In Proceedings of the 8th Young Geotechnical Engineers (YGE) Conference, September 2014. Published South African Institution of Civil Engineering (SAICE).
Technical paper discusses the design and design verification of Screwed-in-Casing Augered Piles (SICAP) for the Mt Edgecombe Interchange at Umhlanga, Durban. In oder to optimism the pile foundation design and verify design assumptions, full-scale pile load tests were undertaken on three test piles using the Osterberg load cell (O-cell). In addition to confirming the ultimate design parameters the results of the load test provided valuable insight to the load settlement performance exhibited by these piles and enabled the design to be adjusted accordingly.
LARGEST TEST FOR AFRICAS MEGA INTERCHANGE
In one of the largest of such tests in South Africa to-date, full-scale pile load tests were undertaken on three test piles using the Osterberg load cell.
Understanding the expected pile behaviour is important as the settlement of foundations is crucial to the stability of the incrementally launched bridges which are sensitive to any differential movements.
ABSTRACT
Screwed-in-Casing Augered Piles (SICAP) are a commonly used foundation type for bridges around the world. This pile type is predominantly favoured on bridges which carry significant loads under strict settlement criteria; SICAP provides higher confidence in the construction methodology and also allows for good quality assurance tests subsequently allowing geotechnical engineers to design for higher shaft stresses and moments being transferred into the piles. Additionally these piles can go to depths exceeding those achievable by many other pile types. Mt Edgecombe I/C will be founded using SICAP of up to 35m length. In order to optimise the pile foundation design and verify design assumptions and expected pile behaviour, full-scale pile load tests were undertaken on three test piles using the Osterberg load cell. In addition to confirming the ultimate design parameters, the results of the load test provided valuable insight to the load settlement performance exhibited by these piles and enabled the design to be adjusted accordingly. Keywords: Pile Testing, Osterberg Loadcell, Settlement, Side Shear, End Bearing
INTRODUCTION
The Mount Edgecombe Interchange is a four level interchange that is being constructed in Umhlanga, Durban, KwaZulu Natal. This will be the third four-level interchange to be constructed in Durban, all within a 30km stretch on the N2, travelling north from the N2/N3 (EB Cloete Interchange) intersection. Mt. Edgecombe Interchange will be the largest four level interchange in South Africa, with five new structures, including two incrementally launched structures of 443m and 947m in length respectively, and three widened structures. The interchange is currently under construction and is scheduled to be completed in May 2016. SMEC SA (Pty) Ltd are the design engineers, whilst CMC di Ravenna is the contractor employed by the South African Road Agency Limited (SANRAL). The geotechnical design was undertaken by SMEC in association with Mr Alan Parrock of ARQ Consulting Engineers.
Aeolian (wind-blown) sands are common to the KwaZulu-Natal north coast around Durban and the site is underlain by Berea Red silty sands to depths of up to 56m. Some more clayey soils associated with a paleo-channel are also encountered in a broad stream and attenuation pond located on the southern quadrant of the site.
Due to the depth to bedrock, the high structural loads and strict settlement criteria, pile foundations were the only feasible solution. The design team was tasked with optimising the foundation design for over 75 bridge piers, whilst adhering to strict settlement criterion. The single pile cap settlements were limited to 20mm and differential settlements to 10mm.
Screwed-in-Casing-Augered-Piles (SICAP) were identified as the most suitable pile foundation for majority of the structures as high shaft stresses could be utilised, and the construction of the pile would not be affected by such factors such as the high water table and collapse of soils within the pile bore. Additionally, the piles easily penetrate and found below a strongly cemented but variable ferricrete layer at 8m and a soft clay layer at about 15m. Finally, there would be a high degree of confidence in the integrity of the pile shaft as the concrete placement would be within a casing and the integrity of the pile shaft could be checked via Cross Hole Sonic Logging (CHSL) tests.
The piles were designed in accordance to the American Petroleum Institute’s Guidelines (API, 1989), a Factor of Safety (FoS) of 1.5 was adopted on side shear and 4.0 on end bearing. This will typically entail a Limit State Design soil resistance factor of 0.4 – 0.5 on the combined ultimate capacity; however, the settlement criterion for the individual piles caps on the incrementally launched bridges need to ensure differential settlement of less than 10mm. In order to optimise the pile foundation design and verify design assumptions and expected pile behaviour, full-scale pile load tests were undertaken on three test piles using the Osterberg load cell. Ultimately the O-cell test had to confirm if these factors of safeties were appropriate.
Three locations were identified to represent the most critical loaded piers on the project in typical Berea Red profiles as well as in the predominantly clayey profile found in the attenuation pond. Figure 1 shows a model of the site. Whilst Figure 2 indicates the position of the test piles TP1 to TP3.
Figure 2 Test pile positions and interchange layout
SUBSURFACE CONDITIONS
Drilling at the site was undertaken in 2011. A borehole was drilled at every pier and abutment position as well as on different positions where MSE walls were proposed. This included 71 positions in total equalling some 3000m of rotary core drilling. In situ tests included Standard Penetration Tests (SPTs), Menard’s Pressure Meter tests and Continuous Surface Wave tests. Laboratory tests included shearbox, triaxial and standard material index tests. Open standpipes were installed in all boreholes.
At TP1 the profile adjacent to the shaft length consisted some 60% silty CLAY and 40% clayey SAND. At TP2 and TP3 the profile consisted 100% silty/clayey SAND (Berea Red sands) adjacent to the shaft length. Based on the subsurface investigation the watertable was found to be at 24m, 21m and 24m depth for TP1, TP2 and TP3 respectively.
TEST PILE AND O-CELL INSTALLATION
The test piles comprised 900mm diameter SICAP piles and were specified to be installed using the same methodology and construction materials as all other project piles and prior to the commencement of any other piles; in this way the test piles would not only verify design assumptions but also intended construction methodologies and conditions.
The pile bore was excavated by screwing in the first length of casing, whilst augering inside the casing. The design was based on the principle that the auger should never advance the casing as the pile’s capacities are based on a coefficient of lateral earth pressure of K=1 (Assuming ϕ = 30̊), entailing the soil mass should remain at rest (K0) before concrete is cast. If the auger is advanced below the casing and the casing is inserted afterwards the coefficient of lateral earth pressure would have changed to an active state. Therefore after concrete placement K would only be around 0.7 to 0.8. Entailing a 20 – 30% reduction in the side shear capacity.
After reaching the required depth, the reinforcement cage with the attached Osterberg (O-cell) assembly was inserted into the pile bores. Concrete was placed in the bore by either a concrete bucket or by creating a ramp up to the top of the oscillator/casing and directly discharging into the bore with a tremmie pipe. The temporary casings were removed during concrete placement.
A loading assembly comprising two 330mm O-cell, with 225mm stroke, was installed in all three of the test piles. The levels of the O-cells are summarised in Table 1. Calculating the exact position for the O-cell position in a friction pile is an iterative process; which involves trying to achieve a balance between the upward capacity, building up only in side shear, and the downward capacity, which is derived from a combination of side shear and end bearing. This is done by deriving an Everett (1991) settlement curve for the predicted side shear settlement and end bearing settlement. If the predicted side shear capacity is almost equal to that of the predicted end bearing the O-cell will be placed as close as possible to the toe, as can be seen for TP1 in Table 1. If the predicted side shear capacity is higher than that of the predicted end bearing, the preliminary estimate for the O-cell depth can be calculated by using the standard pilling formulations together with the predicted Everett load-settlement curves.
The toe deflection was taken to be 10% of the pile diameter, therefore the O-cell with 225mm stroke was utilised (available in 150 and 225mm stroke).
After determining the installation depth, the Everett curves can be reassessed modelling the pile above and below the O-cell as separate piles, whilst trying to match the upward capacity vs Settlement with the sum of upwards and downwards capacity versus settlement curves. Alternatively Fugro recommends as a rule of thumb to place the cell at two thirds (66%) down the shaft length of a friction pile. Table 1 summarises the O-cell and strain gauge positions as well as the pile lengths.
Pile compression and top of pile movement were recorded using telltales monitored by Linear Vibrating Wire Displacement Transducers (LVWDTs) and a Leica automated digital survey level, respectively. Vibrating strain gauges have the advantage over conventional electrical resistance or semi-conductor types mainly in that the sensor output is a frequency rather that a voltage or resistance (Hayes, 2002).
Pressure applied to the O-cell was monitored by a Bourdon pressure gauge and electronic pressure transducers. All the instrumentation and cables where connected through a Geologger to a laptop allowing data to be updated live and stored automatically.
The Quick Load Test method for Individual piles, in accordance with ASTM D1143 were specified but using a smaller load-step procedure according to the manufacturers recommendations.
INTERPRETATION OF STRAIN GAUGE READINGS
In order to determine the side shear developed along the pile shaft, it is necessary to ascertain the load in the pile at each strain gauge. This is quite complicated, as this varies (decreases) away from the O-cell, as load is shed into the surrounding soil. Fellenius (2001) provides such a method which is discussed below and relies on the determination of a composite Young’s Modulus for the pile concrete and reinforcement.
The modulus of concrete will however not be constant through the range of compressive loads during a test. Fellenius (2001) notes that the over large stress range imposed during static testing the difference between initial and final modulus for the pile material can be substantial. This can be overcome by determining the Fellenius tangent modulus from the test data. This can be explained easily by considering a strain gauge placed right above the O-Cell and one 15m up the shaft. For the gauge immediately above the O-cell, the modulus calculated for each increment is not affected by side shear and the tangent modulus is the actual modulus. For the gauge at 15m, the first load increment is substantially reduced by side shear and the actual load change experienced is smaller than the increment of load. The load change will only be the same once the shaft capacity below the strain gauge has been fully mobilised. Therefore at the initial stages of increment the tangent modulus will be much larger (approaching a small strain stiffness) further away from the O-cell than the composite tangent modulus of a strain situated just above the O-cell.
It is a good rule of thumb to place one or two strain gauges right above the O-cell where the strain will not be affected by shaft resistance. Hayes (2002) suggests that although the tangent modulus is a helpful tool it does have some limitations, namely:
- The ultimate side shear values capacities should be reached or exceeded, and
- 50 micro-strain and preferably more than 200micron should be reached during the tests.
Fellenius suggests that once the data reduction is complete one should start by plotting the tangent modulus versus strain for each load increment. This is shown for TP2 in Figure 5 below (the solid line illustrates the ACI derived moduli). The Tangent Moduli will therefore initially be large and as shaft resistance is mobilised, the calculated moduli will become smaller.
Once the tangent modulus has been obtained, the load shed along the pile can be determined as shown in Figure 6. Subsequently, the mobilised net unit side shear for the various depth ranges can also be calculated. These are as summarised in Table 4. It should however be noted that the values reported might not be ultimate values.
From Table 4 one can conclude that the mobilised side shear capacities derived further away from the O-cell with small strains are higher than the values that would have been derived by using a constant Es (ACI derived). Fellenius’ theory thus results in a more optimised design.
LOAD DISPLACEMENT BEHAVIOUR
Besides verifying ultimate load capacity, the O-cell also provides an indication of the load-displacement behaviour of the pile which is critical when one considers the strict settlement required by the bridge designers. The load-displacement behaviour recorded during the test was analysed both in its individual components (i.e. the pile sections above and below the O-cell) and the recombined state. The capacities derived from end bearing and side shear from the O-cell test compared to that predicted by Everett’s formulae are as shown in Figures 7 – 9.
When reviewing Figure 7 the predicted total allowable capacity is 3555kN with 8mm of single pile settlement. From the O-cell test the same settlement can be expected at the allowable capacity. It should however be noted that the ultimate capacity of the pile is much lower than predicted, this would be due to an increase in the expected pile toe settlement from 10% of the pile diameter to some 35% and ultimately a decrease in the ultimate end bearing capacity.
From Figure 8 one can conclude that although the combined settlement behaviour at the predicted allowable capacity (5225kN) corresponds well to that originally predicted the side shear buildup induces larger settlement to this point on the combined curve. The ultimate capacity is however much higher than predicted.
Figure 9 shows that at the allowable capacity the load settlement behaviour is similar to that predicted. The ultimate side shear capacity is however higher than predicted whilst the end bearing ultimate capacity is lower.
DESIGN IMPLICATIONS
Data from two of the tests showed;
- that the ultimate end bearing capacities are lower than the predicted from the API and Everett theory used in design or require settlements equal to 35% of the pile diameter to be mobilised
- the ultimate side shear capacity in two of the piles were larger than predicted
- the combined settlement at the allowable capacities corresponded well to that predicted from Everett’s equation on all three test piles.
Although the ultimate side shear capacities in two piles were higher than predicted, no changes were made to the side shear design assumptions. A number of factors contribute to the calculation of the side shear one of these is for example that the design is based on a “conservatively assessed mean” of SPT values. Schneider (1997) argues that a “cautious estimate” or “conservatively assessed mean” with a 5% chance of a worse overall value would lie about 0.5 standard deviations from the mean. Using an actual mean SPT value would have in all likelihood proved to be a better assumption for these two piles. It was however found that by using the cautious estimate SPT in TP1 the predicted ultimate capacity were almost exactly as predicted. Consequently, no change was made to the assumed FoS on side shear (of 1.5) nor the on the assumption
The FoS on end bearing was however increased from 4 to 8 due to the lower than predicted end-bearing capacities and the significant settlement associated with load build-up in end bearing. This assumption was consequently tested by doing SPTs at the pile-toe/ground interface. These tests indicated a reduction in SPT values in the order of 50%, this corresponds to the reduction in ultimate capacity seen during the testing. The decrease in SPT and end bearing values is considered to be a result of backpressure at the pile toe. Subsequently the piling construction methodology was modified to ensure that pile-hole was excavated and concrete cast on the same day.
CONCLUSIONS
The O-cell testing provided valuable insights on the load-displacement behaviour of project piles and on the verification of design assumptions of the derived capacities. The tests have generally indicated the design to be accurate but highlighted concerns with short piles or piles where there was an over reliance on end bearing. Some of the project piles were subsequently lengthened but this impacted on less than half of the project piles by no more than approximately 10%.
ACKNOWLEDGEMENTS
The pile load testing was funded by SANRAL. The authors thank and acknowledge SANRAL for their support as well as Alan Parrock for his assistance with the design.
REFRENCES
American Petroleum Institute. 1987. Recommended Practice for Planning, Designing and
Constructing Fixed Offshore Platforms. API Recommended Practice 2A (RP 2A) Seventeenth edition, April 1.
ASTM. 1994. Standard Test Method for Piles Under Static Axial Compressive Load. ASTM.
Everett, J.P. 1991. Load transfer functions and pile performance modelling. Proceedings of the Tenth Regional Conference for Africa on Soil Mechanics and Foundation and Foundation Engineering and the Third International Conference on Tropical and Residual Soils. Maseru 23-27 September. pp 229-234
Fellenius,B.H. 2001. From Strain Measurements to Load in an Instrumented Pile. Geotechnical News Magazine, Vol. 19, No. 1, pp 35 – 38
Hayes, J., Simmonds, T. 2002. Interpreting Strain Measurements from Load Tests in Bored Piles.
