Case C

3.3.1 Project overview

Case C locates at Shih-Lin District of Taipei City. The excavation zone is an irregular polygon with a maximum length and width of 47 m and 18 m, respectively. The project site was originally occupied by a 7-story building that the owners decided to demolish and replace it with a 14-story high-rise building. To fulfill the parking requirement of the new building, a 4-level of basement was needed with an excavation depth of 16.1 m using the bottom-up method in 5 excavation stages.

The soil stratigraphy of the site consists of a 3-m thick surface fill, followed by a thick clay layer, and underlain by andesite debris. The top layer is a 3-m thick fill layer.

The second layer is a thick soft clay deposit with a depth varying from GL. -20 m to GL.

-30 m, and the undrained shear strength ratio is about 0.24. The SPT-N values of this clay deposit increase from 2 at GL.-3 m to about 4 at the bottom elevation. Underlain the soft clay layer is a very dense layer comprises mainly of andesite debris. Depth of the andesite debris varies greatly from GL. -20 m to GL. -30 m. The SPT-N values of the very dense layer are more than 50, and it is regarded as the bearing stratum of the project site. The elevation contour of the andesite is shown in Figure 3.11, and simplified soil profile and soil parameters are presented in Table 3.13 and Table 3.14. The ground water level was located at 0.5 m below the ground surface.

The diaphragm wall was 0.8 m in thickness together with 4 levels of H-steel bracing as the retaining system. It is worthy to mention that the depth of the diaphragm wall must be penetrated into the andesite debris at least 1.5 m. Embedment in the dense andesite debris is necessary as the diaphragm wall is required to provide adequate passive resistance to counter the active earth pressure on the retaining side. The diaphragm wall also serves as an integral part of the foundation system as it carries structural loads through columns embedded in the diaphragm wall. Four levels of horizontal bracing were installed stage by stage in the bottom-up method. The horizontal bracing consists of H-steel beams that were pre-stressed to design values immediately after installation.

Moreover, auxiliary measures in the form of cross wall and buttress wall were adopted to reduce the wall deflection. The 0.8-m thick cross walls were extended from GL. -1.5 m to GL. -24.5 m, as a 0.8-m thick buttress wall was extended from GL. -1.5 m to GL. -23 m. The buttress walls and the cross walls above the excavation surface would be removed by step-by-step excavation. There were six inclinometer casings installed in the perimeter diaphragm wall to monitor the wall deflection for each excavation stage. The layout of 4 cross walls, 1 buttress wall, 6 inclinometer casings and the site plan are shown in Figure 3.12 and Figure 3.13. The structural parameters are listed in Table 3.15 and Table 3.16.

As the existing adjacent buildings are at close proximity, the project owner and contractor were very conservative about the design and construction of the basement. Therefore, the

structural engineers were asked to be cautious on the foundation and excavation design.

The excavation sequence of Case C is shown in Figure 3.14, including the sizes and preloads of the horizontal struts.

3.3.2 PLAXIS simulation

1. Boundary condition

The domain boundaries in x-direction and y-direction should extend approximately seven times of the excavation depth as suggested by Khoiri and Ou (2013). A full model was established as shown in Figure 3.15, including the whole analysis domain and the excavation model. In Case C, the domain size is 350 m by 320 m, which is about eight times of the excavation depth, exceeding the requirements suggested by Khoiri and Ou (2013).

2. Soil parameters

In Case C, the Mohr-Coulomb Model was selected with the input parameters listed in Figure 3.13 and Figure 3.14. An effective Young’s modulus was used for the first and third layer, while the undrained Young’ modulus was used for the second clay layer. The estimation of the Young’s modulus has been described in section 3.1.2.

3. Structural parameters

Table 3.15 and Table 3.16 summarized the input parameters of the diaphragm wall, buttress walls, cross walls and H-steel bracings. Diaphragm wall, buttress wall and cross

wall are regarded as plate elements, and H-steel bracings are regarded as node-to-node anchor elements. The calculation of the Young’s modulus has been described in section 3.1.2.

3.3.3 Comparison of results

In Case C, there are four cross walls installed within the excavation. Two conditions with and without the effect of cross walls are both considered in the numerical analyses.

Figure 3.16 shows all results including the field observation, numerical results with and without cross walls. The numerical results are the wall displacements at the location of the inclinometer casings. The red and blue curves respectively represent the excavations with and without cross walls. The blue curve shows larger wall deflection than the others.

With the cross walls incorporated in the numerical analyses, the diaphragm wall deflection should be refrained pronouncedly. At the locations of SI-1 and SI-4, the numerical results show very small wall movements. It is perhaps that SI-1 and SI-4 are significantly affected by corner effect, the wall deformations tend to be small anyway. It is evident that the cross walls are effective in suppressing the wall deflection by comparing the two numerical results. The amount of reduction is also presented in Table 3.18. The amount of reduction for most locations is close to 80% to 90% except for the locations of SI-1 and SI-4, which is enough to justify the effect of cross walls on limiting the excavation induced wall displacement.

Table 3.18 and Figure 3.17 show the maximum values of all results including field data, two numerical results and the results predicted by Clough’s chart. The predictions by Clough’s chart overestimate the wall displacements compared with the field performance, whereas the numerical results with the cross walls are close to the field data.

Essentially, the numerical results without cross wall should be similar to the estimated values by Clough’s chart. However, there is an obvious difference between these two results. The difference can be attributed to that Clough’s chart ignores the three-dimensional effect of a project site.