Piled Raft Foundation

Between the geotechnical design parameters of bearing capacity and settlement, which are independent, if the design just satisfies bearing capacity, settlement will be over satisfied (i.e. settlement less than permissible), and vice versa, if the design satisfies settlement, bearing capacity will be over satisfied (i.e. factor of safety against bearing capacity failure higher than the minimum required value). The latter situation is well known to us in the case of rafts in sand. In either case, the design is not optimum, which is possible only in rare instances when both the requirements are optimally and simultaneously satisfied, i.e., the factor of safety against bearing capacity failure is at the minimum stipulated value, and settlement is just equal to the permissible value. When one requirement is over satisfied, as in a normal case, the resulting geotechnical design is partly over safe, and to that extent, conceptually uneconomical.

The above points need for exploring methods by which the design can be made optimum. For example, in the case of rafts in sand, the question is whether it should be possible for us to design the raft satisfying bearing capacity and look for extraneous means to control settlement which would otherwise be excessive. One such solution available is to use piles in conjunction with the raft, the function of the piles being merely to control settlement. Such a system is called a piled raft.

Fig-1 Piled Raft System
Fig-1 Piled Raft System

However, when such a system is provided, it becomes the combination of a shallow foundation such as the raft and a deep foundation such as the pile, both sharing in the process of load transfer to the soil. Fig-1 attempts to depict this picture. So in theoretical terms, there is a three-way interaction between the raft, the piles and the soil, making it a complex problem for any rigorous analysis, for which the best approach would be numerical analysis such as the ‘finite element method’. For design office use, however, one more often resorts to ‘approximate’ methods.

Example-1 [The Petronas Twin Towers]

Fig-2 Petronas Twin Tower
Fig-2 Petronas Twin Tower

The Twin Petronas Towers in Kuala Lumpur, the capital of Malaysia, when completed in 1998 (at a cost of U.S. $ 1.6 billion), was one of the tallest structure in the world. At 450 m, it was 7 m taller than the Sears (now called Willis) Tower in Chicago, U.S.A., which held the record till then.

The towers are circular in plan. The 395,000 m2 complex has 88 occupied storeys above grade with 5 levels (floors) below grade for parking. Each tower has perimeter columns on a 46 m dia. base with an adjacent 21 m dia. 45-storey bustle. The towers stand 55 m apart and are connected by a bridge at the 41st and 42nd floors (Fig-2). Kuala Lumpur City Centre (KLCC) Bhd. was the developer of the project and the structure is owned by Petronas (for Petroliam Nasional), the national oil and Gas Company of Malaysia.

On the soil side, a depth of 10 to 20 m at the top is water-bearing alluvium underlain by varying thicknesses of residual soils of meta-sedimentary formations, namely siltstone, sandstone, shale and occasionally phyllite (known locally as ‘Kenny Hill formation’), followed by Kuala Lumpur limestone formation which can vary dramatically with regard to surface elevation and solution activity leaving huge cavities. (Rock elevations were found varying by 140 m over a distance of less than 50 m.) The interface is always overlain by erratic ‘slump zones’ where Kenny Hill material has softened and eroded into limestone cavities.

Because of the high slenderness ratio of the structure, the developer and the designer set an ambitious theoretical goal of zero differential settlement, practically limiting it to less than 12.7 mm across the base of the towers. The geologic conditions at the site described above indeed made the job technically very challenging.

Among the different types of foundations considered for the project, the final choice, as dictated by techno-economic considerations, fell on a piled raft consisting of friction piles located in the Kenny Hill formation well above limestone, but with the cavities and slump zones grout-filled, with the pile lengths varied to minimize differential settlement.

An elaborate program of testing was undertaken, which among others included more than 260 pressure meter tests.

At the tower locations, the depth of limestone varied from 80 to 180 m, making it feasible for friction piles in the Kenny Hill above to support tower load of 2680 MN. Adopting a design value of 110 kN/m2 for skin fiction, the final design worked out as 1.3 m dia. piles at 4.7 m spacing extending to a depth of 33 m below a mat (raft) of dia. 53.7 m.

The 3-D finite element analyses gave a differential settlement of 11 mm edge to edge under the tower proper, satisfying the design goal in this respect.

Extensive grouting was undertaken to fill the cavities in the limestone falling within the zone of influence of the towers and to improve the slump zones found immediately above the limestone which were formed by the erosion of Kenny Hill into the cavities and solution channels in the limestone.

Example-2 [Burj Khalifa]

Fig-3 Burj Khalifa
Fig-3 Burj Khalifa

Burj (meaning ‘tower’) Dubai(now called ‘Burj Khalifa’), at a height of over 600 m, is at present the world’s tallest building (Fig.48.3). It is Y-shaped in plan (with three wings at 1200 – see Fig.48.4) and rises to 160 storeys, with a podium at the base which includes 4-6 storey garages.

The tower stands on a piled raft foundation, consisting of a 3.7 m thick raft supported on 1.5 m dia. bored piles extending to a depth of nearly 50 m below the base of the raft.

Hyder Consulting (UK) were the geotechnical consultants for the project. They carried out the foundation design which was independently peer-reviewed by Coffey Geosciences (Australia) under the direction of Professor Harry G. Poulos of the University of Sydney.

For control of differential settlement, optimum performance can be expected to be achieved by the strategic location of a relatively smaller number of piles (Fig-4), rather than using a large number of piles evenly distributed over the raft area, or increasing the raft thickness. The performance of the piled raft can be optimised by selecting suitable locations for the piles below the raft. (This, however, assumes no or limited raft-pile interaction.) In general, the piles should be concentrated in the most heavily loaded areas, while the number of piles can be reduced, or even eliminated, in less heavily loaded areas.

Fig-4 Burj Khalifa (Plan showing piled raft)
Fig-4 Burj Khalifa (Plan showing piled raft)

The Burj Dubai site is characterised by a horizontally stratified subsurface profile, which is complex and highly variable, due to the nature of deposition and the prevalent hot arid climatic conditions. Medium dense to very loose granular silty sands (marine deposits) are underlain by successions of very weak to weak sandstone, interbedded with vey weakly cemented sand, gypsiferous fine-grained sandstone/siltstone and weak to moderately weak conglomerate/calcisiltite. Ground water levels were at 0.0 DMD (Dubai Municipality Datum) which corresponded approximately to 2.5 m below ground level.

The bored piles socketed into weak rock were of 1.5 m dia. and 47.45 m length with the tower raft founded at (-) 7.5 m – DMD. The podium piles were of 0.9 m dia. and 30 m length with the podium raft founded at (-) 4.85 m – DMD. The thickness of the raft was 3.7 m. FE analyses gave maximum loads of the order of 35 MN at the corners of the wings and minimum loads of the order of 12-13 MN within the centre. The minimum centre-to-centre spacing of the piles for the tower was 2.5 times the pile diameter. In all 926 piles were used. The bored piles were constructed using polymer drilling fluid, in place of the more conventional bentonite drilling mud. Settlements predicted by FE analysis were of the order of 70-75 mm in the tower area, reducing drastically to 10-12 mm in the podium area. The final measured settlements were found to be comfortably below the predicted range.

Article written by

Dr. Nainan P. Kurian

One Comment

Add a Comment

Your email address will not be published. Required fields are marked *