الفهرس 
Literature ReviewOnly 14 pages are availabe for public view
 |  | from | 301 |  |  |
| | | from | 301 | | |
Abstract
The presented work herein experimentally and numerically investigates the efficiency of four proposed non-monolithic connections between a new concrete flat-slab with an existing concrete column. Furthermore, an analytical study was done by comparing the experimental results to the American (ACI-318, 2019), Euro (EC2, 2014), and Egyptian (ECP 203, 2018) design codes where their subsequent deviations were computed.
The experimental study comprised of two sets of full-scale flat-slab specimens tested under monotonic static loading. The first set consisted of five flat-slabs that were connected with interior columns, while the other set included two specimens connected with edge columns. The first set included one flat-slab connected monolithically with an interior column to act as a control specimen denoted to as IN/COL, while the other four specimens were connected non-monolithically. These proposed non-monolithic connections were achieved using dowels, weld, brackets, to key-lock systems and denoted to as IN/DOW, IN/WLD, IN/BRC, and IN/FRC respectively. The second set included one flat-slab connected monolithically with an edge column to act as a control specimen and denoted as EDG/COL, while the other specimen, which was denoted as EDG/BST, was connected non-monolithically using the best chosen non-monolithic connection system from the test results of the first set as the best was bracket connection system.
The numerical study was performed by using the ABAQUS program to enrich the test observations and results of the experimental work. The numerical study was done in two stages. The first stage was the validation stage where five models from the total fifty-four performed models were carried out to verify the finite element models against the corresponding experimental specimens. The other forty-nine finite element models were carried out to deeply investigate the influencing parameters of the four proposed non-monolithic connections between flat-slabs and columns in the case of flat slabs with interior columns. The parameters that were studied numerically are:
For the dowel system: the dowels number per each direction (n), the diameter of the dowels (d), the dowels embedded length inside slab (L), the height difference between X-direction dowels and Y-direction dowels (h), and the slab thickness (TH) for the proposed non-monolithic dowel connection system.
For the welded system: the length of welded bars (W) and the slab thickness (TH) for the proposed non-monolithic weld connection system.
For the bracket system: the dimensions of the model’s brackets (leg length B and width A), merging the four brackets of the model to work as one column head, merging shear connectors in different numbers to the bracket leg, and the slab’s thickness (TH) for the proposed non-monolithic bracket connection system.
For the key lock system: the value of the normal stress over the slab-column interface (S), the slab-column interface shape with different coefficient of friction value (F), and the slab’s thickness (TH) for the proposed non-monolithic key-lock connection system.
Conclusions
Based on the results of the experimental, analytical, and numerical works in the current study the following points here concluded:
The experimental results indicate the good efficiency of the four proposed non-monolithic connection systems to connect a newly cast reinforced concrete flat slab with an existing column. Moreover, the experimental results showed that the bracket system was the best connection system from the point of ultimate capacity. The good efficiency of the four proposed non-monolithic connection systems resulted due to the following reasons:
Increasing the punching load due to increasing the critical punching plane of the speciemns that connected by bracket and dowel system where the extension of the bracket leg and dowel bars over and inside the slab respectively led to relocate the critical punching shear sections of specimens far away from their columns’ face respectively.
The additional flexural reinforcement used at the specimens that connected by bracket, dowel, and key-lock systems increased the punching shear strength and the energy dissipation.
Using additional normal stress over slab cross-section by post-tensioning the specimen’s slab as in the specimen that connected by key-lock system enhanced the specimen’s stiffness (initial and ultimate), ultimate punching shear strength, and increased the critical punching shear angle (ψ).
Increasing the concrete characteristic compressive strength from 26MPa for the control (monolithic) specimens to 33.7MPa and 34.1MPa for the non-monolithic specimens with interior and edge columns respectively enhanced all non-monolithic specimens’ behaviors in comparison to the monolithic specimens.
Although the good efficiency of the proposed four techniques as aforementioned but they also have some drawbacks as:
All non-monolithic specimens recorded high column slip related to monolithic specimen especially the specimen that connected by key-lock system where the specimen’s column slipped gradually till the ultimate by 3.9mm.
The high slip rates and values of non-monolithic specimens also affect the specimens’ stiffness degradation where the specimens that connected non-monolithically by using key-lock, weld, dowel, and bracket lost about 83%, 75%, 56%, and 53% from their initial stiffness respectively during them loading lifetime while monolithic specimens just lost about 24%.
The comparisons that were done between the experimental capacities and the corresponding estimated capacities from the American code (ACI-318, 2019), Euro code (EC2, 2014), and Egyptian code (ECP 203, 2018) indicate the followings:
All the tested design codes were conservative compared to the experimental results.
The tested design codes had more reliable predictions for the monolithic connections rather than the non-monolithic ones.
The ECP 203 was the most conservative code followed by ACI-318 and finally the EC2 which showed the most reliable results. The difference in the deviations between the used codes is attributed to using the ACI-318 and EC2 some parameters to calculate the expected ultimate punching load more than ECP 203 as the effective normal stress that affects on slab cross-section and the flexure reinforcement ratio in the slab.
Based on the numerical investigation and results the following points were concluded:
validation process
Finite element mesh is one of many parameters that affect the models’ results accuracy where decreasing meshing size makes the model behavior closer to the experimental specimen behavior. Therefore, program trials must be done with different meshing size at the beginning of the modeling process to determine the best meshing way and size.
The concrete damage plasticity model (CDP) showed better results than separately using the elastic damage theory or the elastic-plastic theory.
Parametric study process
from the parametric assessment of the dowel system (DOW), the followings were concluded:
Relocating the critical punching area far away from the column face by increasing the dowels embedded length inside slab (L) and the height difference between X-direction dowels and Y-direction dowels (h) in dowel’s model, enhances the ultimate shear capacity. where increasing (L) till 200mm instead of 100mm caused 33.5% ultimate capacity enhancement while increased by 42.17% through increasing the (H) till the maximum effective depth of the slab.
Increasing horizontal spacings between dowels at the same column’s face by decreasing dowels number (N) and increase dowels’ diameter (D) caused an enhancement in ultimate capacity expressed by (Pu). where decreasing (N) to two dowels instead of three dowels per face and increasing (D) to 32mm instead of 22mm caused 10.70% ultimate capacity enhancement.
Increasing the area that resists the punching shear stress by increasing the slab depth caused an enhancement in ultimate capacity expressed by (Pu). Setting the slab depth to 300mm instead of 200mm caused an increase for the ultimate capacity with 72.09%.
The main parameter that significantly affected the models’ (Ki) was the slab thickness (TH) as increasing the (TH) to 300mm instead of 200mm increased the (Ki) by 184.72%, while the other studied parameters (n, d, L, and h) affected by very small values didn’t exceed 3%.
(Ku) affected by changing the studied parameters’ values where (Ku) increased and reached 197.28% in comparison to the validated model by the increase of (TH) till 300mm instead of 200mm. On the other hand, (Ku) decreased by 91.70%, 68.72%, and 69.46% in comparison to the validated model by increasing (L) to 200mm instead of 100mm, increasing (H) till the maximum slab’s effective depth instead of 32mm, and increasing the spacings between dowels by decreasing (N) to 2 dowels instead 3 dowels and increasing the (D) to 32mm instead of 22mm respectively.
All the studied parameters affected stiffness degradation at ultimate, as increasing (TH) till 300mm instead of 200mm as in the validated model recorded the highest stiffness degradation at ultimate with a value of 32.88%. Increasing the spacings between dowels by decreasing (N) to 2 dowels instead of 3 dowels and increasing the (D) to 32mm instead of 22mm came after increasing (TH) with 33.15% stiffness degradation at the ultimate. Moreover, increasing (H) till the maximum slab’s effective depth instead of 32mm, and increasing (L) to 200mm instead of 100mm caused 33.59% and 56.23% stiffness degradation at ultimate respectively.
The ductility affected by increasing (L) and (H) where the models’ ductility increased by 18.59% by increasing (L) till 200mm instead of 100mm and the ductility enhancement increased till 70.17% by increase the (H) till the maximum effective depth of the slab. As such, increasing the spacings between dowels by decreasing (N) to 2 dowels instead of 3 dowels and increasing the (D) to 32mm instead of 22mm improved the ductility values by 44.89% however increasing (TH) till 300mm (DOW/TH300) instead of 200mm decreased the ductility 11.09%.
from the parametric assessment of the welding system (WLD), it is noticed that changing weld length didn’t affect the ultimate capacity however changing (TH) from 300mm to 200mm significantly affected where:
(Pu) increased by 69.43% due to increasing the TH from 300mm compared with the validated model with 200mm.
The (Ki) and (Ku) increased by 62.67% and 106.77% due to the increase of slab’s inertia and stiffness by increasing slab thickness.
The ductility decreased by 12.08% due to decreasing the reinforcement ratio where the slab thickness increased to 300mm instead of 200mm without an increase in steel reinforcement.
from the parametric assessment of the bracket system (BRC), it is noticed that:
The dimensions of the brackets’ legs in the bracket connection system must be large enough to avoid the local failure under the brackets’ legs by bearing.
The brackets of the bracket connection system must be stiff enough to avoid the yielding of the brackets’ steel by flexural stress before the slab failure whether by flexure or punching.
Relocating the critical punching area far away from the column face by increasing the extension of the brackets’ legs over the slab enhances the ultimate capacity expressed by (Pu) and the subsequent ductility. Extensionning the brackets’ legs to 200mm instead of 150mm over the slab with constant brackets’ width 280mm caused 12% and 5% enhancements for the ultimate capacity and the ductility respectively. On the other side, the stiffness of the tested specimen did not significantly influenced by the bracket extension.
As in the dowel and welding system, increasing the area that resists the punching shear stress by increasing the slab’s depth caused an enhancement in ultimate capacity expressed by (Pu) where increasing the slab depth to 300mm instead of 200mm caused a 75% enhancement in the ultimate capacity. Furthermore, increasing the slab’s depth increased the model stiffness and decreased the ductility where the Ki and Ku increased by 204% and 93% respectively by increasing the slab thickness to 300mm instead of 200m while the ductility decreased 33%.
Merging the four brackets in the bracket connection system to work as a uni-column head did not significantly affect the ultimate capacity expressed by (Pu) of the bracket model. On the other hand, the stiffness whether Ki or Ku significantly increased by merging the brackets where the Ki and Ku reached 173% and 148% respectively in comparison to the bracket model with separate four brackets and that’s lead to a 5% decrease in the model’s ductility.
Merging shear connectors to four connectors per bracket and embedding the shear connectors inside the slab make the brackets work as one unit with the slab and that significantly improves the behavior of the bracket system model. The Pu, Ki, Ku, and ductility of the model after merging four shear connectors to every bracket from the four system’s brackets increased and reached 123%, 350%, 141%, and 172% respectively in comparison to the model without shear connectors.
The enhancements of the model behaviors by miring shear connectors is limited by shear connectors numbers as increasing the merged shear connectors numbers to 6 instead of 4 connectors per bracket didn’t make a significant change in the bracket model behavior.
The shear connectors must be designed to resist the shear flow between the brackets and the slab to avoid the shear failure at the shear connectors that happened in the current thesis by merging just two shear connectors per bracket.
from the parametric assessment of the key-lock system (FRC), the followings were concluded:
The stress value at the slab-column interface (S) can be controlled in three ways:
Changing the external applied stresses value (Se).
Change the dimensions of the externally applied stress areas (Ae) to increase the effective loads.
Changing the distance between the external applied stresses (Se) to change the overlap values between the distributed stresses inside the slab cross-section.
Increasing the stress value at the slab-column interface (S) improved the ultimate capacity expressed by (Pu). The relations between the (Pu) in kN and the (S) in MPa can be determined by the following equations in the condition of a flat slab with 200mm thickness and 33.7MPa characteristic compressive strength (fcu) connected non-monolithically with an interior square column with 300mm length by using a key-lock system as the interface is shaped according to Euro code recommendations for indented surface
Pu=149.22(〖S)〗^2+20.899(S)+1191.8 (kN)
0 MPa ≤S≤ 0.84MPa
Pu=-8.0652(〖S)〗^2+70.396(S)+1255 (kN)
0.84 MPa <S≤ 4.92MPa
The Eurocode recommendation about the indented surface is valid where performing ribs according to the Eurocode recommendations for the indented surface at the slab-column interface of the key-lock connection system’s model can be represented by a flat surface with a coefficient of friction 0.9 and increasing the stress value at the slab-column interface make the two surfaces’ cases identical.
Changing the ribs numbers and dimensions for the indented surface didn’t make a significant change in the behaviors of the key-lock system’s model.
To avoid the early slippage of the column from the slab when the slab-column interface is flat the friction effect must be significantly increased by increasing the coefficient of friction and the stress value at the slab-column interface (S) or using the key-lock system instead of the friction system.
As aforementioned in all proposed non-monolithic connection systems, increasing the area that resists the punching shear stress by increasing the slab’s depth caused an enhancement in ultimate capacity expressed by (Pu) and the models’ stiffness while the ductility significantly decreased by increasing the slab thickness (TH).
Recommendations For Future Works
from the comprehensive experimental, analytical, and numerical work presented in the current study to evaluate the mechanical efficiency of the four proposed non-monolithic connections, it is recommended to extend the current study by further researches on the following points:
Extend the work presented herein to check the validity of the proposed work to non-monolithically corner columns
Validating the effects of the different parameters that were just studied numerically in the current research by using the experimental ways.
Validating the concluded relations between the (Pu) and the (S) in other conditions like using slabs with other characteristic compressive strengths and dimensions. Moreover, changing the column location locations (ie. edge and corner)
Studying the effects of the concrete characteristic compressive strength on the four proposed non-monolithic connections.
Studying the effect of merging shear connectors to the brackets of the bracket system with different shapes than the used shape in the current study as channel connector, angle connectors, block connector with or without anchors, and block connector with hoop.
Studying the effect of creating a small gap around the column (as done by Zhang et al. (2018) for connecting concrete slab by steel column) for the specimens that connected non-monolithically herein by dowel and weld connection system.
Innovating more ways to connect flat slab by column non-monolithically and study the mechanical efficiency of the innovative connections.
Studying the mechanical efficiency of the proposed non-monolithic connection techniques under cyclic load.