DFI Conference, Geotechnical Challenges in Urban Regeneration, London 2010.

PILE FOUNDATION DESIGN PHILOSOPHY AND TESTING PROGRAM FOR A NEW GENERATION DIESEL FUEL PLANT Maria C. Romano, Technip Italy S.p.A., Rome, Italy; Peter Middendorp, Profound BV, Waddinxveen, the Netherlands; Sikko Doornbos, Terracon BV, Werkendam, the Netherlands. The company NESTE-OIL is building a plant for the production of a new generation NExBTL diesel fuel in the Port of Rotterdam The Netherlands. The fuel is produced from renewable vegetable oil as feedstocks and offers 40-60% lower green house gas emissions compared to conventional diesel fuel. Technip Italy acts as the project's main contractor and is responsible for the design of the plant foundations. Pile foundation subcontractor Terracon installed more than 4000 piles. An extensive pile testing program has been performed by Profound before and during piling production. The paper will describe the philosophy behind the foundation design and present the results and comparison of the pile testing program incorporating Static load tests (SLT) and Statnamic load tests (STN).

INTRODUCTION More than 4000 working piles were driven for the foundations of the Neste Oil Rotterdam NExBLT Plant structures, located in Maasvlakte area in the Port of Rotterdam. About 0.7% of the working piles were tested with compression load tests. Before starting the driving of the working piles a pilot piles test campaign of 26 piles has been carried out in November 2008, 12 piles were tested with compression load tests. Main testing campaign scopes were: • To confirm and improve the quality and method of statement for the execution of the piles; e.g. pile driving energy, pile head finishing, safety procedure during construction. • To confirm the geotechnical design capacity of the pile with load tests both in tension and compression. • To choose the compression load test method, Static SLT or Statnamic STN, to be used on working piles. Statnamic Compression Load Test was adopted at the end to test the working piles. This paper presents the pilot piles campaigns relative to the compression load tests comparing the two testing methods, SLT and STN, results.

SOIL CHARACTERIZATION

industrial activities. For the reclamation, sand originating from the Oostvoorne Lake and nearby harbours (Hartelhaven, Mississippihaven, Europahaven and 8th Petroleumhaven) was used. Originally water depth was about 5 to 8 m below mean sea level (NAP), because of the reclamation and landfill, the current surface level is located at approximately +4.00 ÷ 5.50 m above NAP, at the time of the site investigation. The Site will be levelled to a “top of pavement” finished level of +5.20 m NAP. The Process Unit area shall be raised up to + 6.50 m NAP. Geology overview Large parts of the Netherlands today are below sea level and have in the past been covered by the sea or flooded at regular intervals. The modern Netherlands formed as a result of the interplay of the four main rivers (Rhine, Meuse, Schelde and IJssel) and the influence of the North Sea. The Netherlands is mostly composed of deltaic, coastal and eolian derived sediments during the Pleistocene glacial and interglacial periods. Nearly the entire west Netherlands is composed of the Rhine-Meuse river estuary, but human intervention greatly modified the natural processes at work. Most of the western Netherlands is below sea level due to the human process of turning standing bodies of water into usable land, “polder”.

Site informations The Site area was originally located offshore and it was part of the North Sea, but in 1960 – 1970’s it was reclaimed by the Dutch government for

Soil Investigations A comprehensive investigation campaign was carried out to characterize the site soil and

choose the design parameters to be used for geotechnical calculation. The ground investigation activities consisted mainly in the followings: • 400 CPT’s with measurement of cone resistance and local friction up to a depth of 40 meters below surface level. Some of the CPT’s were instrumented with pore pressure measurement (CPTU). • 15 drillings with undisturbed sampling were performed. • In 4 boreholes 8 standpipes were installed and equipped with divers. The groundwater levels were recorded hourly during the soil investigation. Soil Profile and groundwater level Based on the available boring logs and CPT test results, the soil profile at the Site consists mainly of: • Recent sand unit consists of Holocene deposits. It is encountered up to a depth of about -22 m NAP. Fill present at the site derives from material dredged in the Holocene sand deposits and the two have been considered as a single unit for the porpoises of the geotechnical characterization. • Clay to clayey silt lenses of variable thickness (generally modest) within the above deposits. These lenses are found frequently at about +2 m ÷ 0 m NAP, with a maximum registered thickness of about 1 ÷ 1.5 m, and locally at about -2 m ÷ -3 m NAP, with a maximum registered thickness of about 2 m. A thicker layer, of about 1.5 m to 3 m thick, is often encountered beneath -18 ÷ -20 m NAP. Same lenses of peat have been singled out during boring at various depths; • Pleistocene sands: medium to coarse, densely packed, placed generally beneath -21 ÷ -22 m NAP, but locally the top of this layer deepens up to -26 ÷ -28 m NAP. According to available information two groundwater tables can be localized. A perched groundwater table in the top fills, where clay layers at 0 ÷ -1.5 m NAP are present. According to the groundwater level records, tidal fluctuations do not influence significantly the water level in this layer. The main groundwater table is lower than the shallow ones and is clearly affected by tidal fluctuations. The maximum perched groundwater table elevation ranges from a minimum of +0.5 m NAP (about 4.5 m below present g.l.) to a maximum of +3.0 m NAP (about 2 m below present g.l.). On the basis of the recording of groundwater level the mean deep groundwater table ranges between +0.3 m and +0.9 m NAP (about 4 ÷ 5 m below

present g.l.). The maximum elevation due to the sea tidal fluctuations is in the range of +1.25 m to +1.6 m NAP (about 3.4 ÷ 3.75 m below present g.l.). A big role in sole characterization has been played by the CPT results more than boreholes interpretation. This choice was mostly driven by the National Dutch code and practice. The representative values of the major geotechnical parameters for different soil at the Site are summarized in the Table 1.

Parameters

Recent sands and fill

Clayey silt and clay lenses

Pleistocene sands

γ (kN/m )

18.5 ÷ 20.5

15 ÷ 18

19 ÷ 20

Dr (%)

Locally

-

Locally

3

> 80

> 50

45 ÷ 75 %

30 ÷ 40% φ (°)

40

20 ÷ 34

35 ÷ 37

c’ (kPa)

-

20 ÷ 24

-

20 ÷ 50

Cu (kPa)

-

lower bound 50 ÷ 150

-

upper bound

Table 1 Representative values of the major geotechnical parameters for different soil at the Site. Soil profile along the site has been carefully analysed and compared for the choice of the pilot pile position to be tested as representative of the whole site. Some of the soil investigation results of the main pilot piles testing area are presented in the Figures 1. The results are highlighted by the interpolated soil profiles.

PILE DESIGN Design approach for working pile Pile design for working piles has been developed, in accordance with the Dutch standards and regulations (NEN 6740 and NEN 6743, CUR 2001-8), carrying out for each available CPT a detailed calculation of pile capacity under vertical loads, taking into account negative skin friction and the structural capacity of piles. All results have been grouped in a limited number of areas where similar conditions and bearing capacity are predicted. For the given design value of vertical load in compression and in tension for each pile type in each piling area,

the piles lengths necessary to satisfy the design requirements have been evaluated.

driving out the tube with the piling hammer leaving the footplate in the soil as pile tip. The tubes have been driven with an IHC S90. hydraulic hammer.

Figure 2: Picture of the vibro pile driving installation arrangements Pile diameters used in the site are reported in Table 2.

Pile type

pile shaft diameter

pile foot diameter

1

324

365

2

356

400

3

457

510

Figure 1: CPT’s interpolation in the testing area (cone resistance qc in MPa).

Table 2 Adopted piles diameters (driving steel tube walls thickness is 25mm).

In the testing area length and diameter of the piles have been chosen to be representative of the working piles.

STATNAMIC LOAD ANALYSIS

Pile Information The piling system adopted was a driven cast in situ pile (Vibro pile). Figure 2 shows a picture of the vibro pile driving installation arrangement. A driving steel tube, provided with a larger diameter footplate at the bottom, is driven in the soil; reinforcement cage is then placed into the tube; the steel tube is filled with concrete up to working level; the driving steel tube is then extracted by

Unloading Point Method Statnamic testing is a rapid load testing method and it generates a push load on the pile with a duration of 100ms or longer. During this time the whole pile moves as one unit and stress wave phenomena can be neglected. A standard method for the analysis of statnamic signals to obtain the static load displacement behaviour is the Unloading Point Method. (UPM) (Middendorp. 1992). Its validity has been recently evaluated by the Dutch CUR commission H410

and has resulted in a guideline for Rapid Load Testing (Hölscher et all., 2009). The static resistance Fu at maximum displacement uunl (Unloading Point) is calculated with: Fu (tunl)= Funl - M x aunl

(1)

tunl = the time of unloading, which corresponds with the maximum displacement. Funl = load on the pile at time tunl M = pile mass aunl = acceleration of the pile at time of unloading F, u and a are measured values during a Statnamic test and M can be calculated from the pile properties. It has been demonstrated that the UPM provides reliable results for non-cohesive soils like sands or soils that are less sensitive to the loading rate like rock. For sands a reduction of 10% for the mobilized static resistance Fu is applied to compensate for loading rate effects and pore water pressures. A simplified practical approach to compensate for the loading rate effects is to approximate the Fu static load displacement diagram by a hyperbolic curve and followed by reducing the asymptotic value of the hyperbola with a factor β. The value of the factor depends on the loading rate sensitivity of the soil.

To take into account the loading rate effects, the asymptotic value q is calculated as: q = 1/(β.Funl) - 1/(k.uunl)

(5)

The hyperbolic curve is then presented up to the point (Funl, uunl). It is not advised to extrapolate beyond this point, because that part of the curve is not measured and will be fully based on the assumption that the load displacement behavior will remain hyperbolic.

PILOT PILES TESTING PROGRAM Testing field overview Table 3 summaries the compression load tests performed. Piles from PC1 to PC4 were driven in different locations of the site to spot check the designed bearing capacity for each identified geotechnical area. Piles PC5 to PC12 were driven all in the same area, with soil profile representative of the main process area of the plant. 4 piles were tested with STN and 4 with SLT. Each pile tested with STN had a correspondent pile, same length and diameter, tested with SLT.

A hyperbola is defined its initial tangent p and the asymptotic value Fasym= 1/q and can be written as : F = u / (p + q.u)

(2)

with p = 1/k

(3)

In which F is the load, u is the displacement, k the initial spring stiffness from the load displacement diagram, and q the asymptotic value of the hyperbola. With known points Funl and uunl the asymptotic value q can be calculated with: q = 1/Funl - 1/(k.uunl)

(4)

The initial stiffness k is determined from Fu static load displacement diagram. The stiffness is not reduced because in the initial stage of loading, the loading rate is low.

Table 3 Compression load tests performed. The STN results were supplied before the performance of SLT and the STN results can be considered as a class A prediction. Figure 3 shows the detail plan of the main test field. Piles from PT20 to PT26 and from PT13 to PT19 were either tested with tension load tests or

used as reaction piles for performing the SLT. In Static and Statnamic Load test arrangement are shown in Figure 4 and 5 respectively.

Figure 3: Detail Plan of Main test field.

Loading in Compression: Static and Statnamic Static Load Testing has been performed to assure the minimum capacity required by the structural engineer can be mobilised and to assess the stiffness of the pile-soil system when subjected to loading. The test piles Statically and Statnamically Loaded in compression did not exhibit soil mechanical failure. For four pile specifications (soil profile, pile toe level, diameter) both a Statnamic test and a Static test has been performed. Although the tests have not been executed on the exact same pile, conditions are more or less the same and the results can be compared. For all specifications, except for Vibro 457/510 with a toe level of NAP -7.5 m (piles PC6 and PC8), the (long term) static load-displacement behaviour show very high similarity. The Static test on pile PC8 reveals a low initial stiffness with a sudden increase at approximately 600 kN. This is not regarded as normal pile behaviour, although the Statnamic results also hint a stiffening trend. The hyperbole represents an analytical fit of the measurements and therefore will not indicate the stiffening trend. When removing the initial settlement for both tests a high similarity once more appears.

Figure 4: Static Load test.

Comparing to Koppejan limit state 2 design calculations (NEN6743) incorporating a ξ-factor of 1.0 and positive skin friction over the full pile length a slightly lower stiffness is revealed by testing. In Table 4 average stiffness values over the loading interval are presented.

Table 4 Comparison average pile stiffness.

From Table 4 it can be seen that the resulting stiffness retrieved by Statnamic testing deviates not more than 13 % from the average stiffness retrieved by Static testing. For this comparison should be noted that, though pile specifications are equal, testing has been performed on different piles. Figure 5: Statnamic Load test.

Figure 6: Comparison Static Load Testing and Statnamic Vibro 324/365 -7.5m NAP.

Figure 9: Comparison Static Load Testing and Statnamic Vibro 457/510 -14m NAP. From the comparison above it can be concluded that Statnamic testing presents equal reliability to Static testing, for these pile types in this soil and for the range of load tested. It was therefore decided to perform Statnamic testing rather than Static testing considering also the efficiency and accessibility advantages: •

• Figure 7: Comparison Static Load Testing and Statnamic Vibro 457/510 -7.5m NAP.

4 or 5 statnamic tests can be performed in one day; the duration of one statnamic test is around 1 to 2 hours also including considering mounting and dismounting time of the equipments, while it takes one day for a static test. There is not need of driving extra piles to be used as reaction piles for the test. This also means you can do the test on any pile of the piling plan without worrying of space and time to drive the reaction pile.

Based on the maximum registered Statnamic loads for the separate locations and the mobilized (long term) Static Resistance resulting from the interpretation of the Statnamic testing an indication can be obtained for the maximum (long term) Static Resistance which can be mobilized through Statnamic testing assuming a maximum Statnamic load of 4.4 MN. The Unloading Point method (UPM) was applied for the analysis of the Statnamic results (CUR Guideline H410, draft, 2009).

Figure 8: Comparison Static Load Testing and Statnamic Vibro 356/400 -14m NAP.

CONCLUSIONS The adopted design philosophy and included testing campaign resulted in a successful pile testing program and reliable working piles of Neste site. The pile design was conducted according to the Dutch codes; pile testing load was calculated as

maximum design capacity according to NEN code. The Dutch code does foresee the execution of load tests only to check a new pile technology; it is not foreseen in the code a design based on pile load test. Furthermore no load tests are requested by the code to check the design. The tests were executed to comply with Technip quality system. The pile capacities of the Terracon Vibro piles result higher than the anticipated code designed values. Soil mechanical failure could not be reached for static and statnamic load testing. Therefore the engineering of the pile was confirmed. On this project, for these pile types in this soil and for the range of load tested, Statnamic has proven to be a reliable and economic alternative for compression static load testing.

ACKNOWLEDGEMENTS Studio Geotecnico Italiano (SGI) Milan was Technip geotechnical consultant for the design of the piles and the preparation of the test campaign. REFERENCES (NEN 6740 and NEN 6743) Koppejan - NEN 6743 Guideline on the interpretation of Rapid Load Test on piles (Draft CUR Report commission H410), CUR, the Netherlands, 2009 Middendorp, P. Bermingham, B Kuiper, Statnamic load testing of foundation piles. 4th International Conference on Stress Waves, The Hague, Balkema, 1992. Hölscher, P.; van Tol, A.F, Middendorp, P, European standard and guideline for Rapid Load Test, 17th ISSMGE, International Conference on Soil Mechanics & Geotechnical Engineering, Alexandria, 2009