ACHILLES Lon Long-term De Deterioration of f Lin Linea ear Infr - - PowerPoint PPT Presentation

▶
achilles
SMART_READER_LITE
LIVE PREVIEW

ACHILLES Lon Long-term De Deterioration of f Lin Linea ear Infr - - PowerPoint PPT Presentation

Dec 14, 2022 38 likes •320 views

ACHILLES Lon Long-term De Deterioration of f Lin Linea ear Infr Infrastructure Monday 04 February 2019 Friends House London, UK Agenda 1500 Welcome, background and introduction Prof Stephanie Glendinning (Newcastle) 1520 Modelling

slide-1
SLIDE 1

ACHILLES

Lon Long-term De Deterioration of f Lin Linea ear Infr Infrastructure

Monday 04 February 2019 Friends’ House London, UK

slide-2
SLIDE 2

ACHILLES – an EPSRC Programme Grant

Agenda

1500 Welcome, background and introduction Prof Stephanie Glendinning (Newcastle) 1520 Modelling of weather-driven deterioration Prof Neil Dixon (Loughborough) 1550 Improving inputs to models Prof David Toll (Durham) 1610 Asset behaviour and performance Prof William Powrie (Southampton) 1625 Forecasting and decision support at network scale Prof Darren Wilkinson (Newcastle) 1640 Discussion and next steps Prof Stephanie Glendinning (Newcastle) 1700 end of meeting

slide-3
SLIDE 3

ACHILLES – an EPSRC Programme Grant

Who we are and how we get here

EPSRC funding support - Stakeholders/project partners increasingly embedded

  • 2004-2009 BIONICS
  • Research facility led by Newcastle with Loughborough, Leeds, Bristol, Durham, Dundee,

Nottingham Trent, BGS…

  • A large group of partners including HA (now HE), NR, MottM, Skanska, CIRIA…
  • Constructed a full scale embankment testing facility that continues to deliver data/insights
  • 2005- 3-year start-up funding for CLIFFS
  • A national network led by Loughborough
  • Connected a large group of people interested in climate impact forecasting for slopes

spanning multiple disciplines

  • 2009-2013 FUTURENET
  • Led by Birmingham with Loughborough, Nottingham, TRL, BGS, HR Wallingford
  • Evaluated the look of the UK transport network in the 2050s
  • 2013-2017 iSMART
  • Led by Newcastle, with Loughborough, Queen’s Belfast, Southampton, BGS, Durham
  • Delivered insights into deterioration of transport infrastructure geotechnical assets
  • PLUS MANY OTHER RELATED PROJECTS
  • 2018- ACHILLES – a programme grant building on this previous work

introduction

slide-4
SLIDE 4

ACHILLES – an EPSRC Programme Grant

Context xt

  • The UK’s transport infrastructure is one of the most heavily used in the world
  • The UK rail network takes 50% more daily traffic than the French network
  • The M25 between junctions 15 and 14 carries 165,000 vehicles per day
  • London Underground: Europe's largest metro subway system but also the oldest
  • Much of the rail network is over 100 years old
  • Not just transport assets

introduction

?

00:30

One incident near Birmingham New Street resulted in a total of 4900 delay minutes along the network for the next 12 hours

After Jaroszweski et al. (2015), Meteorological Applications

slide-5
SLIDE 5

ACHILLES – an EPSRC Programme Grant

Model couples hydrological model SHETRAN with FLAC Models the influence of meteorological parameters and climate (change) on slope behaviour introduction

Evid idence at t th the asset scale le – numerical modelli ling Long-term deterioration modellin ing

  • Following equilibrium, deterioration continues at slower rate
  • Influence of extremes will become more significant as FoS approaches 1
slide-6
SLIDE 6

ACHILLES – an EPSRC Programme Grant

Evid idence of materia ial scale le deterioration

Evidence from both the laboratory and field

  • Soil water retention
  • laboratory investigations
  • from field monitoring
  • Permeability/hydraulic conductivity
  • field investigations
  • permeability functions
  • Strength
  • water content
  • suction
  • Cracking
  • at the micro-scale and effect of freeze-thaw
  • at the macro-scale (cracking)

Introduction

slide-7
SLIDE 7

ACHILLES – an EPSRC Programme Grant

Current approaches to asset management

Network Rail – Earthworks technical strategy introduction

slide-8
SLIDE 8

ACHILLES – an EPSRC Programme Grant

Asset deterioration – ACHILLES programme

Generalised deterioration model for transport earthworks (adapted from Thurlby, 2013).

introduction

slide-9
SLIDE 9

ACHILLES – an EPSRC Programme Grant

Modelli ling approach: Key fin indin ings from iS iSMART

  • There is conclusive evidence for seasonal ratcheting progressive failure mechanisms

in constructed slopes

  • However, it remains challenging to model this seasonal ratcheting mechanism
  • Use of an unsaturated framework is critical
  • Key input parameters are:
  • high permeability near surface layer (measured in the field)
  • Soil water retention curves (SWRC)
  • stiffness distribution
  • strength behaviour
  • cracking
  • Non-local strain minimises mesh dependency
  • Weather represented using two approaches: Pore pressure cycling and coupling with

‘weather generator’ to account for current and future climates

  • Able to produce ‘deterioration’ curves and investigate effects of design
  • Our models can replicate measured pore water pressures in a slope and weather

driven progressive failure – the approach has been validated

Modelling deterioration

slide-10
SLIDE 10

ACHILLES – an EPSRC Programme Grant

The near-surface condition and pore water r pressures

  • It critically controls the rate at which an earthwork responds to weather
  • Near-surface permeability measurements for earthworks sparse in the literature
  • Detailed information from Newbury to compare simulations, including

parametric study to define near-surface layers Modelling deterioration

Model calib libration and valid lidation

  • Developed a methodology to allow the influence of meterological parameters

and climate on a slope to be investigated

  • Model makes use of coupling between SHETRAN and FLAC with Two Phase Flow
  • Modelling approach calibrated using Newbury Cutting and Take and Bolton

Centrifuge tests

slide-11
SLIDE 11

ACHILLES – an EPSRC Programme Grant

Model valid lidation of seasonal ratcheting

  • Kaolin slope (modelled/experimental [Take and Bolton 2011])
  • Magnitude and nature of mid-slope and toe displacements are very good
  • This is great news as progressive failure begins at toe
  • And it is supported by independent observations

Modelling deterioration

Numeric ical l models ls of seasonal ratcheting

  • Demonstrates simplistic, transient factor of safety method for two scenarios;
  • Again, shows significance of wet years on the performance of a slope compared to

gradual deterioration under continued seasonal cycles.

slide-12
SLIDE 12

ACHILLES – an EPSRC Programme Grant

Fast-Track Modellin ing – Cut Slo lopes

  • Newbury Cutting
  • Climate Study
  • Intervention / Maintenance Study
  • Geometric Study
  • Embankment Study

Smethurst and Clarke (N.D.) Newbury Cutting (A34) Modelling deterioration

slide-13
SLIDE 13

ACHILLES – an EPSRC Programme Grant

ACHILLES fast-track modelli ling

  • High permeability near surface
  • Evidence from the field shows that cut slope near surface high

permeability zone develop rapidly

  • Vegetation rooting mechanical contribution to soil strength
  • Prior models account for vegetation root influence on suction

generation and effective stress but NOT root influence on strength

  • Used to investigate effects of different remediation

strategies

  • Slope regrading
  • Toe drainage
  • Shear key at toe
  • Soil nails

Fast-track modelling

slide-14
SLIDE 14

ACHILLES – an EPSRC Programme Grant

Fast track modelli ling work: Future

  • Embankment model assumptions
  • Road
  • Impermeable pavement at crest
  • Assume drains at road edges effectively remove runoff
  • High relative strength and stiffness of fill (replicate high quality

compaction)

  • Minimised spatial heterogeneity
  • Rail
  • Permeable crest - Ballasted
  • None existent drainage
  • Lower relative strength and stiffness of fill (replicate end tipped

construction)

  • Increased heterogeneity

Modelling deterioration

slide-15
SLIDE 15

ACHILLES – an EPSRC Programme Grant

Deterioration in inputs to models

Evidence from the laboratory and field

  • Soil water retention
  • laboratory investigations
  • from field monitoring
  • Permeability/hydraulic conductivity
  • field investigations
  • permeability functions
  • Strength
  • water content
  • suction
  • Cracking
  • at the micro-scale and effect of freeze-thaw
  • at the macro-scale (cracking)

Improving inputs to models

slide-16
SLIDE 16

ACHILLES – an EPSRC Programme Grant

Soil il water retention

Field and lab experiments

  • wetting and drying cycles
  • progressive loss of suction for same water content

Improving inputs to models

Hydraulic ic conductiv ivit ity

Large test programme to characterise hydraulic conductivity

  • in the top 1.5 m several orders of magnitude observed

Unsaturated soil il strength

  • cycling of wetting and drying leads to progressive loss of strength

Soil il cracking and deterioration

Observations: micro (lab: env SEM) vs macro (field scale crack measurement)

  • cracks persist and soil deteriorates
slide-17
SLIDE 17

ACHILLES – an EPSRC Programme Grant

Key messages

  • Rate of change (SWRC and strength) is non-linear - greatest change observed

after primary drying. Subsequent rate of change and magnitude is lower BUT cycling effect is continued through extreme events

  • Macro-scale cracking increasing exposure and influence – it renews and

perpetuates W/D cycle effect – deterioration at nano to macro scale.

  • W/D is a pre-cursor to the initiation of progressive failure - causing the soil at

the near surface of an engineered clay slope to reduce in strength without any change in external load. Implications for slope condition (stability) assessment

  • the need for non-stationarity of soil parameters and ground model, with

changes occurring both seasonally and gradually over time.

  • there is a need for new constitutive soil model(s) that can account for soil

deterioration due to wetting and drying Improving inputs to models

slide-18
SLIDE 18

ACHILLES – an EPSRC Programme Grant

New in investigations of deterioration processes

Material-scale testing, modelling, performance and mitigation

  • Experimental testing at a range of scales

Improving inputs to models 10-7 10-6 10-4 10-5 10-8 10-1 10-2 10-3 10 1 100 1000

Im Imaging Elem lement tes estin ting

Large scale laboratory Field testing

Laboratory and Synchrotron X-ray computed tomography to observe porosity and fracture networks supported by mercury and helium porosimetry Investigate effects of environmental cycles (dry-wet and thermal stress cycles). Triaxial testing and tension testing with suction measurements Larger scale flow-based lysimeter testing Field-based wetting experiments to determine infiltration rates

1.2m m

slide-19
SLIDE 19

ACHILLES – an EPSRC Programme Grant

Asset behaviour and performance

Field site data

  • Long-term monitoring to include deterioration and reaction to extremes
  • Field-scale experiments to study particular phenomena
  • Monitoring campaigns to determine heterogeneity with time and space

Asset behaviour and performance

?

A new site – A flood embankment in clay Newbury - highway cutting in London Clay BIONICS - Test embankment in intermediate plasticity clay

slide-20
SLIDE 20

ACHILLES – an EPSRC Programme Grant 50 100 150 200 250 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 01/01/2003 02/05/2003 01/09/2003 01/01/2004 02/05/2004 31/08/2004 31/12/2004 02/05/2005 01/09/2005 31/12/2005 02/05/2006 01/09/2006 01/01/2007 02/05/2007 01/09/2007 01/01/2008 02/05/2008 31/08/2008 31/12/2008 02/05/2009 01/09/2009 31/12/2009 02/05/2010 01/09/2010 01/01/2011 03/05/2011 01/09/2011 01/01/2012 02/05/2012 01/09/2012 31/12/2012 02/05/2013 01/09/2013 01/01/2014 02/05/2014 01/09/2014 01/01/2015 03/05/2015 01/09/2015 01/01/2016

Water content 0.3 m Water content 0.6 m Water content 0.9 m Water content 1.5 m SMD water balance Neutron Probe A Neutron Probe B Neutron Probe C

Long-duration observations - Newbury

Asset behaviour and performance

Volumetric water content (m3/m3) Soil moisture deficit (mm) 13 years

slide-21
SLIDE 21

ACHILLES – an EPSRC Programme Grant

Forecastin ing and decis ision-support

  • Investigation of ‘real’ route data (M4 & London-Bristol rail line)
  • Consider actual slope geometries and geologies for use in modelling
  • Use (surrogate, statistical) models to generate deterioration curves accounting

for heterogeneity and uncertainty

  • Develop a rational approach to understanding the future behaviour of

significant proportions of a network, and prioritisation of investment decisions forecasting, decision-support

slide-22
SLIDE 22

ACHILLES – an EPSRC Programme Grant

Fie ield ld scale le: asset data for modellin ing

Rail slope geometry (London-Bristol) Total rail slopes: 19,506

  • 7,903 cut slopes

443 cuttings in London Clay

  • 11,603 Embankments

405 embankments constructed from London Clay

[data: Mott MacDonald/NR Lidar data]

forecasting, decision-support Road slope geometry (M4 London-Bristol) Total road slopes: 2,039

  • 505 cut slopes

26 cuttings in London Clay

  • 624 Embankments

24 embankments constructed from London Clay

[data: HAGDMS]

  • Aim is to balance the number of modelling to give useful coverage of slopes on

the network (our approach covers 86 % of the slopes on the network)

slide-23
SLIDE 23

ACHILLES – an EPSRC Programme Grant

Deterioration state 180 years after cutting construction + NR failed slopes and vulnerability

FoS ≥ 1.3 FoS < 1.3 ≥ 1.2 FoS < 1.2 > 1 FoS ≤ 1 Vulnerability descriptor, bounding lines and failed slopes as per Network Rail Earthworks Technical Strategy 2018 Real data more conservative than model

  • Site specific failure drivers – not captured

by model?

  • Lack of high permeability near surface

behaviour and desiccation. This second point to be addressed in ACHILLES

Network Rail il data overlain on model results

slide-24
SLIDE 24

ACHILLES – an EPSRC Programme Grant

Dealing wit ith uncertainty/variability/heterogeneity

  • Material variability
  • Properties vary from slope to slope, and from slice-to-slice within a

slope.

  • Incorporate material spatial heterogeneity within a slice using

spatially correlated Gaussian random fields.

Random field for cohesion (Nuttall, 2013)

forecasting, decision-support

slide-25
SLIDE 25

ACHILLES – an EPSRC Programme Grant

Dealing wit ith uncertainty/variability/heterogeneity

  • Material variability
  • Include parameter variability
  • Latin hyper cube methods – Allow efficient coverage of parameter

space

  • Statistical methods to derive appropriate ranges of parameters and

understand their relative frequency of occurrence

forecasting, decision-support

Lin inking model l outputs to fin inancial im impacts

  • Link models to whole life costs
  • Investigate effect of maintenance vs remediation on costings led by

John Preston (Southampton University – Transport Economist) – RC3

slide-26
SLIDE 26

ACHILLES – an EPSRC Programme Grant

Hie ierarchical Bayesian modelling for r data synthesis

  • Use emulators/surrogates for the numerical codes in order

to efficiently characterise network state

  • Use hierarchical spatial Bayesian modelling to pool

information and correctly propagate uncertainty due to missing data and the future

  • Use Monte Carlo ensembles characterising network

performance to make probabilistic forecasts and compare cost and risk of different intervention strategies

forecasting, decision-support

slide-27
SLIDE 27

ACHILLES – an EPSRC Programme Grant

Conclusions and ongoing work

  • We have considerably advanced the numerical models of climate driven

slope failure and their inputs, including a novel deterministic approach to use UKCP09 data.

  • We have successfully demonstrated the likely mode of deterioration and

failure, and created deterioration curves that reflect these.

  • The time to failure is still not correct, but we are working to correct this.
  • Further work is also continuing to incorporate more extreme weather

events.

  • The model can be used to demonstrate that future climate effects have

an adverse impact on slope stability.

  • We have demonstrated the use of a simplified model in investigate

remediation strategies and are working on coupling this with future climate effects

  • We are now linking the performance curves to investment and design
  • ptions and considerations of uncertainty and heterogeneity

Conclusions

slide-28
SLIDE 28

ACHILLES – an EPSRC Programme Grant

Questions

Fast track modelling

  • What types of interventions should we include?
  • How could this work complement and extend current work being undertaken for

asset owners? Material scale

  • What types of interventions should we include at the material scale?
  • What site investigation/lab test data is available that might help gain additional

insights?

  • How might we use the performance curves to better inform remediation and design

decisions? Asset scale

  • What is the evidence for asset deterioration/loss of performance?
  • Are there particular data sets that we should be using?

Network scale

  • What system/network scale performance data/indicators should we use?
  • How might we use the network performance forecasts to better inform investment

and operational decisions?

Modelling deterioration