INTRODUCTION
Single cantilevered poles, having polygonal or circular crosssections, can
be made of wood, aluminum, steel, reinforced concrete, Fibre Reinforced Polymer
(FRP) or combinations of these materials. The main factors that influence the
choice of a power supply post are mechanical properties, fabrication, transport
and erection expenses, average life and maintenance cost. Permanent load acting
on electric poles is from self weight and the weight of electric cables, while
the variable load is produced by wind storm, temperature variation and hoarfrost;
it has been shown (Kudzys, 2006) that 70% of the electric
lines failure have been caused by strong wind and ice storm inducing lateral
loading.
Nonprestressed or prestressed steel reinforcement for concrete poles have
been successfully used but the corrosion of steel reinforcement has caused high
maintenance and replacement cost. The presence of salts, chlorides and other
severe aggressive environments such as combination of moisture, temperature
variations and sulfate attack are weakening the protection provided by the alkalinity
of the concrete on the steel reinforcement thus resulting in corrosion at high
rates. Alternatives to deal with this issue have been previously applied using
repair techniques of damaged RC poles based on FRP external wrapping (Chahrour
and Soudki, 2006). Civil engineers have studied the possibility of mitigating
corrosion problems related to steel by replacing it with glass fibrereinforced
polymer composite materials in case of columns (Choo et
al., 2006) and beams (Almusallam and AlSalloum,
2006). Introducing of such new advanced composite materials underlined the
possibility of avoiding the disadvantages of corrosive, electromagnetic interfering
traditional steel reinforcement. Additionally, superior mechanical properties,
high strength to weight ratio, revealed the suitability of FRP composite materials
for structural uses (Fib Bulletin 40, 2007).
Experimental programs carried out on high strength concrete pylons reinforced
with prestressed Carbon Fibre Reinforced Polymeric (CFRP) resins of a 27 m height
subjected to bending and to freeze thaw cycles have proven advantages like durability
and low self weight by minimizing reinforcement amount and concrete cover (Terrasi
et al., 2001).
An alternative to enhance the flexural behavior of power transmission hybrid
poles is the development of spun cast concretefilled FRP tubes. Important experimental
tests (Naguib and Mirmiran, 2002; Fam
and Rizkalla, 2003; Qasrawi and Fam, 2008) and finite
element modeling (Fam and Son, 2008) on hybrid concretefilled
FRP tubes with or without steel longitudinal bars have emphasized significant
high performance characteristics in bending. Increasing in flexural strength
of spuncasting elements utilizing fiber reinforced cementitious composites
such as short carbon fibers and polyvinyalcohol (PVA) embedded in a cement matrix
has been reported by Kaufmann and Hesselbarth (2007).
The study is focused on the theoretical and experimental analysis of centrifuged concrete poles reinforced with GFRP longitudinal bars and flexible composite spiral reinforcement to avoid the corrosion of steel reinforcing products.
FABRICATION OF THE CENTRIFUGED HYBRID HOLLOW COLUMNS
Generally, the fabrication technology of the centrifuged hybrid hollow concrete columns reinforced with GFRP bars and composite spiral for confinement is based on the one established for columns with steel bars and spirals implemented by Somaco SA, Romania. However, some particularities due to the nature of the materials making up the hybrid product are taken into account and correspondingly treated. The following materials characteristics have been utilized based on the data determined in supplier laboratory (Table 1). GFRP bars for longitudinal reinforcement produced by Fiberline Composites A/S, Denmark and GFPP spiral strip designed by our team and fabricated by CEPROPLAST SRL, Romania. A presentation of the geometric characteristics of all elements is shown in Table 2.
A summary of the technological steps to fabricate the tapered columns follows:
the reinforcement cage (Fig. 1) has been set up on rigid fixing
rings and this assembly has been placed in the formwork using distancekeepers;
six fixing rings have been utilized to setup the longitudinal reinforcement
(8 GFRP bars); the formwork was steelmade (Fig. 2), from
two pieces joined by screws; on the formwork length there are transversal circular
ribs, guided by the wheels of the centrifuge machine (Fig. 3).
The reinforcing cage has been placed in one half of the formwork and then the
concrete has been poured; the amount of concrete has been determined so as to
fill exactly the formwork; after that, the formwork was closed and placed on
the centrifuge machine; a centrifuging time of 15 min, similar to that utilized
in case of steel reinforced columns has been selected; about 1 h after the centrifuge
operation the formwork has then been kept in place; the formwork has been transported
in the treatment chamber and kept for 2 days; the formwork was then removed
and the column stored for 28 days until the complete curing of concrete.
Table 1: 
Summary of mechanical materials properties 

Table 2: 
Summary of geometrical characteristics 


Fig. 1: 
The FRP reinforcement cage 

Fig. 2: 
The steel formwork for centrifuged column 

Fig. 3: 
The rotating wheels of the formwork 
THE EXPERIMENTAL PROGRAM
The columns have been tested in horizontal position to bending at an age of 28 days, according to usual procedures, determining the maximum forces and displacement recorded at the free ends of concrete poles reinforced with GFRP bars. Columns with two reinforcing solutions have been fabricated and tested: the first column type specimen with a length of 3 m has been reinforced with smooth GFRP bars and the second type column, 4 m long, has been reinforced with ribbed GFRP bars; the longitudinal profile of both bars are shown in Fig. 4.
An adequate instrumentation has been designed and installed to characterize the structural response of the tested specimens. A specially designed testing stand was constructed to provide the adequate experimental conditions (Fig. 5). The locations of the Linear Voltage Displacement Transducers (LVDTs) to determine the lateral deflections are positioned as shown in Fig. 5. The columns were clamped at one end and the action was exerted by pulling the free end. A loading cell installed on the pulling device has been utilized to measure the corresponding applied load. Successive loading and unloading cycles were applied up to failure and the corresponding displacement for each applied load were measured.
A loaddisplacement diagram for the 3 m long column reinforced with plain GFRP bars is shown in Fig. 6 for all five loading cycles applied to the specimen. It has been noticed that the first crack occurred at about 4000 N during the third loading cycle; after each cycle the permanent deformation was recorded as shown in Fig. 6. A recorded peak load equal to 6400 N was accompanied by a maximum horizontal displacement equal to 84 mm of the free end. Four meter long specimen has been subjected to eight loading cycles reaching an ultimate peak value equal to 17000 N. In Fig. 7, the repeating loaddisplacement curves for this specimen are shown. In case of the second column a first crack occurred at 4500 N and a more uniform cracking pattern has been obtained and a maximum horizontal deflection equal to 257 mm has been recorded (Fig. 8). A significant difference have been observed between the behavior of columns reinforced with smooth round GFRP rods and GFRP ribbed rebars, with the geometry visualized in Fig. 4.
Some failure modes typical for the bent cantilevered columns reinforced with
GFRP rebars having smooth or ribbed geometries have been observed.

Fig. 4: 
Types of GFRP longitudinal rebars for the hollow concrete
columns 

Fig. 5: 
Instrumentation of the test samples with LVDTs (a) the 3 m
long column and (b) the 4 m long column 
Experimental
study carried out by other research teams (Hao et al.,
2007) has revealed the sharp difference between the nature and the values
of the bond strength for smooth and ribbed GFRP rebars, explaining the overall
behaviour of composite reinforced columns.
The cracking process developed progressively, as shown in Fig.
9, until the failure loads have been reached. It has been noticed that the
posts failed by concrete crushing, typical for over reinforced elements. The
GFRP rebars did not attain their ultimate strength because of the poor bond
stress developed between the concrete and the composite bars. A better distribution
of cracks (Fig. 9) has been noticed in case of the column
reinforced with ribbed bars as a result of better bonding strength between concrete
and the GFRP reinforcement. A direct effect of this aspect was the increased
failure load (17000 N) and a limited plastic behavior before failure.

Fig. 6: 
Forcedisplacement curve for the 3 m long column corresponding
to the end column LVDT (wired transducer 1) 

Fig. 7: 
Forcedisplacement curves in case of the 4 m long column
(wired transducer 0) 

Fig. 8: 
The total displacement at free end of the 4 m specimen 

Fig. 9: 
Crack development in the 4 m long post 
Cracks development in the tension part of concrete poles reinforced with GFRP
bars due to bending depends on parameters such as crack spacing, bonding strength
between composite bars and concrete and strain values in transverse and longitudinal
reinforcements (Newhook et al., 2002). The spuncast
concrete pole show quite a large displacement at the free end from bending,
with ultimate crack observed at around 1.4 m distance from the fixed end of
pole, feature that is common in case of experimental pole tests (Kaufmann
et al., 2005).
THEORETICAL ANALYSIS
Power supply posts are special structural elements whose design although apparently simple raises specific problems. The axial load due to the self weight and power supply installations weight is undertaken exclusively by the concrete area, the FRP considered reinforcement will be subject only to flexural elements design rules, in terms of strength requirements. The determination of bending moment capacity depends on the location of the neutral axis, whose position is determined by equation of equilibrium between the concrete compression and the FRP tension on the cross section. As stated before the post is idealized as a cantilever beam fixed at the bottom end, subjected to flexure caused by the transversely acting loads such as wind pressure or, accidentally, earthquake. Since, the GFRP bars resist only to tensile stress occurring over the crosssection, on the tension side, their contribution in compressive loading resistance is negligible.
In many cases it is considered that the design of such elements is controlled by stiffness, corresponding to serviceability limit state design, by checking the effective deflections versus allowable ones. As far as the ultimate limit state is concerned, there are two accepted failure modes in case of flexural FRP reinforced concrete elements: either concrete crushing or FRP rupture. Both concrete and FRP longitudinal reinforcement are brittle materials determining sudden failure. However, in case of concrete crushing there is a little plastic behavior making this type a slightly more desirable one. Generally, if the design is performed to induce this failure mode, the serviceability requirement concerning the deflections is met.
A particular approach has been used to determine the stiffness of the cantilever
column. It has been considered that both concrete and GFRP bars contribute to
the total stiffness:
where, K is the total bending stiffness of the reinforced column, K_{c} is the bending stiffness provided by the concrete gross section and K_{f} is the bending stiffness provided by GFRP bars.
For uniformly tapered poles the moment of inertia of concrete section may be
conservatively taken as the gross moment of inertia at a distance onethird
of span from the smaller end of the column. The appropriate value of the pole
concrete stiffness is also given in the above mentioned report:
where, E_{c} is the concrete modulus of elasticity and I_{g} is the gross moment of inertia of concrete as specified before.
With geometrical and mechanical data already presented K_{c} = 8.544×10^{11} N mm^{2}.
The stiffness provided by the GFRP longitudinal bars has been calculated considering an equivalent composite tubular section having a total area equal to the sum of the cross section of 8 composite bars. Using the characteristics of GFRP bars (Table 1) and the geometry of the reinforcing scheme (Fig. 10), the stiffness value of the composite equivalent tube has been evaluated, K_{f} = 5.917×10^{11} N mm^{2}. Compared the experimental and theoretical elastic deflections determined with total stiffness, K, an approximation of about 7% has been noticed.
Prediction of the most probably way to produce a failure, leading to the ultimate
limit state, may be done by comparing the effective FRP reinforcement ratio
to the balanced FRP reinforcement ratio. These two parameters are calculated
with the following relations:
where, ρ_{f} is the FRP reinforcement ratio, ρ_{fb}
is the FRP reinforcement ratio producing balanced conditions, A_{f}
is the GFRP reinforcement area, A_{g} is the crosssectional gross area,
β_{1} is the reduction factor for concrete strength, or stressblock
factor for concrete (ISIS Canada, 2001), f’_{c}
is the specified compressive strength of concrete C45/50 (Table
1), f_{fu} is the design tensile strength of GFRP bars, (Table
1), E_{f} is the guaranteed modulus of elasticity of GFRP, (Table
1) and ε_{cu} is the ultimate strain in concrete, (Table
1).
The ratio of the balanced neutral axis depth c, to the effective depth of the section d_{1}, can be expressed using strain compatibility shown in Fig. 10:

Fig. 10: 
The reinforcing pattern and the stressstrain distribution
in concrete pole reinforced with GFRP 
where, ε_{cu} is the ultimate compressive strain in concrete, c is the depth of the neutral axis, d_{1} is the effective depth and ε_{frp, 1} is the corresponding value of the strain in FRP.
It has been found by trials that the depth of the neutral axis exceeds the column wall thickness (c = 60 mm). In the next step the compressive and tensile resultants, C_{c} and T_{frp} are determined. The compression force, C_{c}, on the concrete section is calculated with:
where, A_{c} is the area of the concrete in compression. The coefficients
α_{1}, β_{1}, Φ_{c} are determined according
to (ISIS Canada, 2001): α_{1} is the stressblock
factor for concrete α_{1} = 0.85–0.0015 f’_{c}≥0.67,
β_{1} is the stressblock factor for concrete β_{1}
= 0.97–0.0025 f’_{c}≥0.67 and Φ_{c} is the
material resistance factor for concrete = 0.65.
The area of concrete in compression A_{c} shown in Fig.
10, it is equal to the area of the annulus, A_{a} calculated with
(Wang and Salmon, 1979; Kocer and
Arora, 1996):
where, D is the external diameter of concrete pole and d is the internal diameter of concrete pole. The values for θ_{1}, θ_{2}, the angles figured out in Fig.10, are calculated with:
The distance denoted by y, Fig. 10, measured between the neutral axis and the horizontal axis of entire cross section is determined with:
Using the mechanical and geometrical characteristics of the GFRP reinforced column a value was found for C_{c} = 93.85 kN. The total tensile force T_{frp} of GFRP bars can be expressed as:
where, Φ_{frp} is the resistance factor for GFRP, 0.4 (ISIS
Canada, 2001), A_{f, i} are the areas of the GFRP bars located in
the tension region (Fig. 10), ε_{frp, i} are
the strains in the tension GFRP bars (Fig. 10) and i is the
current number of GFRP bars (Fig. 10).
After all algebraic calculations, the tensile force was determined: T_{frp} = 94.69 kN. These values of C_{c }and T_{frp} have been evaluated after a number of trials, selecting proper values of c. Since, the column external and internal diameters are known, D = 250 mm and d = 150 mm, the neutral axis depth is also known and the bending moment capacity M_{r} can been evaluated as:
where, f_{rp, i} is the tensile stress in the corresponding GFRP bars,
denoted 1, 2, 3 in Fig. 10 and e_{i} represents e_{1},
e_{2}, e_{3}, respectively (Fig. 10).
Table 3: 
Tensile stress in GFRP bars 

Using the mechanical properties of GFRP bars and corresponding distances, the following effective stress calculated in the reinforcing elements (Table 3).
The bending moment capacity determined with Eq. 11 using
the data from Table 3 is M_{r} = 52.124 k Nm. This value of the column
flexural capacity satisfies the condition:
required by ISIS Canada (2001). In Eq.
12 the significance of notations are: M_{cr} is the cracking moment,
f_{r} is the modulus of rupture of concrete equal to ,
I_{t} is the transformed section moment of inertia and y_{t}
is the distance from the neutral axis to the extreme tension fibre.
Using the mechanical and geometrical properties of the GFRP reinforced column,
a cracking moment value M_{cr} = 3.206 k Nm has been calculated, that
satisfies Eq. 12.
DISCUSSION
Since, the density of steel is 7850 kg m^{3} and the density of GFRP bars 2200 kg m^{3}, a saving in weight of about 40% of the reinforcement has been obtained. The result is based on the same number of reinforcing bars utilized for centrifuged hollow columns fabricated with the same technology.
The short term tensile strength of GFRP bars (1100 MPa) is much higher than
that of steel bars (435 MPa), but the ultimate long term tensile strength of
GFRP bars should be affected by a retention coefficient not exceeding 0.5 for
a span life equal to 100 years (Fib Bulletin 40, 2007).
The GFRP bars contribute to the stiffness of the member with 44% less than
the steel bar, therefore the total deflection of the hollow column reinforced
with composite bars is larger than that reinforced with steel bars.
With a maximum 257 mm lateral deflection, representing a relative displacement equal to 6.67% the column shows a good flexibility under transverse loads.
The two different geometries of the longitudinal GFRP reinforcing bars determine
significant differences in structural responses (Fig. 6, 7),
in terms of peak force values and lateral displacements. The forcedeflection
curves underline the superior behavior of the ribbed composite bars hollow column
with respect to the smooth bar reinforcing. This superiority can be also noticed
in terms of compactness of loading/unloading curves and of residual deformations
as well.
CONCLUSION
Hollow concrete columns reinforced with GFRP longitudinal bars and flexible composite spirals can be designed, installed and exploited in a similar manner with the columns reinforced with steel products.
The centrifuged procedure can be utilized to fabricate GFRP reinforced hollow concrete columns. When concrete poles are reinforced with smooth GFRP bars the bonding is not sufficient and this solution does not enable an efficient use of materials characteristics. Ribbed GFRP bars provide a much better bonding enabling the loading of the column to much higher values. Since, the elastic modulus of GFRP bars is only slightly higher then that of concrete, the reinforcement contribution to the stiffness of the element is not substantial. Existing formulas in design norms and codes are suitable for this type of element, having in mind the particularities of composite bars as reinforcements. Finally, all advantages of FRP composite bars (corrosion resistance, lower specific weight, convenient electric and magnetic properties) can be met and valorized in this structural element.
ACKNOWLEDGMENT
This study has been partially supported by a research project on hybrid structures made of polymeric composites and traditional building materials, Program PN IIIdeiCod 369.