الفهرس Only 14 pages are availabe for public view
 |  | from | 165 |  |  |
| | | from | 165 | | |
Abstract
To better understand the interaction between the soil and the geosynthetic material, two series of laboratory California bearing ratio (CBR) tests had been carried out. Focusing on the relationship between the embedment depth, the surcharge level, and the geogrid stiffness, the improvement mechanism and the optimum depth at which the geosynthetic material perform best have been investigated. The soil had been reinforced with a single geogrid layer. Three types of biaxial geogrid with different stiffness have been embedded at four depths 2, 12.5, 25.4 and 50.8 mm. Three surcharge levels were used in the loading test namely, 0, 22.5, and 45 N. The second series of loading has been conducted with constant vertical stresses at the level of the geogrid to omit the effect of the vertical stresses level on the behavior of the geogrid.
Another series of loading tests was carried out in a 400 x 400 x 400 mm wooden box with a 100 mm diameter loading plate using a hydraulic jack. The geogrid was embedded at depths of 2, 25.4, and 50.8 mm. The surcharge used was 0.0, 0.17, and 0.33 kN to yield vertical stresses of 0, 1.11, and 2.11 kPa respectively. The scale effect as well as the boundary condition has been investigated to check their effect on the results.
A numerical model was developed to simulate the laboratory loading tests using commercially available finite-element based code “Plaxis”. To simulate the reinforcement layer in the numerical model an apparent cohesion factor (C) has to be entered along with an apparent modulus of elasticity factor (E). The values of E and C have been back calculated in a way that the stress-displacement curve obtained from the numerical model matches the experimental stress-displacement curve at the range of the CBR tests (2.54 mm displacement). The relationship between C, E, and both the surcharge level and geogrid type was further studied. The tension force in the geogrid was recorded at displacement of 2.54 mm, at peak and at failure.
The tension force was recorded and integrated along the surface of the reinforcement. The vertical component has been calculated and divided by the loading plate area to get the corresponding vertical stress. The vertical stress developed at the geogrid surface was compared with the total bearing stress improvement, which had been recorded in the experimental tests.
The cohesion and modulus of elasticity factors obtained from the back calculation were used in the numerical model. The reinforcement was modeled at different embedment depths under different surcharge level. The purpose of this model was to validate the cohesion factor and modulus of Elasticity parameters used.
A neural network model was created, trained, validated, and tested to simulate the CBR tests of soil defined with its C, E, and Phi parameters.
The loading tests in the CBR apparatus showed a CBR improvement up to 3.25 times by embedding one layer of geogrid at appropriate depth. That CBR improvement depends on the stiffness of the geogrid, surcharge load, and depth of embedment. This CBR improvement could be equivalent to 59% saving of the material in highway pavement applications. The improvement curve shows a peak at an embedment depth of 0.25 to 0.50 times the loading area diameter depending on the surcharge load. The improvement was found to be directly proportional to the surcharge load and stiffness of the geogrid and was found to be decreasing with increasing the depth of embedment. The loading tests with a small size loading plate showed an improvement ratio up to 1.9, which was less than the similar cases tested in the CBR apparatus by about 10%. This difference might be attributed to the geometry effect and the boundary conditions.
The results showed that membrane effect contribution to the total bearing-capacity improvement is only ranging between 10 to 29%. The most contribution to the bearing-capacity improvement is coming from other mechanisms. Changing the failure mechanism and spreading the load over wider area had the most effect on bearing capacity improvement.