Offshore Pile DesignA Brief Project Presentation

April 23, 2015Pagkratios Chitas

offshore geotechnical engineering has formed an increasingly developing concept for the recent decades, as a result of the requirements of the oil and gas industry to move its activities to deeper waters, and thereby, introduce more challenging environments to geotechnical engineers

most of the available foundation design methodologies, are based upon empirical approaches in order to capture the pile behaviour, making the design reliability questionable

any over-prediction of pile’s capacity can be proved risky and lead to significant financial costs for the companies and sometimes to irreparable damage

in response to the above assumptions, the presented report attempts to provide an assessment of the performance of the most relevant pile design techniques, on sites with different soil profiles

Overview

Aim: the assessment of the behaviour and response of offshore driven piles, in normally

consolidated (i.e. Pentre site) and over-consolidated (i.e. Tilbrook site) clayey soils, through the use of both O-Pile software, as well as the findings from available literature, in order to make a comparison and evaluate the pile performance in different soil types

Objectives: to interpret and extract required design parameters from real measured field data, as

they have been derived from in-situ and lab tests for Pentre and Tilbrook sites to insert the parameters from data sampling into the O-Pile software, which will

produce the axial and lateral analysis outputs, through different pile design approaches, so as to assess the performance on the two sites

to bring the data together for presentation, so as to provide a reasonable comparison and commentary on the approach followed

to select, according to the results, the most suitable method for pile design for each site case

Aims and Objectives

Undrained Shear Strength Profile Bulk Unit Weight Profile

Atterberg Limits (Borehole Data)

Relative Void Index at Yield (Oedometer Data)

Unit Skin Friction (CPT Test)

Interface Angle of Friction (Ring Shear Test)

In-situ and Lab Test Data for Pentre Site

Undrained Shear Strength Profile Bulk Unit Weight Profile

Lateral Resistance at 2.3 and 2.8 m Depths (UU Triaxial Test)

Atterberg Limits

Pile Head Load-Displacement

In-situ and Lab Test Data for Tilbrook Site

Additional Design Considerations  Pentre Tilbrook

Pile diameter (m) 0.762 0.762Total pile length (m) 58.5 33.5Pile tip depth (m) 55 31Pile wall thickness 20mm for top 4.5m to withstand driving

stresses, remaining pile length 15mm pile wall thickness

40mm for top 6.5m to withstand driving stresses, remaining pile length 30mm pile wall thickness

Pile unit weight (kN/m3) 78 78Yield stress (MPa) 470 770Poisson’s ratio 0.3 0.3Young’s modulus (GPa) 230 210Comments: Soil parameters for Pentre, will be

considered from 15m below ground surface and under, up to 55m, which is the penetration depth, as the pile used is sleeved for the first 15m, reducing, thus, the skin friction in this zone to a negligible value.

Soil parameters for Tilbrook, will be considered from 1.8m (i.e. due to excavation reasons) to 31m below ground surface, where the penetration reaches.

Pile Design Considerations for both Sites

Pile Design: In general, piles represent deep foundations that gain capacity through the distribution of

loads throughout the soil layers, utilising hence, the interface shearing resistance between the soil mass and the shaft or the tip of the pile.

The axial resistance to both uplift (tension) and bearing (compression) loads, arises from the skin friction and the end bearing pressure at the bottom.

Capacity provided by skin friction is of primary importance when designing in an offshore environment.

Offshore pile design relies mainly on CPT data. The installation occurs with the driving of the pile. During offshore design on clayey soil, undrained conditions need to be considered.

Important Design Aspects

Factors to consider during capacity design: soil features (plasticity, particle-size distribution, stress history, relative density) installation method post-installation effects loading rate loading type (tensile, compressive) pile length pile roughness friction fatigueAdhesion factor (α): Alpha forms an empirical parameter, which is often used during pile design in order to

account for the reduction of skin friction after pile driving process in clays. This occurs due to the decrease of some part of soil’s undrained shear strength, after the lateral vibration of the shaft which is accompanied by pore pressure generation and strain softening that take place during the pile driving procedure.

Factors Affecting Pile Capacity

Main Features of Design Methods

Main Features of Pile Design Methodologies

Method API KOLK NGI ICP F05

(1) skin friction is a function of undrained shear strength based on CPT data CPT-based method CPT-based method CPT-based method

(2) simplistic approach accounts for friction fatigue accounts for soil plasticity

requires lab tests (shear ring and

oedometer)

links skin friction with cone tip resistance

(3) no consideration of soil sensitivity

accounts for stress history (OCR)

reliable for sensitive soils

accounts for stress history

accounts for long-term pile capacity performance

(4) no consideration of friction fatigue

accounts for pile length effects

alpha-based method

accounts for friction fatigue

reliable but overconservative

(5) no consideration of interface friction angle (δ)

no consideration of soil plasticity

designed to address defects of

APIreliable mostly for sand

main disadvantages: measurement

uncertainties and interpretation difficulties

Parameters used:

end bearing resistance (qbf)

undrained shear strength (su)

Skempton’s bearing capacity factor (Nc)

effective overburden pressure (σv0’)

effective unit weight (γ’)

 

end bearing resistance (qbf)

undrained shear strength (su)

effective overburden pressure (σv0’)

effective unit weight (γ’)

pile geometry (D and L)

 

undrained shear strength (su)

effective overburden pressure (σv0’)

plasticity index (Ip)

overconsolidation ratio (OCR)

relative void index (lv0)

relative void index at yield (Δlvy)

interface angle of friction at failure (δ)

 

cone tip resistance (q)

cone skin friction (f) 

Axial Capacity Results

Total Capacity Comparisons for Pentre

Alpha vs Depth for Pentre

Axial Capacity Results

T-Z Curves at 54.75m Depth for Pentre Q-Z or Pile Head Load-Displacement Curves at 0.05m Depth for Pentre

T-Z and Q-Z curves are used during the assessment of the performance of axially loaded piles, to provide axial load transfer analysis in relation to the pile vertical displacement (Z).

T is the mobilised shear stress at the shaft-soil interface. Q is the end bearing resistance of soil beneath the pile tip.

Axial Capacity Results

Total Capacity Comparisons for Tilbrook

Alpha vs Depth for Tilbrook

Axial Capacity Comparisons

Axial Capacity Comparisons for both Sites

Method Skin Friction

(kN)

End Bearing

(kN)

Total Peak Capacity

(kN)

Average Unit Skin Friction

(kPa)

Average α (alpha)

Pentre Site (NC Clay)API 7936 513 8449 68 0.973

APICOOM 4409 513 4922 38 0.557

KOLK 6341 513 6855 54 0.746

NGI 4697 513 5211 40 0.540ICP 6150 435 6585 52 -F05 4235 513 4749 36 -Measured 5170 860 6030 54 0.617

Tilbrook Site (OC Clay)API 13143 1728 14871 186.6 0.352APICOOM 16839 1728 18567 239.2 0.500

KOLK 14084 1728 15812 199.8 0.846

NGI 12529 1728 14257 177.8 0.346

Measured 14681 1450 16131 204.4 0.428

Lateral Capacity Results

P-Y Curve Comparisons at 2.3m Depth Pile Head Load-Displacement Curve for Peak Displacement 0.1m

The P-Y curves displayed above, are used for the lateral analysis for the Tilbrook site, and present a comparison between measured results and the estimated values according to API and Reese methods.

Where ε50 is the axial strain at 50% of the peak deviatoric stress during triaxial undrained test, as recommended by Reese.

The value 0.0025 is taken from the UU tests, whereas 0.005 is generally used in the absence of triaxial test data for stiff clay (Clarke, 1993).

Pentre site (normally consolidated clay): Axial Capacity and Alpha: In the case of Pentre, NGI forms a relatively reliable method to

predict the axial pile capacity in soft clayey soil (as it does account for soil’s sensitivity), as well as the adhesion factor α. In contrary, Fugro method was found to be over-conservative in this case and wouldn’t be recommended as it would result in overdesigning and unnecessary high costs, whereas the API approach was found to overestimate the axial capacities.

T-Z and Q-Z Curves: With regards to the T-Z curves (i.e. the skin friction against displacement), the methods that capture best the behaviour are KOLK and ICP, whereas for the Q-Z curves, the most reliable results were derived from NGI method.

Result Interpretations for Pentre

Tilbrook site (over-consolidated clay): Axial Capacity and Alpha: In the case of Tilbrook, KOLK is found to offer relatively reliable

results for the total capacity modelling (as it accounts for the stress history of the soil and is reliable for clays), while the API would be recommended for the estimation of the adhesion factor α, as KOLK overestimates it. On the other hand, NGI was found to be over-conservative, thus it wouldn’t be recommended, and APICOOM can be regarded as the least reliable method during axial capacity estimations.

P-Y Curves: For lateral pile design, interpreting both the results from the P-Y and the pile head load-displacement graphs, it can be assumed that API method is able to model the lateral response of the clay in the most approximate way. However, for more conservative purposes, average values between API and Reese results could be considered.

Result Interpretations for Tilbrook

Assumptions: Predicting offshore pile capacity forms a complex procedure, and therefore, more than one

methods should be considered during the design process, to enable comparisons between the various obtained results, while different factors should be taken into account depending on soil type and condition, in order to provide more accuracy and reliability.

The results are found to be in a good alignment with the findings of related literature review. During pile capacity estimation, skin friction should be regarded as more important than end

bearing resistance, due to the fact that higher loads occur at the top layers, while soil’s undrained shear strength there is lower than the respective value at the bottom.

It should be born in mind that for soft clays, sensitivity factors need to be considered during design, whereas over-consolidated stiff clays offer higher capacity but stress history is important.

Conclusions