1. Water Statistics Users Group – Spring meeting
Long-life, Low Probability of Failure Assets – Deterioration Modelling and Reliability Assessment
Sue De Rosa – Halcrow
12 May 2011
11/01/2019
11/01/2019
11/01/2019

2. WSUG Spring Meeting 12 May 2011
11/01/2019
11/01/2019

3. UKWIR Project on Long-Life Assets
WSUG – the project in context
Assets covered
Project background and objectives
Approach and deliverables
Key elements in constructing the toolkit 3 3
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

4. WSUG – The project in context
Considers asset reliability
Probability of asset failure
Consequence of failure to asset serviceability
Risks
Modelling methods
Deterioration modelling based upon materials degradation processes
Probabilistic methodologies
Uncertainty modelling
Undertaken jointly by Halcrow and ICS 4 4
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

5. Assets covered Concrete structures (general e.g. treatment plant)
Water Towers
Boreholes
Service reservoirs – general & roof membranes
Tunnels
Bridges
Aqueducts
Large diameter pipelines (e.g. man-entry size)
Large sewers
Tanks – general & cess pits/septic tanks
Embankments
Dams
Hydraulic structures (e.g. Siphon, In/Outlet houses)
Retaining walls
Buildings - operational
Roadways 5 5
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

6. Project Background & Objectives
To develop a toolkit based on science and technology, rather than expert judgement, which will allow water companies to develop robust capital investment plans for maintenance of long-life, low probability of failure civil engineering structures which will satisfy the regulator
Tools to support investment planning
Deterioration models to assess asset life
Auditable and robust
Issues
Limited failure data
Lack of deterioration data
Often high consequence (critical) 6
11/01/2019
11/01/2019
Long Life Assets

7. Toolkit characteristics
Approach
Integration of asset deterioration “know-how” into a risk-based toolkit that informs management of long-life assets
Alignment with the Common Framework – forward looking analysis, asset failure and serviceability forecasts
Deliverables
Knowledge of model options and alternatives
Spreadsheets with deterioration curves, compatible with company systems, applicable to different long life assets
A toolkit to enable robust capital investment plans for long-life, low failure-rate civil assets
Toolkit characteristics
Generic use to shape and focus
Specific use on a particular asset, to decide on actions
Variety of models to cover asset types
User-assessment of their own assets
Use as a decision support tool to restrict subjectivity in the process 7
11/01/2019
11/01/2019
Long Life Assets

8. Key elements in constructing toolkit
FMEA – basis for understanding deterioration
Functional Limit States – triggers for action
Model development – the science behind the curves
Link to Serviceability – risk management process
Planning intervention – commensurate with risk and level of uncertainty
Challenges:
Need for robust deterioration models
Assets with low failure probability
Diverse materials
Relevance to industry regulation – serviceability & risk
Common Framework aligned 8 8
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

9. FMEA
Root Cause Analysis to align asset deterioration mechanisms to failure modes
Analysis included:
Identifying deterioration mechanisms
Identifying factors influencing deterioration
Identifying principal failure modes
Determining the models applicable to the deterioration/failure modes identified 9 9
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

10. Functional Limit States
For the purposes of this project a series of material- and physical condition-based Functional Limit States were defined
Examples:
Chloride concentration of 0.4% at reinforcement
25% loss of structural section of ferrous component
At Limit State:
A point in the deterioration process which represents a significant change in the asset’s vulnerability/resilience
Structure/structural member deemed vulnerable to experiencing a specific failure mode (linked to the functional limit state)
This could potentially lead to a loss of function or service 10
11/01/2019
11/01/2019
Long Life Assets

11. Model Design Models are designed to:
Predict the progression of material degradation for a selection of material/environment combinations
Predict the progression of the asset through various Limit States, e.g.
Superficial surface cracking
Loss of material through cracking/spalling
Leaks and seeps (perforation)
Significant loss of structural section
Predict when Limit States are achieved in asset life 11
11/01/2019
11/01/2019
Long Life Assets

12. Model Types
Material-based models for concrete and ferrous structures – the principal components of the toolkit
Weibull model – provides alternative where materials model not suitable, e.g. masonry structures
Markov chain model – another option where materials models not suitable 12
11/01/2019
11/01/2019
Long Life Assets

13. Materials models Material Deterioration mechanisms
Factors affecting deterioration rate
Concrete
Carbonation
Sulphate attack
Chloride attack
Reinforcement corrosion
Relative humidity
Sulphate content
Chloride content
Steel/cast iron/ferrous
Corrosion
Aggressivity of water/soil
Concrete Models
Environment/ Location
Model Applicability
Carbonation
Chloride
Sulphate
Marine (outside/external) 1
Internal/inside
External/outside (not marine)
Buried
Immersed
Score 0 means model not applicable
Score 1 means model applicable 13
11/01/2019
11/01/2019
Long Life Assets

14. Concrete Models I Focusing on most important deterioration mechanisms
Carbonation
Chloride
Sulphate
Traditional two-stage model typically used to define the modelling process:
Prediction of time to onset of corrosion (Initiation)
Prediction of rate of corrosion to serviceability limit state (Propagation)
Deterministic and probabilistic modelling capability 14
11/01/2019
11/01/2019
Long Life Assets

15. Traditional model for corrosion (Tuutti, 1982)
Concrete Models II
Traditional model for corrosion (Tuutti, 1982)
11/01/2019
Long Life Assets

16. Toolkit – Carbonation model16 16
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

17. Ferrous Model I Deterioration rates according to environment: Buried
Atmospheric
Immersed
Rates taken from Standards/Codes (BS8004, BS6349, Eurocode 3)
Exposure Classification
Mild (L)
Moderate (M)
Severe (H)
Model scenarios: Total of 27
Buried structures containing water
Internal (water)
External Buried H M L HH HM HL MH LM ML LH LL
Immersed structures
Corrosion both sides H M L 17 17
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

18. Ferrous Model II Example: Immersed Structures Immersed structures
Exposure classification H M L
Corrosion rate (mm/y)
0.050
0350
0.020
Corrosion both sides (mm/y)
0.100
0.070
0.040
Example: Immersed Structures 18 18
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

19. Toolkit – Ferrous model19 19
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

20. Link to service
Link between: deterioration – asset failure – service failure
Risk Management Process 20 20
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

21. Deterioration  Asset failure  Service I
Functional Limit States - examples:
LS1 - Superficial surface cracking
LS2 - Loss of material through cracking/spalling
LS3 - Leaks and seeps (perforation)
LS4 - Significant loss of structural section
How vulnerable to failure is the asset at the Limit State?
Depends upon:
Precise Limit State
Loading on asset
Qualitative e.g. 1 to 3 classification (H, M, L)
Deterministic e.g. actual Stress value 21 21
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

22. Deterioration  Asset failure  Service II
How vulnerable to failure is the asset at the Limit State?
Distribution of R ( t )
Distribution of S
Service life distribution
Service period tp R , S
Time
Failure probability
Mean service life
Capacity to resist load
Load 22 22
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

23. Deterioration  Asset failure  Service III
Consequence Modelling
What is the service impact of asset failure -type and extent?
Types:
Leakage
Supply Loss
Water Quality
Flooding
Pollution
Damage to people/property
Influencing factors:
Asset type and function
Population served
Redundancy in system
Location
Estimate extent of asset function loss and length of time to restore service 23 23
11/01/2019
11/01/2019
11/01/2019
Long Life Assets

24. Intervention strategy I
Intervention options include:
Increased monitoring
Minor repairs
Moderate repair/refurbishment
Replacement
Solution option determined primarily by:
Type of asset, its condition and material of construction
Functional limit state reached
Accessibility/repairability of asset
Criticality of the asset (through consideration of consequences of failure) – risk to service
Costs
Uncertainties in knowledge/data
Other broader factors influencing intervention selection
synergies within investment programmes
additional drivers (e.g. Supply/Demand/Quality)
obsolescence
Is action/intervention cost-beneficial? – WLC/CBA 24
11/01/2019
11/01/2019
Long Life Assets

25. Intervention strategy II
Model
Population served Increasing proximity to people
Repair
Replacement
Do minimal
Monitor
Higher load
Decreasing uncertainty
Qualified impact/load
Model
Inspect
Model 25
11/01/2019
11/01/2019
Long Life Assets

26. Long-Life Assets Questions DerosaS@halcrow.com 26 11/01/2019