verification

Using the experimental properties of the timber and reinforcement steel bar, the verification of the theoretical model was carried out for different specimens with various reinforcement ratios to check the capability of the theoretical model to simulate the case under consideration as shown in Figure 14 and Table 6. A satisfactory agreement has been achieved between experimental and theoretical model results with 4 and 11% values of the largest variation coefficient COV of comparison results for ultimate loading and deflection, respectively.

Figure 14

Experimental and theoretical Comparison of load–deflection response of different RG beams.

Table 6

Verification of theoretical and experimental results

Symbol	P u-exp (kN)	P u-the (kN)	P u ‒ exp P u ‒ the	Δ u–exp (mm)	Δ u–the (mm)	∆ u ‒ exp ∆ u ‒ the
B1con.	77	76.8	1.00	14.1	15.0	0.94
B2-T1	89	87.6	1.02	15.5	15.8	0.98
B3-T1-C1	101	98.9	1.02	18.2	15.7	1.16
B4-T2	105	97.7	1.07	17.1	16.5	1.04
B5-T2-C1	110	109.6	1.00	19.9	16.4	1.21
B6-T2-C2	113	120.9	0.93	19.5	16.4	1.19
Mean			1.01			1.09
COV			0.04			0.11

5 Conclusions

Based on the results of the experiments and the theoretical model presented in this study, it is possible to draw the following conclusions:

1. The NSM techniques using ordinary steel bars are effective ways of strengthening the glulam members in terms of flexural stiffness while increasing ultimate load capacity in a range of 16–49%.

2. The RG beams showed more strength as the tensile reinforcing ratio was increased as a result of the timber section tending to compression failure mode.

3. Mainly three parameters affected the flexural behavior of the RG beam, which were the reinforcement ratio, the timber stress ratio of ultimate tension to yield compression ( σ gt − u / σ gc − y ) , the ratio of axial stiffness ( E r A rt / E g A g ) , and the placement of tensile reinforcing to the timber section ( d t / h ) .

4. A brittle tension failure mode dominated the failure mode in the unreinforced glulam member (i.e., reference beam) due to the absence of reinforcement in the tension zone.

5. The reinforcement ratio is the major factor in increasing the flexural strength of the RG member in failure mode. In the case of the horizontal shear, failure mode becomes dominant. There is no improvement in the strength properties of the RG beam, and the influence of the reinforcement ratio has vanished.

6. Though several theoretical predictions of flexural capacity were understated when compared to experimental predictions, these disparities were frequently about 10%, confirming that the proposed theoretical model was accurate, and the mean of P _{u_exp}/P _{u_the} and Δ _{u_exp}/Δ _{u_the} were 1.01 and 1.09, respectively.

7. Fully wrapped strips of CFRP sheet 40 mm × 600 mm at 150 mm c/c as shear reinforcement for RG members slightly improved flexural performance and ultimate load capacity by 2–7%.