FACULDADES INTEGRADAS DE CACOAL - UNESC

ENGENHARIA CIVIL

NOME COMPLETO DO(S) AUTOR(ES)

TÍTULO DO TRABALHO (EM LETRAS MAIÚSCULAS) FONTE TIMES

NEW ROMAN – TAMANHO 16: e subtítulo, se houver, em letras
minúsculas

CACOAL
2019
 NOME COMPLETO DO(S) AUTOR(ES)

TÍTULO DO TRABALHO (EM LETRAS MAIÚSCULAS) FONTE TIMES

NEW ROMAN – TAMANHO 16: e subtítulo, se houver, em letras
minúsculas

Artigo apresentado como parte dos requisitos

necessários para aprovação no componente curricular
Trabalho de Conclusão do Curso de Engenharia Civil das
Faculdades Integradas de Cacoal – UNESC.

Orientador: Prof(a). Nome do(a) orientador(a)

CACOAL
2019
 3

Pull Out Strength of Bonded-in Steel Rods

Perpendicularly to the Grain in Corymbia citriodora
and Pinus oocarpa Shiede Timber
Júlio C. Pigozzo1, Felipe N. Arroyo2, Diego H. Almeida3, André L. Christoforo3,*, Eduardo
Chahud4, Francisco A. R. Lahr5

1
Department of Civil Engineering, Technology Center, State University of Maringá (UEM), Maringá, Brazil
2
Faculdades Integradas de Cacoal (UNESC), Cacoal, Brazil
3
Department of Civil Engineering (DECiv), Federal University of São Carlos (UFSCar), São Carlos, Brazil
4
Department of Civil Engineering, Federal University of Minas Gerais (UFMG), Belo Horizonte, Brazil
5
Department of Structural Engineering (SET), São Carlos Engineering School, São Paulo University (EESC/USP), São
Carlos, Brazil

Abstract: The study of pull out strength of glued in rods perpendicularly to the grain is presented. Three types
of resin were used in order to evaluate their efficiency. Additionally, the effects of the variation of the moisture
content and of the glue line thickness were considered. The beams were made of wood from Corymbia
citriodora, (ρ12% = 1000 kg/m3) and Pinus oocarpa Shiede (ρ12% = 550 kg/m3). The wood specimens (four
samples for each type of test) were seasoned to the expected moisture contents of 12, 15, 18 and 22%. The
anchorage length of the glued in rods was 80 mm, the hole diameters were 7.5, 8.5, 9.5 and 10.5 mm and their
respective glue line thickness, 0.6, 1.1, 1.6 and 2.1 mm. The deformed reinforcing bars used as rods, with the
minimum yield strength of 500 MPa, had nominal diameter of 6.3 mm. The obtained results confirmed the
assumption that type of resin, moisture content, glue line thickness are strength determining factors, while
density is less than others. The results are presented in a comparative form. The failure modes are discussed and
the need of a broad understanding of resin behavior on wood was emphasized.
Keywords: Adhesives, glued-in steel rods, structural resin, anchorage strength.
1. Introduction

The use of glued steel bars on structural pieces of wood was initiated by making it clear the need to fix bolts that could
receive axial, lateral or combined loads. Such a connection has received attention and recognition by: excellent
performance when well designed and executed; aesthetic appearance; and low cost [1]. The bars used as connectors are
preferably: threaded; galvanized and high strength; with deformed surfaces; and scored or threaded with high strength.
Anchorage adhesion, initially, is the combination of chemical and mechanical adhesion. From a request level the chemical
adhesion breaks, remaining only mechanical adhesion [2].
The use of bonded steel bars is characterized by the sticking of bars into holes of larger diameters. It depicts an innovative
and improved method of connections, being an important aspect for connections using adhesives. Its main advantages are:
connections allow higher levels of effort transfers; top connections with glued bars can withstand large bending moments;
the holes used in the connections do not weaken the structural members, as in the connections with bolts; possible errors in
construction sites are avoided, such as improper drilling; makes it possible to join large pieces, having larger free spans; the
structural members become more aesthetic, avoiding apparent connectors like toothed plates or screws and bolts; the
connections are easily protected against fire; the connections are potentially cheaper, compared to the finger-joint system,
as it does not require special machines for the execution; and the connections have less material and lower cost of
production, compared to the bolted connections [3].
Bainbridge and Mettem [4] report that there are still no general technical standards governing the use of steel bars bonded
to wood structures, although they have been used for more than twenty years. Performance requirements and project
regulations differ between them. Due to the uncertainties in the behavior of these connectors and the lack of reliable
calculation methods, they have not yet been introduced into the main part of the European Standard. It is currently listed in
EUROCODE 5:1993 Part 1-1 [5] as recommendations for use in its annex.
Structural synthetic resins most commonly used in timber structures are: phenol resorcinol formaldehyde (PRF),
polyurethanes (PUR) and epoxies (EP). Currently PRF does not require a hot cure and has fewer retractions and the EP is at
a competitive cost. There were previously restrictions on the use of EP as a structural resin, suspected of having brittle
ruptures with increasing temperature or against long loads. Currently, this adhesive is the most suitable for anchoring steel
bars in structural timber, because they have better properties and fewer defects over time. A significant change appears
when the EP is no longer considered as a set of resins with uniform behavior, but it is possible to compare different brands
and manufacturers [3].
The variables that influence the anchorage strength of bonded-in steel bars are currently being studied by several
researchers in Europe and New Zealand. The results, conclusions and suggestions presented are related to the
characteristics of the adhesives used, however, few authors cite the commercial name and the manufacturer of these
adhesives, impairing comparisons of results and, in many cases, preventing general conclusions.
About the influence of the mechanical properties of the wood on the anchorage strength. Riberholt [6] proposed an
expression estimating average values of the anchorage strength, considering rupture model as the shear of the wood around
the hole. This expression, adopted by EUROCODE 5:1993 Part 2 [5] and item A.2.2, considers the effect of wood density
with a significant variable. On the other hand, Buchanan and Moss [8] and Bengtsson et al. [9] found no significant
influence of wood density on the anchorage strength of bonded-in steel bars.
The aim of this work was to evaluate the behavior of two bi component epoxy resins and a bi component polyurethane
resin, for the study of anchorage strength of steel bars bonded to beams of Corymbia citriodora and Pinus oocarpa Shiede
in the direction perpendicular to the grain, without considering the natural variations of the mechanical properties. In
addition to the anchorage strength, the objective is to know how some variables influence the anchorage strength, such as:
wood density variations, comparing different species; variations of wood moisture at the time of bonding; and variations in
the thickness of the glue line.
2. Material and Methods
Table 1 shows the structural resins used, their consistencies and commercial suppliers. Polyurethane resin developed
by the Institute of Chemistry of São Carlos (IQSC/USP) was composed of prepolymer A249 and polyol 25015C,
produced from castor oil.
Batchelar and Mcintosh [10], reviewing rupture models in experiments due to improper blends and/or incorrected
epoxy adhesive applications in situ, concluded that the entire sizing operation should be done in a suitable
environment with adequate quality control and specialized people.

Table 1. Resins employed in experiment

Commercial Commercial
Consistency Type
Name Suppliers

Compound Liquid Epoxi Otto Baungart S/A

Sikadur-32 Pasty Epoxi Sika S/A
Polyurethan
Polyurethane Liquid IQSC/USP
e

Specimens manufacture – Corymbia citriodora

In each beam, a resin was used and the mechanical properties of the wood were admitted constant along its length.
The studied beams of Corymbia citriodora (ρ12% = 1000 kg/m3) had dimensions of 5 x 20 x 300 cm with moisture
content initially between 25 and 30% (Fig. 1). From each beam six specimens were obtained, with a length of 25 cm
each. In order to obtain the expected humidity of 15, 18, 20, 24, 28 and 34%, the mass of water contained in each
specimen was controlled, where some were immersed in water and others were air-conditioned in a kiln with forced
ventilation and temperature control until they reach the desired weight. Then, the specimens were packed separately
in transparent plastic bags, where they remained for a period of 40 days to homogenize the mass of water throughout
its volume.

Figure 1. Specimens obtainment in each Corymbia citriodora beam, anchoring bars and diameters of holes for liquid resins.
The experiment was designed so that there were two replications in each observation, considering the possibilities of
excessive defects, as a result of forced drying. For the polyurethane resin, the low results and the two anchorage
rupture behaviors presented initially suggested the possibility of errors in the preparation of the resin, leading to the
repetition of the experiment. Therefore, the specimens were obtained in four beams and the results correspond to the
average of four replications. For Compound resin, the specimens were obtained in two beams, and the results
correspond to the average of two replications. For these resins, each specimen, according to Fig. 1, received four
holes in the direction perpendicular to the grain, with depth of 8.0 cm and diameters of 6.5, 7.5, 8.5 and 9.5 mm.
For the Sikadur-32 resin, specimens were obtained in a single beam. The series of specimens corresponding to the
second beam was neglected, due to excessive retraction defects due to drying. In these specimens, the diameters of
the holes were 8.5, 9.5, 10.5 and 11.5 mm, considering the difficulty of injecting this resin into holes with smaller
diameters. The thicknesses of the glue lines were, respectively, 0.58, 1.08, 1.58, 2.08, and 2.58 mm for the diameters
used.

Specimens manufacture – Pinus oocarpa Shiede

In each beam, a resin was used and the mechanical properties of the wood were admitted constant along its length.
The beams of Pinus oocarpa Shiede (ρ12% = 550 kg/m3) were initially air dried, with moisture content between 11.5
and 12%. From each beam, with dimensions of 5 x 20 x 150 cm, four specimens with a length of 25 cm each were
obtained, according to Fig. 2. In order to obtain the expected moisture, the mass of water contained in the specimens
was checked were immersed in water until reaching the desired weights, corresponding to the expected moisture
content of 12, 15, 18 and 22%. Then, the specimens were packed separately in transparent plastic bags, where they
remained for a period of 30 days to homogenize the mass of water in its volume.

Figure 2. Specimens obtainment in each Pinus oocarpa Shiede beam, anchoring bars and diameters of holes for liquid resins.

Each resin was studied in a single beam, considering the least possible loss of specimens with defects due to air
conditioning. For the liquid resins (Polyurethane and Compound), each specimen received four holes in the direction
perpendicular to the grain with depths of 8.0 cm and diameters of 6.5, 7.5, 8.5 and 9.5 mm.
For the Sikadur-32 resin, in each specimen the holes had depths of 8.0 cm and diameters of 8.5, 9.5, 10.5 and 11.5
mm, considering the difficulty of injecting this resin into holes with smaller diameters. The thicknesses of the glue
lines were, respectively, 0.58, 1.08, 1.58, 2.08, and 2.58 mm for the diameters used.

Steel bars characterization

In all of the specimens, the steel bars received surface cleaning treatment by applying a rotating steel brush until it
reached the white color at the end in contact with the resin. Subsequently, thinner (commonly used for cleaning) was
applied as solvent to remove oil residues. CA-50 steel bars (f yd = 500 MPa) with a diameter of 6.3 mm, axially
requested in two load cycles with monotonic loads, were used, the first cycle being with up to 70% of the ultimate
strength. The adhesion surfaces of the steel bars were evaluated from the average size of the surface fillets,
obtaining, on average, 2.03 cm2 of surface area for each unit of bar length.

Statistical analysis
The results were submitted to statistical analysis to develop the most adequate statistical model, with the purpose of
inferring average responses of anchorage strengths in the 90º direction from the experimental data for each resin.
The statistical methods used were: multiple linear regression analysis, analysis of variance, residue analysis and
normality test for residues.

3. Results and Discussion

Compound resin (liquid consistency) was partially absorbed or drained by small cracks in the wood, thus altering the
anchorage length in the specimens. Even after the initial hardening, the specimens showed small variations in the
anchorage areas, which became significant in the statistical model. For the purpose of comparing the results using
the statistical models for this resin, the mean value of the anchorage area was 16.24 cm², which corresponds to an
average anchorage length of 8.0 cm.
Statistical models present significant variables: moisture content of the wood at the time of bonding and the
thickness of the glue line. Densities for each wood measured in the specimens showed small variations, not being
significant in the studied models, considering that few beams were used for each resin and it was tried not to include
the natural variability of the wood as variable in the experiment.
In some models for the epoxy resins, the residue analysis shows that the variable moisture content could be
presented in quadratic form, but, due to the small number of observations, the linear model is also significant and
valid for statistical models presented. Through the analysis of variance, it can be observed that all the models can be
considered highly significant.
The studies of the residuals versus independent variables showed symmetry indicating zero mean and show that
there is no correlation between the residues and the observed variables, indicated by the uniform dispersion of the
points around zero.
Normal probability tests for residues indicate that the residues and hence the responses follow a normal distribution,
so the tests used in each analysis are appropriate and possible to use.

Corymbia citriodora members

With the polyurethane resin of castor oil, for glue lines with thicknesses of 0.58, 1.08, and 1.58 mm and hole
diameters: 7.5, 8.5, and 9.5 mm respectively, the anchorage rupture were characterized by shearing of the resin in
the glue line. Presented average value to anchorage strength (RA) equal to 3.0 kN and did not depend on the
variations of humidity or the thickness of the glue line. For the 0.08 mm glue line, corresponding to the 6.5 mm
hole, the anchorage rupture also occurred with the resin shear, although another factor is influencing these results
that clearly stood out in relation to the other glue line thicknesses (Fig 3.). Average anchorage strength equal to 10
kN was obtained for air dry wood, with small linear influence decreasing for increases in humidity.
 Figure 3. Anchorage strength and moisture content for polyurethane resin of castor bean oil – Corymbia citriodora member

For the epoxy resins, the statistical models presented in Table 2 were obtained, with increasing linear variations for
increases in the thickness of the glue line and decreasing linear variations for the increase of moisture content.
 Table 2. Anchorage strength and coefficient of determination (R²) in 90º toward in the direction to Corymbia citriodora.

Resin Type Statistical Model R²(%)

Compound RA = 23.23 + 5.050∙e – 0.580∙MC 81.70

Sikadur-32 RA = 18.70 + 3.560∙e – 0.258∙MC 72.80

Where: RA = anchorage strength (in kN); e = glue line (in mm); MC = moisture content (in %).

Fig. 4 presents the average values to anchorage strengths with glue lines thickness variations, for wood moisture content of
12 and 22% and anchorage strengths with wood moisture variations of 15 to 30%, for a glue line thickness equal to 1.58
mm.

Figure 4. Comparative anchorage strength values for tree structural resins varying the glue line thickness and moisture content in Corymbia citriodora:
(1) Theoretical values; (2) Bar rupture limit; (3) Bar yeld limit.

Pinus oocarpa Shiede members

Anchorage strengths with the polyurethane resin in Pinus oocarpa Shiede specimens (Table 3), showed markedly greater
values for the thickness of the glue line 0.08 mm, with tendencies to decrease rapidly as this variable grows, (Fig. 5).
Regarding to moisture content, it presented a behavior similar to that observed in some analyzes, with the variable MC 2
behaving better in the statistical model than the variable MC. Considering the few observed results, the model presented in
Table 3 was chosen for this resin.

Table 3. Anchorage strength and coefficient of determination (R²) in direction to 90° at Pinus oocarpa Shiede.

Resin Type Statistical Model R²(%)

Polyurethane RA = 12.60 – 0.0276∙MC² - 1.65∙ln(e) 97.80

Compound RA = 28.64 – 1.04∙MC + 3.62∙e 92.00
Sikadur-32 RA = 27.10 – 0.559∙MC + 1,23∙e 79,70

Where: RA = anchorage strength (in kN); e = glue line (in mm); MC = moisture content (in %).
Figure 5. Anchorage strength in relation to glue line thickness for polyurethane resin of castor bean oil in various moisture content for hole diameter equal
to 6.3 mm.

For the epoxy resins, the statistical models are similar to those obtained with Corymbia citriodora wood, in which the
anchorage strengths present increasing linear variations for increases in the thickness of the glue line and decreasing linear
variations in the increase of moisture content presented in Table 3.
Fig. 6 presents the average values to anchorage strengths with glue line thickness variations, for wood moisture content of
12 and 22% and anchorage strengths with wood moisture variations of 12 and 22%, for a glue line thickness equal to 1.58
mm.

Figure 6. Comparative anchorage strength values for tree structural resins varying the glue line thickness and moisture content in Pinus oocarpa Shiede:
(1) Theoretical values; (2) Bar yeld limit.

Figure 7 shows, comparatively, the effect of the apparent density difference between Corymbia citriodora and Pinus
oocarpa Shiede, also considering the effects of moisture content variation and epoxy resin type effects on anchorage
strength, for glue line thickness equal to 1.58 mm.
 Figure 7. Comparative anchorage strength values for two epoxy resins in Corymbia citriodora and Pinus oocarpa Shiede.

It is observed that the density variation among the wood species was not very significant. The results should be evaluated
by considering the type of epoxy resin used and the differences in moisture.
For the dry-to-air woods, Sikadur-32 resin showed higher strength, about 5% in Pinus oocarpa Shiede compared to
Corymbia citriodora. At 22% of moisture content, the anchorage strength in Pinus oocarpa Shiede was lower by 10%,
compared to Corymbia citriodora. For Compound resin, the strength of anchorages in Corymbia citriodora were 11%
higher, with moisture content of 12%, and 40% higher, with moisture content of 22%, compared to Pinus oocarpa Shiede.

4. Conclusions

The results of this study permit us to conclude:

- All anchorage rupture models to the polyurethane were instantaneous, characterizing brittle breaks with resin
shear. Thickness of the glue line 0.08 mm stands out because it has higher anchorage strengths compared to other
thicknesses. The polyurethane resin of castor oil studied is not suitable for the proposed purposes, considering that
the reaction with the moisture contained in the wood incorporates CO 2 bubbles and, consequently, decreases its
shear strength in the resin;

- Epoxy resins exhibited: glassy consistency after hardening and anchorage rupture occurs initially due to loss of
adhesion on the steel surface and subsequent loss of mechanical adhesion, with the anchorage strength becoming
progressively smaller as the bar is removed;

- Sikadur-32 resin presented better anchorage results at high moisture contents, although the resin was more difficult
to use;

- The use of liquid epoxy resin requires replacement in the holes after initial hardening due to absorption or cracking
in the wood;

- Variations of the anchorage strength with epoxy resins, relative to the independent variables, are as follows: linear
negative variations with increasing moisture content and positive linear variations with increasing glue line
thickness;

- The difference in apparent density between Corymbia citriodora and Pinus oocarpa Shiede did little to influence
the anchorage strength;

- Behavior of all epoxy resins should be evaluated in different moisture content levels for each type of wood,
considering the anchorage strength, the rupture model, the time available to use the mixture of the components and
the difficulties of application of the product.

ACKNOWLEDGMENTS
 For all the provided support, the authors thanks the Fundação de Amparo à Pesquisa do Estado de São Paulo
(FAPESP) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES).

REFERENCES
[1] Madsen, B. Timber connections with strength and reliability of steel. International Wood Conference, Toronto, Canada, p. 4-504, 4-
511, 1996.
[2] Molina, J. C. Análise do comportamento dinâmico da ligação formada por barras de aço coladas para tabuleiros mistos de madeira
e concreto para pontes. São Carlos, 2008. Tese (Doutorado em Engenharia de Estruturas) – Escola de Engenharia, Universidade de
São Paulo, São Carlos, 2008.
[3] Aicher, S.; Gustafsson, P.; Wolf, M. Load displacement and bond strength of glued-in rods in timber influenced by adhesive, wood
density, rod slenderness and diameter. 1 st International RILEM Symposium on Timber Engineering, Stockholm, Sweden, Sep. 13–
14, p. 369–381, 1999.
[4] Bainbridge, R. J.; Mettem, C. J. Bonded in rods for timber structure: a versatile method for achieving structural connections. The
Structural Engineering, v. 77, n. 15, p. 24-27, 1999.
[5] European Committee for Standartization. EUROCODE 5: Desing of timber structures, Part 1-1: General rules and rules for
buildings, ENV 1995-1-1, European Committee for Standardization, Brussel, Belgium, Dec, p. 110, 1993.
[6] Riberholt, H. Glued bolts in glulam: proposal for CIB Code, Proc, Of the CIB – W18 Meeting, Timber Structure Meeting Twenty
one, Parkville, Vancouver Island, Canada, p. 21-72, 1988.
[7] European Committee for Standardization. EUROCODE 5: Design of timber structures, Part 2, Bridges. UK, London, 45 p, 1993.
[8] Buchanan, A.; Moss, P. Design of epoxied steel rods in glulam timber. Pacific Timber Engineering Conference, Rotorua, New
Zealand, March, p. 286–293, 1999.
[9] Bengtsson, C.; Kemmsies, M.; Johansson, C. J. Production control methods for glued-in rods for timber structures. 6 th World
Conference on Timber Structure, Vancouver, Canada, Jul 31-Aug 3, p. 7,4,1, 2000.
[10] Barchelar, M. L.; Mcintosh, K. A. Structural joint in glulam. 5 th World Conference Timber Engineering, Montreux, Switzerland, v.
1, p. 289-296, 1998.
 APÊNDICE A – Título (se houver)

Segundo a norma NBR 14724 (2005), os apêndices constituem os textos ou documentos elaborados pelo

autor, a fim de complementar sua argumentação, sem prejuízo da unidade nuclear do trabalho.
 ANEXO A – Título (se houver)

Segundo a norma NBR 14724 (2005), os anexos constituem os textos ou documentos não elaborados pelo

autor, que servem de fundamentação, comprovação e ilustração.