Geotechnics and Energy 56th

Richard Jardine

Geotechniekdag 2017, Breda

7th November 2017

The energy conundrum

Global energy demand rising 20% per decade

Imperative to cut greenhouse gas emissions

Geotechnical contributions towards resolving grand engineering challenge

W J M Rankine: 1820-1872 Thermodynamics & geotechnical pioneer

Three main Parts, each with paired topics

I – Maintaining supplies: Offshore oil & gas platform foundations & deepwater landslide risks

II - Climate change: Geotechnical impact & engineering adaptation

III – Renewable energy: Mono & multi-pile offshore windturbine structures for shallow & deeper water

Broader aspects

Transferring offshore research to civil engineering

And three general geotechnical themes Integrating geology, experiments, analysis & field observations

Collaborative research & engagement with Industry, as cited throughout

Practical tools: “As simple as possible, but no simpler” Measurements & predictions for North Sea foundations

BP Magnus 1982 Wind 9 MN

Waves 130 MN Hutton TLP 1984 Norway 186m

Scotland Denmark Magnus: pile strain gauges & accelerometers Hutton TLP: special sensors Groups of 9x2.1m steel tubes driven 82m, hard glacial tills Similar groups, but under tension High resolution static & storm data Dynamic storm loading data Revolutionary floating Hutton Tension Leg Platform Load-cells, ±0.03mm IC gauges, static & storm monitoring Jardine, McIntosh & Hight 1988

Sea bed 0 Depth, m Stiff clay 10 Dense sand

20 Hard clay

30 Dense to v. dense sand 40

Hard Clay 50

Dense to v. 60 dense sand 1.83m OD piles driven through Courtesy Conoco deep glacial clays & sands Predictive methods Conventional API Pile FE, t-z & p-y soil ‘spring’ curves, elastic group factors

Small strain & FE Stress-path tests, local instruments Non-linear stiffness

Shear G/p΄ = f (εs) Bulk K΄/p΄ = g (εvol)

ICFEP; Mod Cam Clay & Mohr- Coulomb models

Installation & Coulomb interface

Non-linear group interaction Smith 1992

Jardine & Potts 1988, 1993 Foundation stiffness: measurements & predictions Conventional: 4 x too soft Axial Hutton case Group load tonnes “IC” predictions accurate 3500 “Small-Strain” ICFEP

Also at Magnus, axial, 3000 rotational & lateral Jardine & Potts 1993 2500 Kenley & Sharp 1993 2000 Ganendra 1994 Observations

1500 ‘IC’ approach used widely onshore 1000

500 Less offshore use.. Conventional

Poor axial capacity reliability 0 CoVs up to 70% 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 Uplift mm Field tests with Imperial College Pile

4.0 Axial load, local σr , τrz & pore pressure, h/R=72 multiple h/R levels Bond, Jardine & Dalton 1991 h, m 3.0 h/R=50 Installation, equalisation & loading Six clay & sand sites in UK & France 2.0 Bond 1989, Lehane 1992, Chow 1997 h/R=27 ‘ICP method’ from field, lab tests & theory; Jardine & Chow 1997, Jardine et al 2005 1.0

h/R=8 Checks: 149 test ICP-05 database Yang et al 2017 Radius R Recent updates: With ZJU, 0 With NGI, UWA & Fugro, Lehane et al 2017 Greatly reduced CoVs, predictive bias eliminated ICP précis

Field SI: CPTu qc profiling & sampling Laboratory: interface shear δ́ , σv́ 0 & clay OCR, Sensitivity St

Base: qb = f (qc)

Shaft: effective stress failure τrz = σ́ rf tan δ́

Sand Clay

Pre-loading σ́ rc = qc f(σv́ 0, h/R*) σ́ rc = σv́ 0 f(OCR, St, h/R*)

2 2 0.5 R* = [Router - Rinner ]

At failure

Compression σ́ rf = σ́ rc + ∆σ́ rd σ́ rf = 0.8 σ́ rc Tension ≈ 25% lower same

Dilation ∆σ́ rd = δr 2G/R δr = pile roughness G = f(qc, σv́ 0)

Courtesy P van Esch, Heerema Applications & checks All Shell platforms from 1996

Instrumented driving & storm response monitoring

Field performance assessed: 13 case histories, Overy 2007

Better reliability & economy, reduced installation risks Critical to marginal projects Overy & Sayer 2007

Sand method:

© API RP2 GEO and ISO Goldeneye Courtesy R Overy, Shell Storm performance?

Shearwater A; piles driven in 1997 Response to ≈20 year storm, December 27th 1998

2.13m OD, 51m: hard clays & dense sand Waves at Tern Single piles push & pull, deck sways; Hunt 1999

Model: API t-z, p-y curves 300 Measured maxima Horizontal sway, mm 200

100

0 0 4 8 12 Significant wave height, m As at Magnus & Hutton, field far stiffer than API model 1+ year after driving & cyclic loading Research questions posed by field experience

Ageing after full pore pressure equalisation? Stiffness & capacity?

Cyclic loading? Impact & assessment for design?

Full stress regime around driven piles? Poorly understood, beyond accurate analysis?

Field: ageing study in dense Dunkirk sand

Steel pipe piles 457mm x 19m Parker, Jardine, Standing & Xavier 1999

Driving is key: Bored piles behave differently! Puech et al 2013

With HSE & PMC

Jardine & Standing 2000, 2012; Jardine, Standing & Chow 2006 Shaft capacity from 1st time tests

Capacities grow markedly over months after driving

Re-tests: more complex & disrupted trends

Tension, kN 3000 235 days

81 days 2000

9 days 1000 Initial stiffness constant

0 0 10 20 30 Pile head displacement, mm Adding tests by Gavin et al 2013 & Karlsrud et al 2014

1st time tension 0.3

3.0 Larvik Meanupdated trend IAC 2.5 Dunkirk

Blessington 2.0 QS(t)/

QS(ICP) (ICP) s 1.5

(t)/Q s Q 1.0

End of driving 0.5 Dunkirk - Jardine et al. (2006) Blessington - Gavin et al. (2013) Larvik - Karlsrud et al. (2014) 0.0 1 10 100 1000 10000 TimeTest afterage, installation: days days Governing processes, necessary conditions? Rimoy, Silva, Jardine, Foray, Yang, Zhu & Tsuha 2015 Model: Grenoble calibration chamber experiments with Pierre Foray

Electric Jack Jardine et al 2009 Load cell 36mm OD ICP;

Fontainebleu NE34 & GA39 sands 0m Wet & dry, Dunkirk ID range

-0.33 m C (46R)

-0.52 m 150 kPa surcharge

B (26R) Constant stress, rigid or -0.79 m ‘active’ radial boundaries

Up to 36 sand A (9R) -1.0m stress cells Zhu et al 2009 Stainless & mild steel piles, -1.50m Driven & jacked

Post-docs: M Emerson, B Zhu, Z Yang, C Tsuha; PhDs: Rimoy 2014, Silva 2015

Model: radial stresses in NE3450 50

0 0

40 4.0 40 2.0 Heavily pre-loaded sand 0.50 8.0 0.50 4.0 30 12 30 6.0 Intense loading under tip 16 0.50 8.0 0.50 radius r & h = 0 20 20 20 10

0.75 0.75 R σ´ ≈ q ≈ 20 MPa R

z c 10 10

/ /

1.0 / h σ´r ≈ qc/3 h 1.5 1.0 2.0 0 14 3.0 8.0 4.0 0 3.0 σ´r/qc contours in % 6.0 4.0 2.0 decay with r/R & ± h/R 1.5 0.75 -10 0.75 -10 0.50 And pre-cycled 0.50

-20 -20 Quite unlike bored piles 0.25 0.25 0.25 -30 -30 0 5 10 15 20 0 5 10 15 20 Jardine, Zhu, Foray & Yang 2013 r / R r / R Pile penetrating Load-off pauses Model: arching round shaft after installation

Adding SST shaft σ′r data Reveals σ′r maxima at 2 < r/R < 4

σ′r/qc % σ′θ/qc % 3.0 3.0 Radial Hoop 2.5 2.5 h/R=5.6 h/R=5.6 2.0 2.0

h/R=16~21

(%)

c

(%)

c q

1.5 h/R=31.1

h/R=16~21 1.5

q

/

/

' rs

s 

 1.0 h/R=31.1 ' 1.0 

0.5 0.5 h/R=40.6 h/R=40.6 0.0 0.0 0 5 10 15 20 0 5 10 15 20 r /R r /R

If arch relaxes through creep σ′r & shaft capacity rise

Benchmarks for numerical analyses Yang et al 2014, Rimoy et al 2015 Dataset includes micro-mechanical observations

Dense fractured ‘crust’ forms around shaft - as in field

From above Side view

0.5 to 1.5mm thickness Laser analysis of breakage

Grows with d50 & h Porosity measurements

Pile edge

10 mm

36 mm

Yang, Jardine, Zhu, Foray & Tsuha 2010 Analysis: Arbitrary Lagrangian-Eulerian FE, with grain crushing

Breakage Base q MPa Axis b contours 0 5 10 15 20 25 ~0-2mm 0 ~2-5mm 0.30 0.20 15 ~5-8mm 0.10 0

Pile

30 S/R 1: high crushing Experiment Theory 2: significant 45 3: moderate z, mm

60

S = tip penetration depth Grain breakage R = pile radius None B = 0, Full B = 1

Zhang, Nguyen & Einav 2014 ‘ALE’ FE predictions of arching

σ′ /q % σ′ /q % r c θ c 6.0 6.0

h/R=3 4.5 4.5 h/R=6

h/R=3

: % :

:%

c

c

3.0 3.0 q

q h/R=6

/

/

' 

' r

  1.5 h/R=9 1.5 h/R=9

0.0 0.0 0 5 10 15 20 0 5 10 15 20 r /R r /R Similar to σ measurements, maxima within 30%

Add shaft abrasion & cycling to improve predictions?

Or tackle with Discrete Element Method?

Zhang, Yang, Nguyen, Jardine & Einav 2014 DEM analysis: Ciantia et al 2017

5 0.4m x 1m ‘sand mass’ - 5x10 crushable d50 =8.5mm grains

Matches qb – penetration curve & arching stresses

σ′r/qc % σ′θ/qc % 3,0 3,0

2,5 2,5

h/R=6.6 2,0 h/R=6.6 2,0

1,5 1,5

1,0 1,0 h/R=17.9 h/R=17.9 h/R=33.6 h/R=33.6 0,5 h/R=40.3 0,5 h/R=40.3 r/R r/R 0,0 0,0 0 5 10 15 20 0 5 10 15 20 Better h/R trends

Analysis converging towards experiments Influence of pile & grain diameters?

Interface shear zone controlled by d50

Mini-ICPs: less capacity growth than field

New field tests: NGI, UCD & Grenoble-IC

Also offshore re-drive checks

Confirm shaft capacity growth over time

Next objective: stress regime around tubular driven piles? 2.13m OD, 38.5m, very dense sand

Borkum Riffgrund, German North Sea

Jardine, Thomsen, Mygind, Liingaard & Thilsted 2015 Parallel axial cyclic research

Field: Dunkirk piles’ global response

Model: local stresses in calibration chamber

Laboratory: cyclic element tests

Analysis: simplified procedure

Practical application: sands & clays

Dunkirk field piles: Unstable, Stable or Metastable global response?

Q Depends critically on:

Qmax N, Qcyclic & Qmean Q cyclic Qmean

Related to tension capacity QT Time 45 Qmin 40 Q /Q = Q /Q = 0.42 cyclic T mean T Drained pore pressures 35 30 Marginally Metastable 25 N=200 example 20 N=150 Little effect until N>50

Uplift, mm Uplift, 15 N=100 N=50 10 5 Failure at Nf = 206 0 0 1 2 3 4 5 6 16% capacity loss Time (hours) Global response: 14 Dunkirk field tests

1.0 Unstable Jardine & Standing 2012 Nf < 100 Qcyclic/QT Rapid degradation 0.8 1 41

12 0.6 Meta-stable

13 100 < Nf < 1000 24 206 Modest damage 0.4 1 27 9 3 >221 over tens of cycles >200 345

0.2 >1000 Stable N > 1000 0.0 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Capacity grows, Qmean/QT stiffness constant Local stress response? Mini-ICP experiments Model: mini-ICP in dry NE34 sand at Dunkirk ID Local interface stress paths

Stable: 1000s cycles, capacity grows Unstable: rapid degradation

No drift in σ´r or displacements under displacement cycling

Intermediate load controlled cases: σ´r drift rates tracked precisely Tsuha, Foray, Jardine, Yang, Silva, & Rimoy 2012; Jardine 2013 Analysis: simple approach from ICP tests

Metastable & Unstable τrz cycles compact sand near shaft, σʹr unloads, paths drift towards interface failure

Locate metastable, stable boundaries? Relate σꞌ drifts to cyclic loads & N? r  Element, model or field pile tests? 'r

Design storm: Rainflow & ‘Equivalent Number of cycles’

Bored piles far more susceptible to cycling Tests must model driven pile installation

Triaxial tests: modelling ‘pile’ paths

Conditioning Pre-stressed ‘OC installation’ stress path: ABD

Creep & ageing at B & D

Pre-cycling at C

Constant volume cycling at D Vary Cyclic Stress Ratio CSR = Δq/ p´

Δq = half peak to trough

Drifts in effective stress Δp´ strains & stiffness?

‘Driven pile’ triaxial tests: pꞌ drifts over 4500 cycles

Seven CSRs: 0.05 ≤ qcyclic/pʹ0 ≤ 0.5

High resolution sensors: consistent strain & stiffness trends

∆pꞌ/pꞌ % NE34 ∆pꞌ/pꞌ0 % Dunkirk 0 Stable CSR <0.30 Stable CSR <0.30

Metastable Metastable

Unstable CSR >0.45 Unstable CSR >0.60

Aghakouchak, Sim & Jardine 2015

HCA - simple shear cycling, Hollow Cylinder Apparatus

Track changes in σ´n DisplacementDisplacement transducerTransducer BelloframBellofram cylinder cylinder matching pile conditions

Ram Ram Measure σ´1 , σ´2, σ´3 ClampClamp SprocketSprocket and / torque & σꞌ1 angle α torquetransmission transmission chain StepperStepper motor motor for for torsiontorsion RotaryRotary tension tension cylinder cylinder

Tie rodTie rod HardinHardin oscillator oscillator

Proximity transducers CamCam Prox. transducers α AcrylicAcrylic chamber chamber wall wall

SpecimenSpecimen σꞌ1

LoadLoad cell cell Outer cell pressure Outer cell and pore water Pwppressure transducers transducers

To foundation 72mm ODTo foundationsample HCA Nishimura 2006 Liu 2015 HCA simple shear: drifts in σꞌn over 4500 cycles

‘Driven pile’ cyclic HCA tests, five CSRs, 0.05 ≤ τcyc /pꞌ0 ≤ 0.45

Basis for predicting pile response

∆σꞌn/σꞌn0 % Dunkirk ∆σꞌn/σꞌn0 % NE34 Stable CSR < 0.20 Stable CSR < 0.20

Metastable Metastable

Unstable > 0.30 Unstable > 0.30

Aghakouchak 2015 Analysis: projecting pile response zone limits

Aghakouchak 2015

Dunkirk field & triaxial predictions NE34 model & HCA predictions

Black curves: pile test results Black curves: pile test results

Metastable zone Metastable zone predicted from predicted from triaxial tests HCA tests

Parallel laboratory work with clays Jardine, Puech & Andersen 2012 Application in cyclic storm assessment

Courtesy BP CourtesyCourtesy Chevron Chevron

Clair & Clair Ridge Extreme Captain: EOR project North Sea; new West of Shetland storms; hard & reconfigured platforms; dense sand, glacial tills: Hampson et al 2017 tills & clays: Argiolas & Jardine 2017

Effective contributions on continental shelf

What about deepwaterCourtesy? Chevron Deepwater

Off the continental shelf

Large offshore landslides

Gulf of Mexico example – Sigsbee Escarpment

Pushing the envelope: 1997-2005, after Evans 2007

Ram Schiehallion Ursa Holstein Hoover Thunder Atlantis Powell Horn Foinaven Diana Horse Mars Marlin Kizomba Girassol Mountain Nakika

0

500 450 500

1000 896 980 988

1225 1300 1311 1350 1500 Depth, m 1463

1646

1844 2000 Landslide risks? 1920 2133

2400 38 Sigsbee escarpment: Gulf of Mexico

Kovacevic, Jardine, Potts, Clukey, Brand & Spikula 2012

North 30° N Houston New Orleans

29° N

Shelf 28° N

27° N

200 km 26° N Sigsbee Escarpment 94° W 92° W 90° W 88° W Basin floor Sigsbee geomorphology

Water depth ≈1500m

Developers’ risk analysis: mechanisms & recurrence?

Controlled by active geology & sedimentation

Spreading, uplifting salt diapirs & thick Quaternary clay layers 40 SEISMIC BASEMAP WITH PROPOSED ATLANTIS LATERALS OIL AND GAS ROUTES Geology: geophysics & deep boreholes MARDI GRAS WEST, GC PHASE XII, 3D SPEC SEISMIC° INLINE 5534, V.E. =2 NORTHWEST SOUTHEAST

≈ 4km

≈ 700m

Jurassic salt

Lower continental Sigsbee Upper continental

Lower Continental shelf Slope EscarpmentSigsbee Upper Continental rise Rise Escarpment

Uplifting & spreading salt: constrains & distorts Quaternary sediments 3D SEISMIC LINE ILLUSTRATING GEOLOGIC STRUCTURE OF THE SIGSBEE ESCARPMENT GEMS Figure 4-4 Ground modelling Maximum Depth borehole below1200 sea depth level, m

1800 Horizon 4

2400 Top of Shale Top of Salt 3000

3600 0 1000 2000 3000 4000 Distance, m

Quaternary Horizons H6 to H0 include high Ip low residual φꞌ clay layers

Routine stability analysis cannot model delayed progressive failure

Or predict historical and future slide recurrence periods Numerical ICFEP ‘bottom-up’ modelling of last 600,000 years

Alternating medium & high Ip layers added sequentially: Modified Cam Clay

Sedimentation on 1o slope towards initial 10o escarpment face

10 slope H3 – 550,000 bp

High Ip layers 250m ~100

H4 – 600,000 bp 1,000m

Parabolic basal uplift applied from H3 onwards

1st principles modelling: slope genesis & failure

Maximum slope ~230 prior to last landslide

Rapid run out failure unloads slope

700m Swelling over 1,000s years sediment

250m 236m slow uplift

1,000m

Sets conditions for future stability Analysis of slope since last slide

Long term swelling, overconsolidation, progressive failure

Coupled, strain softening Mohr Coulomb model

From triaxial & IC ring shear tests

0 High Ip clay Peak φ’peak =18 c’peak=10kPa 0 Residual φ’res =12 c’res= 0

0 0 Less plastic clays & sand: ductile φ’crit=25 & φ’=30

Non-linear permeability & stiffness from lab tests

Predicting future risks

1.3km long failure likely, progressive & delayed mechanism

Rupture surface

Residual shear zones Plastic layer

500m

Recurrence interval: c. 5000 years

Critical to project risk assessment

Field verification? Local geomorphology & sediment dating Part I – Maintaining oil & gas supplies

Contributions to continental shelf & deepwater field developments

What about climate consequences & risks? Intergovernmental Panel Climate Change, IPCC, model predictions

‘Business as usual’: 3-5.5°C mean warming by 2100

Mean +1oC from 1750 non-uniform

Future trends? NOAA

Very likely: less Arctic sea ice & shallow permafrost

Likely: more frequent intense precipitation

Medium confidence: storm track shifts

Increasing flood risks Part II – Addressing climate change

1st Modelling geotechnical response Greatest impact: permafrost

Integrated approach, field verification

New design tools for adaptive engineering

2nd Practical adaptive engineering

Flood defence strengthening on difficult ground Permafrost: depths up to 1km Air Annual mean

Snow & vegetation Temp oC

0 Active UK Mean Top Temp Permafrost Siberian study Pole Seasons

Analogue Zero Amplitude Temp USASRCPTF 2003

Geothermal Continuous Discontinuous Sporadic gradient

Unfrozen

Depth North American permafrost

1oC/decade N Alaskan permafrost warming, Romanovsky 2015

0.5 to 1oC/decade air rise in Yukon

Thermal creeping landslides in sporadic permafrost

‘YT’ landslide, glacio- 350m fluvial soils Little Salmon Lake, Yukon Lyle et al 2014

NW Territories: thaw-slump features, up to 25m deep & 30ha www.nwtgeoscience.ca/project/summary/permafrost-thaw-slumps Predicting fate of Siberian permafrost

Imperial College project for BP

Climate: Global IPCC models & local adjustments Imperial College Physics: Reifen & Toumi 2009

Local design tool: coupled Thermo-Hydro-Mechanical THM modelling tool with UPC Barcelona Nishimura, Gens, Olivella & Jardine 2009

Regional: Depths & rates of change in permafrost state? GIS, Engineering Geology & Thermal FE Nishimura, Martin, Jardine, & Fenton 2009 1000 x 1000km region, Lake Baikal, Siberia Maps, atlases & Google Earth

Digital Elevation Models & ground reconnaissance

Five ground model stereotypes Rolling hills, taiga forest Cambrian to Ordovician sedimentary rocks Permafrost, from 60m thick to absent Three geotechnical profiles

A: Alluvial valleys B: Bare rock hills C: Colluvial slopes

Photo M Smith Photo: M Smith BP

Detailed characterisation: Martin 2009

Geothermal FE Analysis: Nishimura, Martin, Jardine & Fenton 2009

Key ground properties: Top boundary: air temperature Porosity & grading snow cover & surface transfer

Porosity Snow 0 0.1 0.2 0.3 0.4 0.5 0 B 11Zoom1 nodes 2 S now CA Zoom 1 55 elem ents AC 20 0.6 m Ground Ssnownow elem ents

] elements

40 A – Alluvium G round m

[ 100m100 m 10m10 m

B - Bare rock h

t C - Colluvium p e 60 D 1m1 m

80

100 Base boundary: regional flux Analysis

Aim: predict temperature T variations with depth x & time t

∂u(T) ∂ ∂T 1-D heat equation - λ(T) -q=0 ∂t ∂x ∂x

q = heat production/sink u = specific internal energy

∂u(T) ∂ = c ρ 1-φ +c ρ S T φ+c ρ (1-S T )φ T+lρ S T φ ∂t ∂t s s l l l i i l l l c = specific heat ρ = density l = specific latent heat s = solid l = water i = ice

Sl = degree of pore liquid-ice saturation φ = porosity

λ = thermal conductivity

Analysis

Conductivity (λ) combines solid (s), liquid (l) & ice (i)

1-φ Slφ (1-Sl)φ λ(Sl)=λs λl λi

Freezing function: Sl = f (T) depends on porosity & grading:

αρ (1-φ) s -β Sl= T ρ lφ α & β specified for Profiles A, B, C

Batch-processing for FE database, hundreds of cases Thermal maps: point-by-point `speed-dating’ matching

Surface: Air temperature, snow & ‘nt’ transfer factor, slope & aspect Ground Profile: A, B or C Base flux: 0.02 to 0.05 W/m2 Climate Geology Topography

Global AOGCM models Desk, Space Shuttle radar reconnaissance data & satellite Local adjustments Digital Elevation Model Slopes, surface Elevation

Surface Profile A, B or C Base flux

W/m2 Air temp, snow cover, nt factor φ, Sl

Database search Query Return Thermal FE Database of solutions 100s of runs

Regional maps Local analysis Results for one 60x80 km area 3 6170000 2

1 6160000

Permafrost top temperatures 0

g

n i

o Year:h t -0.5 C r

o N 4 19406150000 -1

1940 & 2000 predictions: -1.5

-2 Validated against field survey 6140000 3 -2.5 -3 570000 580000 590000 600000 610000 620000 TTOP [oC] 2 Easting 3 6170000 2059 climate: SRES A2 case 2

1 1 6160000

0

g

n

i h

Melting landscape, CH release t -0.5

4 2000r o N 6150000 0 -1 -1.5

Slope & foundation failures -0.5 -2 6140000 -1 -2.5 -3 -1.5 570000 580000 590000 600000 610000 620000 TTOP [oC] Infrastructure distress Easting

-2 3 6170000 -2.5 2 1 6160000 -3 0

Adaptive design g

n

i h t -0.5

2059r

o N New THM analysis tool 6150000 -1 Validated: Calgary pipeline tests -1.5 -2 Naples metro ground-freezing 6140000 -2.5

-3 570000 580000 590000 600000 610000 620000 TTOP [oC] Easting Climate change – 2nd topic

‘Nuts & bolts’ of practical geotechnical adaptive design

Strengthening flood dikes on weak peat foundations

In The Netherlands

With Deltares, Rijkswaterstaat & HHNK Water Authority

Zwanenburg & Jardine 2015

Uitdam, north of Amsterdam Markermeer Polder, 1.5m below Dike sea level (NAP) 30 km dike fails stability checks

Improve at minimum environmental & economic cost?

Geotechnical controversies: Models for peat? Zwanenburg 2014 Effects of past loading? Failure mechanics? Drained or undrained?

Resolve by field, lab & theoretical studies Google Earth 2014 Uitdam 1.6ha test site Density ρ, Mg/m3 0.8 1.2 1.6 2.0 2.4 2.8 3.2 0 18 boreholes, geological logging -1 Bulk in-situ profiling & lab testing -2 Solid Level -3 m NAP 4.5m -4 Peat 4.5m peat, H2 - H3 Von Post, over clay -5

-6 Clay -7 85% organic, 750 < w < 1200% -8

Yield stress σ’vy, kPa 0 5 10 15 20 25 30 Unit weight ≈ water, low σ´ 0 -1 -2 Thin surface ‘crust’ Level -3 m NAP -4 Mean Oedometer yield: 8 < σꞌ < 14 kPa vy -5 ≈11 kPa

-6 B3 Triaxial & DSS: 5 < su < 10 kPa -7 B16 -8 Load tests to failure

Six instrumented foundations

Range of different geometries Controlled loading to failure

Short & long term, Single & multi-stage tests

Highly compressible; 1.5m settlement in 6 months under 35 kPa FE Analysis supported by lab & field testing

Analysis capturing details of undrained failures

Loaded slab Test 1 Loaded slab Test 2

5 m Mechanisms, % shear mobilisation

Calibrate laboratory & field tests to iterative back-analyses

Develop range of parameter selection routes

Apply in dike remedial works

In-situ su test calibrations

Nkt =CPTu 15.3 NBallBall Penetrometer= 16.5 μ =Field 0.5 Vane Test -1 -1 -1

-2 -2 -2

-3 -3 -3

Level -4 -4 -4 CPTS34 Ball Field m NAP S32 B33 FVT 33 Penetrometerfit fit

-5 fit S32 -5 -5 Vane

levelm][NAP levelm][NAP levelm][NAP

-6 -6 -6

-7 -7 -7

-8 -8 -8 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 s [kPa] s [kPa] s [kPa] Suu kPa Su kPau Su ukPa Test 6 Area: pre-loading

Apply to profile consolidation su gains for multi-stage cases Three calibrated su selection routes for adaptive stability design 1. Factored in-situ tests: CPT, Ball Cone & Field Vane

2. ‘SHANSEP’ su/σ́ v – OCR function 3. Lab: interpretation scheme 20 Field Test 18 1 2 16 3 0.9 Field su = 0.6 σ́ v OCR Operational 14 4 su kPa 6 12 Oedometer mean σꞌ 10 vy

8

6

4

2

0 0 5 10 15 20 25

σꞌv kPa Part II

Addressing climate change

Modelling ground response & developing adaptive design

Simple & complex methods for range of climates, verified in field

Could add many more layers of complexity

Internationally agreed scientific conclusion:

Modelling plus adaptation is not sufficient, greenhouse emissions must fall Part III Supporting renewable energy

Aim: reduce offshore wind costs & enable deeper water projects

Foundations: up to 30% of capital cost, 22% on average

1st Deeper water sites Apply oil & gas research Address ‘problem’ geomaterial: Chalk

2nd Shallower monopile projects Modernise design: PISA Joint Industry Project

With Industry, UK Innovate, Carbon Trust, PISA partners 1st - Multi-pile structures for deeper water

Jacket, tripod & floating support structures

Sand & clay sites ICP from SI stage: East Anglia One Small strain testing & analysis 102 three-leg platforms, Rattley et al 2017

Borkum West II 40 large tripods Borkum West Wikinger New research for ‘Problem’ sites EAOW Chalk: Wikinger, German Baltic 70 four-leg jackets Borkum West II - German N Sea: Merritt et al 2012, Jardine et al 2015

40 tripods, 2.48m OD piles Dense sands & stiff clays 50 year storm cyclic loading

ICP: large cost reductions

≈30m

Driving 2.48m dia. piles Photos courtesy Trianel Chalk sites: special problems for large driven piles

Exposure under glacial till near Wikinger: Rugen, German Baltic

Weak CaCO3 easily damaged by impact & cycling

HighlyBlah blah uncertain driving for Wikinger: Advance trials

Load bearing? Advance offshore tests

From Wikimedia commons Tests on 1.38m OD piles for 70 jacket turbines in chalk: Wikinger

Three ground Barbosa,profiles, Geduhn trios ,of Jardine, 1.37m Schroeder OD piles & Horndriven 2015 at each

Penetration % chalk WK38 16.2m 20 WK43 30.7m 66 WK70 31.0m 78 WK38

Three piles driven at each Dynamic monitoring WK70

11 to 15 weeks’ set-up WK43

At each location Static tension to failure 2km Instrumented restrike Approx. Cyclic test at WK38 72

Test pile installation, October 2014: 17m Euro budget First fully remote stage-loaded large scale seabed tests; 40m water Driving response: WK43, 1.38m OD pile

Blows per 0.25m Instrumented dynamic monitoring:

Till Signal matching analysis: h/R* trends

Penetration Routine Worrying low driving m design resistances? prediction

Re-drives show marked ageing

Chalk Linked to onshore ageing & cyclic: plain & ICP piles

Measured Barbosa et al 2017

Buckley et al 2017 Deploying test frame: December 2014

Remotely controlled static and cyclic tension load tests

Remotely operated seabed load frame

For scale

16.66m WK38 turbine location test, 108 days after driving

ICP analysis: 10.3MN Buckley 2015

Designers’ prediction

Outcome: Wikinger foundations re-designed Chalk: summary

Dynamic, static & cyclic offshore tests Initial design proven conservative: good return on investment

Research in progress

Production pile monitoring

Supporting ageing, cyclic & ICP chalk tests onshore in UK

Integrated with field studies on monopiles

PISA Joint Industry Monopile Project

Wind With Oxford & UC Dublin Nacelle Cut costs, enable deeper water Beam use in sands & clays elements Tower Analysis, laboratory & large field tests Wave

Transition Replace standard p-y methods peice Low L/D: add extra components

Monopile Calibrate: FE & stress path tests

Recognise: cyclic response P-y springs

Byrne, McAdam, Burd, Houlsby, Martin, Zdravković, Taborda, Potts, Jardine, Sideri, Schroeder, Gavin, Doherty, Igoe, Muir Wood, Kallehave & Skov Gretlund 2015 PISA design method: ‘Simple as possible, but no simpler’

Conventional PISA M M H H

z z p p τ

τ p p

S

Mb Only lateral p-y ‘springs’ Four sets of ‘springs’, check by Little scope to capture detailed 28 instrumented driven steel piles soil properties Dunkirk & Cowden test sites 2m diameter piles under cyclic loading at Dunkirk 3-D ICFEP analysis calibrated to advanced lab tests

Dunkirk: dense marine sand Two PhD testing programmes: stress path & HCA experiments

Cowden: Humberside Sandy glacial till with stones & fissures HCA ‘specimen sculpting’ aided by CT scanning

Photos: Liu Research by Ushev & Liu

38 & 100mm D triaxial tests

High-resolution local strains

Mid-height pore pressure

Dual-axis bender elements

72mm OD HCA experiments

Control over σ1, σ2, σ3 & α

Compressibility, non-linear stiffness & shear strength Anisotropy, strain rate & cyclic dependency 3-D ICFEP soil models Calibrated to laboratory & in-situ tests

Cowden: expanded Mod. Cam Clay Dunkirk: Bounding Surface Plasticity σʹ /pʹ Non-linear small 1 Strength variation in deviatoric plane G/p' strain stiffness

Log deviatoric strain, %

J J/Jcs

Effect of Effect of OCR density

p' p'/p'cs Zdravković, Taborda, Potts, Jardine, Sideri, Schroeder, Byrne, McAdam, Burd, Houlsby, Martin, Gavin, Doherty, Igoe, Muir Wood, Kallehave & Skov Gretlund 2015 Modelling 2m OD, 10.5m deep, Cowden piles

2500

Field Test

2000

1500

1000

Routine API/ISO/DNV analysis Horizontal Load, kN Load, Horizontal

500

0 Horizontal displacement Routine analysis: far softer than field

Modelling 2m OD, 10.5m deep, Cowden piles

2500

Field Test

2000

ICFEP Class A prediction Far better match, 1500 Next task – cyclic loading

1000 Routine API/ISO/DNV analysis

Horizontal Load, kN Load, Horizontal

500

0 Equally applicableHorizontal to onshore displacement foundations & caissons

One of our ‘broader themes’; the other three? First, integrated approach

Geology, experiments, analysis & full scale field behaviour

Photo c. 1996

4th Rankine Lecturer, Professor Sir Alec Skempton 1914-2001 Next, broad collaboration Steve Ackerley Andrew Merritt Atkins Amin Aghakouchak Michael Mygind BP Cambridge In-situ Ricardo Argiolas Rory Mortimore Chevron Pedro Barbosa David Nethercot Alan Bolsher Satoshi Nishimura Conoco Phillips Andrew Bond Robert Overy Deltares Róisín Buckley Eric Parker DONG Energy Duncan Parker EPSRC Byron Byrne David Potts Fiona Chow ESG, formerly PMC Mark Randolph Jim Clarke Michael Rattley Fugro Clive Dalton Siya Rimoy Geotechnical Consulting Group, GCG Bethan Davies Way Way Sim Itai Einav Grenoble Tech, Laboratoire 3S-R Carlos Santamarina Trevor Evans Health & Safety Executive, HSE Felix Schroeder Pierre Foray Matias Silva Iberdrola, Scottish Power Renewables Clark Fenton Philip Smith Innovate UK Liana Gasparre Jamie Standing Lankelma UK Antonio Gens David Taborda Joanna Haigh Offshore Wind Consultants, OWC Ralf Toumi David Hight Oxford University Michael Hotze Christian Le Blanc Thilstead Royal Haskoning DHV Rupert Hunt Cristina Tsuha Emil Ushev Shell UK Graham Keefe Stavroula Kontoe Robert Whittle UPC Barcelona Barry Lehane Zhongxuan Yang University College Dublin Tingfa Liu Lidija Zdravkovic University of Western Australia Christopher Martin Bitang Zhu Zhejiang University Ross McAdam Cor Zwanenburg Finally

Fit for purpose practical tools “As simple as possible, but no simpler”

Wide spectrum of complexity considered

Future challenges, one example per main part

I – Stress regime inside & around open driven piles?

II – Creeping thermal landslides?

III – Cyclic loading of monopiles & caissons? Geotechnical progress towards resolving the energy conundrum?

Maintaining safe & efficient offshore oil & gas supplies

Addressing climate impact & developing effective adaptation tools

Tackling problem at source: improving renewable energy economics

Thank you for your patient attention