Santanu Saha, Haldia, W.B. and Dr. Sudhendu Saha, W.B. highlight the need of a good design of ground improvement with stone columns with a major focus on ground improvement using stone columns and band drains for foundation of large diameter storage tanks.

IMPROVEMENT of strength and compressibility characteristics of soft or weak subsoil, by use of various forms of vertical drains with preloading and/or by installation of stone columns as load bearing elements have been identified as an effective means of ground improvement technique. The field application of technology has developed faster than the design methodology, as the composite behaviour of stone columns and the surrounding soil present a complexity of behaviour, both in terms of load sharing capacities and settlements. The vertical drains including geosynthetic band drains, when associated with pre-loading improves the shear strength and reduces the compressibility of clayey soils by achieving accelerated consolidation under imposed loads. The basic principle involved is that of three dimensional consolidation involving a combination of vertical and radial drainage. The most significant work in this field came from Barron (1948) who incorporated the effects of radial drainage. Later, Hansbo (1979) gave solutions considering effects of smear and well resistance. A method for calculating the degree of consolidation under combined effects of vertical and radial drainage was also presented by Carillo (1942). This paper presents the salient features of the design method adopted for ground improvement for foundations of large crude oil storage tanks. The existing subsoil deposits, the scheme of innovative optimal ground improvement technique executed, and tank performances during hydro testing have been presented.

Subsoil conditions

The subsoil at the site (Haldia), as revealed in soil investigation, comprised compressible clayey soil strata. The upper soil strata of soft silty clay with decayed vegetation extended up to average depth of about 9 m from existing ground level in low- lying area, which was proposed to be developed by about 1.5 m to achieve the finished ground level. The soil strata below 9 m were about 6 m thick non plastic gray clayey silt and fine sands, underlain by about 11 m thick soft silty clay with decayed vegetation. The soil strata below up to about 26 m were stiff to hard silty clay and dense silty sands. The undrained cohesion of subsoils varied from 25 to 45 kN/m2, and coeff. of volume compressibility varied from 2.3 x 10-4 to 3.9 x 10 -4 sq.mIkN corresponding to pressure range of 50 to 100 kN/sq.m. The N-value varied from 2 to 5.

The coefficient of consolidation for pressure range 50 to 100 kN/sq.m varied from 6.70 XlO -4 em- /sec to 11.2 X 10 -4 em / sec.

Design approach

The load bearing capacity of the virgin ground under proposed uniform circular loading below large tank foundations were estimated to be about 80 kN/sq.m, which was far less than the required design bearing capacity of 160 kN/sq.m under proposed construction of large crude oil steel floating roof storage tanks of capacity 60,000 kL, 79 m in diameter and 13.5 m high with total empty weight of 1375 tonne. A number of ground improvement techniques including piles were available which could be used for design of foundations for large oil storage tanks. However, for selection of an appropriate design for techno-commercial assessment in respect of each alternative turn out to be in favour of ground improvement using stone columns, since:

• The length of stone columns would significantly be shorter than piles as it is not necessary to extend the stone columns to a hard stratum ( Bhandari 1998 ).
• Stone columns can withstand large drag forces without getting their load transfer characteristics hampered unlike piles (Madhav 1994).

The beneficial effects of installation of stone columns in weak or difficult subsoil deposits is manifested in the form of increased load carrying capacity and significant reduction in settlements. In similar situations, in recent past, stone columns have been successfully used (Bhandari 1983, Hughes & Withers 1974) for improvement of ground, particularly for storage tank foundations.

Alternatively, vertical drains like sandwicks, band drains, etc, associated with pre-loading, could also be used. Such vertical drains themselves do not share any part of superimposed loads, except providing only drainage paths for accelerating consolidation of the ground under pre-loading. The preloading technique, although quite effective, have major limitation of long time duration together with high cost of pre-load materials, and the environmental hazards associated in its use and disposal, particularly in a running industrial plant areas. For the proposed construction of floating roof crude oil storage tanks of 79 m diameter and 13.5 m high, the total expected average settlement of the virgin ground at centre, at R/2 and at periphery of tanks were approximately estimated to be 950 mm, 900 mm and 465 mm respectively. Such long-term large settlements are not acceptable for the satisfactory performance of storage tanks. As such, the ground improvement scheme had to be so designed to reduce the possibility of excessive settlements and at the same time such reduced consolidation settlements to occur prior to installation of the tanks to operation.

Ground improvement technique

It was observed that subsoils upto average depth of about 9 m was highly compressive with very low bearing capacity. For improvement of load-bearing capacity of the ground, installation of stone columns was considered to be appropriate. Since the load-bearing capacity of stone columns and the treated ground do not depend on the length of stone columns beyond critical length, which is about five or six times the diameter of stone columns, and as the upper compressible strata extended only about 9 m below existing ground level, underlain by fine silty sand layer, the length of stone columns which was considered adequate is only 8 m. The lower soil strata extending up to about 26 m below ground level was also highly compressible, underlain by stiff to hard silty clay and dense sands. As such, treatment of the ground upto at least 26 m was considered essential. This could be achieved only by installation of vertical drains like very cost effective geosynthetic band drains up to depth of 26 m.

The unique combination of stone columns and geosynthetic band drains for improving the ground for foundations of large storage tanks was adopted for the first time in reducing the depth of stone columns. The deep installation of band drains helped in reducing the time of consolidation process of soil under surcharge loads during construction and also during hydro testing of tanks.

Load-bearing capacity of treated ground

After installation of stone columns, and geosynthetic band drains, sand pad foundations were constructed. The steel storage tanks are generally constructed in place over the sand pads.

The hydrotesting of tanks are taken up subsequently. In the present case, about 25 per cent of the design load was actually applied during construction of sand pads and steel tanks. As a result, during this period, the treated ground got partially consolidated under construction activities. Due to rest time after ground treatment, the consolidated undrained cohesion of soil would reasonably be increased to at least 45 kN / sq m. With this value of cohesion, the safe lead bearing capacity of stone columns was estimated as (Saba 1992) P.lFS= Psafe = Ap[(Yz +q,)K+qs (1+2K)/3+ 4C]Net>

where,

Y = submerged unit weight of soil,

z = depth of bulge = 2d

d = finished diameter of stone columns

qs = surcharge on surrounding soil

K = Earth pressure coefficient

A = cross sectional area of stone column ES. = facor of safety

p Net> = tan? (45 + 4>/2)

4> = angle of internal friction of compacted stones

The safe bearing capacity of treated composite ground may be estimated as

Q _, = [(A – A ) q + Psaf ] / A

S where,

A = Influence area of each stone column = 0.868 S2

S = spacing of stone columns in triangular grid

Peripheral concentration of stone columns was provided to prevent any possibility of lateral movement of subsoil and to provide extra stability to edges of sand pad foundations.

Settlement analysis

A typical estimate of total probable consolidation settlement of virgin soil strata, under tank loadings were calculated using the average thickness of different strata , and the average values of coefficients of volume compressibility. The long-term consolidation settlement is normally calculated as follows:

S = A. L H. mv . ilp

where,

A. = factor depends on pore pressure parameter

H = thickness of respective soil strata,

m, = coefficient of volume compressibility of respective soil strata

ilp = increase in effective pressure at mid-depth of respective soil strata.

The total probable consolidation settlement of the ground treated with partially penetrating stone columns up to 8 m below existing ground level, and band drains upto 26 m below GL in the tank pad areas was estimated as S, = &+ilH .where, & = probable settlement of stratum reinforced with stone columns

Settlement criteria

Large steel storage tanks are fairly flexible structures and transmit the weight of the liquid content to the foundation as uniformly distributed load. The bottom plates can easily withstand considerable differential settlement. But the vertical shell because of thinness may be distorted by differential settlement along the periphery, and this may lead to ovality of floating roof tanks. To avoid tension in the bottom plate, the safe permissible change of slope between edge and centre of tank is about 2.23 per cent (Penman 1977). The initial slope of top of sand pad was provided accordingly. To minimise the possibility of shell distortion, the shells are constructed on annular bottom plates which in turn founded on crushed stone ring beam.

In view of above considerations and past experiences, the following settlement criteria were recommended for the 79 m diameter 13.5 m high floating roof storage tanks:

• The average total settlement at periphery during and at end of hydro test shall be limited to 400 mm.

• The differential settlement along tank periphery measured at cleats on shell shall not exceed 1 in 300.

• The maximum differential settlement between diametrically opposite points on the tank periphery shall not exceed 150 mm .The hydrotestings of tanks had already been completed successfully satisfying the above design criteria to demonstrate that the present innovative design concept is very much cost effective and based on sound theory and practice. Since about 25 per cent of load was applied during construction period causing about 25 per cent of expected settlement to occur before actual hydrotesting was taken up. As such the total average settlement that was recorded at tank periphery at the end of hydrotest was only about 200 mm.

Conclusions

Design of ground improvement with stone columns has not been standardised yet. Many authors have attempted various semi-emperical methods of design using stone columns. A few theoretical approaches have also been attempted by the researchers, idealising the soil-stone column system.