A. INTRODUCTION
A large mat foundation constructed to support a 13-story dormitory on Piedmont residual silty soils in Altanta experienced large settlements as reported by Prof. Paul Mayne (1). The dormitory building was constructed on a 104m (341ft) long by 18m (59ft) wide by 1.08m (3.5ft) thick reinforced concrete mat foundation was modelled with 3D finite elements and results in this study closely matched the reported settlements (DeepFND software program). The design was based on a combined total live and dead loading of approximately 200 kPa (4.18 ksf), although the settlements noted herein are based on the actual dead load plus reduced live loading evaluated to be about 150 kPa (3.13 ksf). As excessive settlements were reported in comparison to the settlement calculated from the initial design, an independent geotechnical study took place with on-site measurement of the actual settlement of the building. The model of the mat foundation is recreated with the DeepFND FEM module.
B. MODEL OF DORMITORY FOUNDATION
The DeepFND Finite element method is used for the simulation of the settled foundation. The concrete pilecap is simulated with shell finite elements with thickness equal to 1.08m. Soil borings confirmed the presence of shallow fill and alluvial soils overlying natural residual fine sandy silts (ML) and silty fine sands (SM) with an ambient groundwater table approximately 6 to 7 meters (20 to 23 feet) deep. These materials exhibited N-values in the range of 5 to 25 blows per 0.3 meters, generally increasing with depth, as shown in Figure 5. A thin zone of partially weathered rock (PWR) exists at the site (generally 1 to 3 meters thick) and SPT refusal was generally encountered at depths ranging from 12 m (40 feet) to 24 m (80 feet) below existing grade. Presumably, the SPT refusal reflects the top of unweathered Piedmont rocks. At a nearby Georgia Tech site northwest of this property, rock coring in five borings confirmed the presence of granitic gneiss bedrock with 38 < RQDs < 45 percent [1].
The constructed model in DeepFND is illustrated in figure 3a. A single homogeneous soil layer is used up to a depth of 18m with a Modulus value equal to Es=8500Kpa as proposed in [1]. The strength properties selected as c=5Kpa and φ=30o. Water elevation is located at 2m below the foundation. The mesh discretization generated by the DeepFND software is illustrated in figure 3b.
C. ANALYSIS RESULTS
The analysis results for the mat foundation are illustrated in the figures bellow. The deformed shape and contour of displacement of the FEM model is depicted in figure 4 and a comparison of the simulated settlement contour and measured settlement contour are depicted in figure 5.
Figure 4: Displacement Results in DeepFND
Finally, the displacement to pile cap distance graph is illustrated in figure 6.
D. CONCLUSIONS
The Finite element analysis method implemented in DeepFND is utilized to capture the settlement of a dormitory building in Atlanta. The finite element method is capable of accurately capturing the flexibility of the slab, the slab to soil interface and nonlinear soil behavior near the surface of the foundation.
E. REFERENCES
[1] Paul Mayne, " Unexpected but foreseeable mat settlements on Piedmont residuum.", International Journal of Geoengineering Case Histories, Vol. 1, Issue 1
Book A free web presentation:
