Confined length of reinforced concrete columns at various axial load levels

The seismic performance of reinforced concrete columns subjected to various levels of axial load is assessed in the paper. An equation relating the confined region to the applied axial load is proposed and compared to the Algerian seismic code RPA99/V2003 specifications, based on the concrete spalling length measured on 16 specimens under different axial load ratios, obtained from the experiments. Results indicate that the length of the confined region of a column subjected to high axial load is underestimated in the above mentioned specifications.


Introduction
The most frequent damage observed during earthquakes is the crushing of concrete, buckling of longitudinal reinforcement, and opening of transverse reinforcement. This damage is attributed to poor concrete confinement at the RC columns near the column-beam joints, inadequate horizontal reinforcement detailing, and high axial loads. Over the past 30 years, numerous researchers have conducted investigations aimed at estimating flexural behaviour of reinforced concrete columns. Many parameters, such as the axial load ratio, transverse reinforcement ratio, configuration of transverse reinforcement, main reinforcement ratio, concrete strength, and yield strength of steel reinforcement, can influence seismic performance of reinforced concrete columns [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16]. The purpose of these research endeavours has been to investigate seismic behaviour of RC columns by analysing the effect of some crucial parameters on the overall performance of RC columns. Numerical models for specimen testing have been developed and analysed. The analytical results show reasonable agreement with experimental ones. The analysis does not only accurately predict the stiffness, load, and deformation at the peak level, but captures the post-peak softening as well. It has been shown that both factors, axial load intensity and transverse reinforcement ratios, have an important influence on the strength, maximum sustained displacement, and energy dissipation capacity of columns. The effect of the axial load ratio is thoroughly investigated in this paper through analysis of experimental results. Furthermore, sixteen cantilever column models with a square section as shown in Figure 1 and Figure 2 are tested under quasi-static unidirectional and bi-directional displacement, combined with different axial load level [17][18]. Finally, equation relating the confined region to the applied axial load intensity is proposed and compared to those recommended in the current Algerian seismic code RPA 99/v2003, Eurocode 8, and ACI 315. [19][20][21].

Effect of axial load on normalized horizontal load -rotation angle hysteresis loops
An increase in constant axial load from to for specimens under a unidirectional horizontal load will result in the reduction of deformability capacity, defined in this study as the deformation corresponding to 20 % drop in flexural strength, and in the increase of flexural strength degradation after the peak. However, the hysteresis loops for specimens under the high axial load will be fatter than those for specimen under a moderate axial load, which show some pinching as illustrated in Figure 3.

Effect of loading path on the normalized horizontal load -rotation angle relationship
Bi-directional loading reduced the flexural strength capacity for small-scale specimens under and increased the dissipated energy as illustrated in Figure 4(a) and Figure 4(b). However, for specimens under the axial load and bi-directional horizontal loading, flexural strength, deformability and dissipated energy, were reduced as illustrated in Figure 4(c) and Figure 4(d).This may be attributed to concrete damage, since for bi-directional loading, concrete located in the 4 faces at the column base will be subjected to high compression force, whereas only 2 faces are subjected to high compression for unidirectional loading.

Axial strain-normalized curvature relationship
The axial load intensity had a significant effect on the column shortening as shown in Figure 5. The axial strain was defined as the elongation/shortening measured at the column base for a distance equal to the column depth, divided by the column depth. As can be seen, specimen D1N3, unidirectionally loaded and subjected to a moderate axial load, exhibited more elongation, 0.35 %, than shortening, 0.15 %. In the same manner, specimen D1N6, subjected to axial load two times greater compared to specimen D1N3, showed only shortening throughout the test process. It is worth noting here that the bi-directionally loaded specimen D2N6 exhibited only shortening from the beginning of the test, even though the specimen was under a moderate axial load, as shown in Figure 6. It can be seen that the loading pattern, as well as the axial load level, have a significant influence on the seismic performance of columns.

S train distribution in longitudinal and transverse reinforcements
During the test, concrete cover spalled first, which was followed by buckling of longitudinal corner reinforcement. As the test progressed, concrete started crushing at the corners and gradually the load carrying capacity was reduced as damage penetrated toward the column core. This state can be seen in Figure 8, which shows strain distribution at the east side of shear reinforcement at 1 % and 3 % drift, respectively, at 3 different hoops locations along the column height, as shown in Figure 9.b. The strain of the external hoop started to reduce with an increase in strain of the internal hoop. This means that concrete at the periphery of the core was severally damaged and, hence, the effective concrete area reduced considerably. Taking into account the observed damage and the results found using the shear reinforcement strain distribution, the column section was classified into several distinct areas. These areas are shown in Figure 10 and classified from 1 to 4. The number in each area indicates its crushing order. Strain distribution for longitudinal corner bars over the height, is shown in Figure 11 at 2 % drift. The maximum-recorded strain took place either in the second or third layer shown in Figure 8.a.

Observed damage
For specimen L1D60 subjected to unidirectional loading in the north-south direction and presenting an axial load ratio of 60 %, the spalling of the cover concrete reached a height of 1.5 of the column depth. Damage to the cover concrete (spalling) was observed at the base of specimen L1N60 for a height of 1 to 1.5 of the column depth. The north-east and north-west corner longitudinal rebar was buckled at the height of 0.5 of the column depth from the base. A slight buckling of the corner longitudinal rebar was observed for specimen D1N6 and spalling of the cover concrete reached a height of 1.0 to 1.4 of the column depth from the base. For specimen L1N6B, only two longitudinal rebar out of 12 reinforcement bars buckled at the end of the test, and 4 to 5mm crack widths were measured at the west side of the column. Spalling of the cover concrete reached a height of 0.5 to 1.4 of the column depth. For the specimen subjected to bidirectional loading, L2N6B, concrete spalled at the column base between 1.1 to 1.6 of the column depth at 3 % rotation angle. Vertical cracks were observed during the test reaching 1.6 of the column depth in height. Ten longitudinal rebars out of twelve reinforcement bars buckled at the end of the test. Buckled bars exhibited an "S" shape between 0.2 and 0.8 of the column depth from the base.
The following remarks can be made based on the observed damage of specimens with different scale (small/large) of columns: -The spalled concrete zone increases significantly with an increase in the scale of columns (large scale) as shown in Figure 12. It can clearly be seen that damage is concentrated at the lower part for a small scale column. -Buckling of the longitudinal rebar is more important in case of large scale columns for the same displacement, as illustrated in Figure 13.
It can be concluded that the scale effect also has a significant influence on the seismic performance of columns, especially on the damage pattern.

Confined length of reinforced concrete columns at various axial load levels
Where, h e is the clear height of the column, and D is the largest crosssectional dimension of the column. As given in eqs. (1), (2) and (3), the confined zone is independent of the axial load ratio, which is not consistent with the observed experimental results.
In 1989, S. Watson [22], proposed a confined region that takes into account the axial load intensity given by: A required confined length is proposed using the measured concrete spalling length for 16 specimens as shown in Table 2 for different axial load ratios. The governing equation can be written as:      The observed damage for large-scale specimens indicated that the length of potential plastic hinge to be confined for a column with high axial compression force, 0,6f c 'A g specified by the current Algerian seismic code RPA99/v2003, is underestimated.
The confined region needs to be extended to prevent failure of column outside the code's specified plastic hinge region. As an example, the Algerian seismic code, ACI 315, and Eurocode 8, gave for specimen L2N6B, a confined region equal to 1,07D, 1,0D and 1,5D, respectively, which is less than 1,6D found in the experiment, as shown in Table 3.

Conclusion
Some of the main results, as obtained by the authors during realisation of several experimental programs focusing on sixteen reinforced concrete columns with different axial load intensities, are presented and discussed in this paper. The main conclusions of this research program are summarized as follows: -Damage to large-scale specimens observed during the testing shows that the length of potential plastic hinge to be confined for a column with high axial compression force,, specified by the current Algerian seismic code RPA 99/ v2003, is underestimated. The confined region needs to be extended to prevent column failure outside of the confined region specified in the code. Based on the data obtained in the scope of this experimental program, an equation relating the confined region to the column size and the intensity of the applied axial load is proposed.