Modeling and verification of response of RC columns strengthened in flexure with mechanically fastened FRP

2.1. Test specimens and materials

Three RC Column Specimens designed and constructed with a height of 1 meter and dimensions 20×20 cm. A 100×30 cm foundation designed to fix the column to the reaction frame and 30×50 cm concrete element used to apply the axial and lateral loadings.

Fig. 1Geometry of the RC columns

Fig. 2Columns strengthened using FRP

2.2. Specimen layout

In this research, a new mechanical connection is proposed. The new proposed mechanical connectors fastened the FRP layer to the RC column substrate and was designed using steel angles and bolts. Designing the MF connections was based on initial forces developing in the members during the loading. Fig. 3 shows the initial forces was used in design approach.

In the laboratory, First the surface of the RC columns prepared and the same layer of longitudinal FRP used on two opposite sides of the RC columns. In C1 column FRP jackets and in C2 column MF used to fix the FRP layers to the RC column. After the FRP system was allowed to cure for about 24 hours the anchorage system assembled. For retrofitting the C2 column first three holes spaced 30 centimeters drilled along the column and the bolts placed in these holes and fixed with two nuts in both sides. An angle with two stiffeners designed and constructed for each side of the RC column. Fig. 3 shows the details of MF systems were used in strengthened C2 column.

Fig. 3Mechanical fasteners used in C2 column

The steps for implementing the mechanical connection included:

1) The locations of bolts were first marked on the surface of the concrete.

2) Holes to the depth of 15cm were drilled at the column-foundation connection.

3) Hardeners welded to the angles and the angles placed in pre-defined positions.

4) Bolts placed in the holes and the holes filled with epoxy adhesive.

2.3. Test setup and loading

The test setup consisted of reaction frame supporting lateral and vertical hydraulic actuators. so, all columns tested under combined axial and lateral loads. Preceding the application of the lateral load, RC columns first loaded with constant axial load using a hydraulic jack at top. The 200 kN axial load applied primarily is approximately 25 % of the ultimate axial load capacity. After initial axial loading, lateral load applied in a displacement control mode.

The applied displacement amplitude was a fraction of the estimated tip yield displacement of the C0 column, $\mathrm{\Delta }y$ (~10 mm) reached from the numerical simulation. The lateral cyclic displacement level selected as multiples of 0.5$\mathrm{\Delta }y$ and three cycles applied at each displacement level. Fig. 4 shows the details of the loading system and the loading protocol applied similarly to all the specimens.

Fig. 4Test setup details and loading protocol

4.1. Properties of materials

4.1.1. Concrete and steel

Nonlinear damaged plasticity model for inelastic behavior of concrete assumed and concrete isotropic damage plasticity model is designed and used in nonlinear analysis of the concrete columns. The continuum, plasticity based, damaged model used for concrete modeling. Damage plasticity assumed to characterize the compressive and tensile response of concrete as shown in Fig. 7. At the beginning, the stress-strain relationship is linearly elastic under axial tension until the value of the failure stress $f_{to}$ reached. Failure stresses in concrete block converted to replace micro cracks in it. Beyond the state of the failure stress in concrete, stress-strain response designed by softening characteristic. Under compression, the response is linear until the value of initial yield $f_{co}$. After attaining the ultimate stress $f_{cu}$ in the plastic zone, the response of concrete characterized by the stress hardening followed by strain softening. Therefore, concrete stresses determined unloading from any point on the strain are [11]:

1

$f_t=E_c\left({\epsilon }_t-{\epsilon }_t^{pl}\right)\left(1-d_t\right),$

2

$f_c=E_c\left({\epsilon }_c-{\epsilon }_c^{pl}\right)\left(1-d_c\right),$

where $E_c$ is the modulus of elasticity of concrete. Then, the effective tensile and compressive cohesion stresses of concrete were estimated as:

3

${\stackrel{-}{f}}_t=\frac{f_t}{(1-d_t)}=E_c\left({\epsilon }_t-{\epsilon }_t^{pl}\right).$

Linear Kinematic hardening model which is a bilinear model for steel was assumed for rebar and stirrup.

Fig. 7Concrete damage plasticity model [11]

a) Tension behavior associated with tension stiffening

b) Compressive behavior associated with compression hardening

Fig. 8Alternative stress-strain laws for steel reinforcing bars: a) accurate, b) simplified

a)

b)

5.1. Evaluation of behavior of not-retrofitted column

In the first numerical model, the concrete column was considered without any retrofitting. This column is similar to the Laboratory specimen and used as the benchmark model. Fig. 9 shows the finite element and lateral load-displacement diagram from the finite element analysis along with the experimental results.

Fig. 9Finite element model of C0 specimen and lateral load-displacement diagram

5.2. Evaluation of behavior of EB-FRP strengthened column

In this model, the concrete column was retrofitted with longitudinal layers of FRP. Based on the fact that the longitudinal layers of the FRP can be restrained by FRP jacketing, FRP wrapping was used at the bottom and top of the column. The height of the wrapping was 16 cm according to the plastic hinge length. Table 2 shows the characteristics of the used polymeric fibers. The lateral load-displacement curve and finite element model of this column shown in Fig. 10.

Fig. 10Finite element model of C1 specimen and lateral load-displacement diagram

Although numerical modeling indicates an increase in the flexural strength of the column reinforced with longitudinal FRP and wrapping in both ends, the results illustrate that with de-bonding of FRP at first cycles of lateral loading the stiffness and overall behavior changed, and the strengthened column could not reach the preferred resistance. Moreover, with de-bonding the FRP layers both RC column and FRP layers resist load individually and could not improve the desired composite action.

5.3. CFRP-strengthened column with proposed mechanical connection

In this section, a new detail proposed for fixing CFRP at the column-to-foundation connection and column substrate. The L Profile St37 Angle used in the proposed mechanical connection. The angle has a tensile strength of 2400 kg/cm² and a final stress of 3600 kg/cm² and the modulus of elasticity is equal to 2.1×10⁶ kg/cm², according to the material tests. In Fig. 11, the finite element model of the column is shown.

As seen in the results, the proposed retrofitting detail changed the failure mode and increased the bending capacity and lateral displacement of strengthened column. The computational results show that the fracture did not occur in the linear section of the curve as de-bonding of FRP layers did not occur during the loading and the uniform load-displacement curve indicates the ideal composite behavior for this specimen. Therefore, based on the results new proposed mechanical connection had an effective role in overall seismic behavior of retrofitted column.

Fig. 11Finite element model of C2 specimen and lateral load-displacement diagram

5.4. Numerical modeling and experimental results

By observing the results obtained from the finite element analysis and comparing them with the experimental results, close proximity is obtained. The ultimate lateral load and lateral drift obtained from the finite element analysis and the experimental tests compared in Table 3. Also, Fig. 12 shows lateral load-displacement curves of all specimens. By comparing the capacity curves of the columns C0, C1 and C2, the inadequacy of traditional methods in retrofitting of concrete columns is determined. Results indicate that the new proposed anchorage system improved the overall lateral strength in comparison with the similar column strengthened with traditional externally bonded system.

According to Fig. 12 the combination of longitudinal FRP and anchorage system displayed the best response characteristics. Using anchorage system improved the overall flexural capacity and the lateral displacement of FRP-strengthened RC column. But the most important point is the ductile failure of the strengthened RC column due to the anchorage system.

As can be seen, the value of the ultimate lateral strength of C1 column increased 64 % compared to the non-reinforced column, but the ultimate lateral drift does not change significantly in this column. The ultimate lateral drift is 2.7 % in column C0 while it is 3.9 % in the C1 column. Besides, as shown in Fig. 12 the stiffness of the C1 column at the beginning of the linear section has increased due to FRP layers. However, after separating the layers from the concrete surface, the linear section of the capacity curves for C1 and C0 column extend in parallel. Based on the results, the C2 column had a better performance than the other specimens. The ultimate lateral strength and the ultimate lateral drift have been significantly enhanced and the capacity of the column has increased. The C2 column has experienced a final resistance of 19.92 kN, which was nearly double that of the control specimen. On the other hand, the ultimate drift has increased to 61 mm in this specimen. So according to the mentioned points, it seems that the flexural retrofitting of concrete columns with the conventional methods do not satisfy the seismic retrofitting criteria.

Fig. 12Lateral load-drift curves