Pipeline Risk Management Modeling of Corrosion NACE ECDA Seminar Jan 2009
What can a risk assessment do? where will the next failure happen? when will a failure occur? how many failures next year? where are the hot spots? what is the best use of my resources? what are the priorities?
Key Concepts Risk = (hazard likelihood) X (hazard consequence) Probability = Degree of Belief Risk Assessment -- Risk Management Management = choices in resource allocation
Historical (Informal) Risk Mgmt ADVANTAGES: simple/intuitive consensus is often sought utilizes experience and engr judgment successful
Historical (Informal) Risk Mgmt REASONS TO CHANGE: more at stake from mistakes inefficiencies/subjectivities lack of consistency need to consider complicated factors regulatory mandates
Gas IM Rule Objectives Prioritize pipeline segments Evaluate benefits of mitigation Determine most effective mitigation Evaluate effect of inspection intervals Assess the use of alternative assessment Allocate resources more effectively ASME B31.8S, Section 5
Gas IM Rule Points to consider Account for relevant attributes Use conservative defaults for unknown data Identify significant risk-driving factors Sufficient segment discretization or resolution Predictive or what-if capability Updateable to reflect changes or new information Populating risk model is resource intensive Validate model, show to be plausible with respect to known history and significance of threats ASME B31.8S, Section 5
Threat Categories ASME B31.8 Supplement considers 3 categories of threat: Time Dependent May worsen over time; require periodic reassessment Time Stable Does not worsen over time; onetime assessment is sufficient (unless conditions of operation change) Time Independent Occurs randomly; best addressed by prevention
Bathtub Curve Failures Time
Threat Categories Time Dependent Threats External corrosion Internal corrosion Stress-corrosion cracking (SCC) (Fatigue)
Threat Categories Time Independent (Random) Threats Third-party/Mechanical damage Immediate failure Delayed failure (previously damaged) Vandalism Incorrect operations Weather related Cold weather Lightning Heavy rain, flood Earth movement
Threat Categories Time Stable Threats (resistance) Manufacturing-related flaws in Pipe body Pipe seam Welding / Fabrication-caused flaws in Girth welds Fabrication welds Wrinkled / buckled bend Threads / couplings Defects present in equipment Gaskets, O-rings Control / relief devices Seals, packing Other equipment
ASME B31.8s Subject Matter Experts Relative Assessments Scenario Assessments Probabilistic Assessments
Better Way to Conceptualize Types of Models Absolute Results Relative Results Tools for All Models Probabilistic methods Scenarios Trees SME (input and validation)
Relative, Index, Scoring Models intuitive comprehensive ease of setup and use optimum for prioritization mainstream served us well in the past
Scoring Model Issues Difficult to anchor Potential for masking Technical compromises weightings scale direction interactions of variables (dep vs indep) Validation (reg reqmt) New Uses
Index Sum vs. Fail Probability Index Score Scenario 1 Scenario 2 Probability Index of Failure Score Score Third Party Damage 60 90 Corrosion 70 10 Design 80 90 Operations 70 90 280 280 Probability of Failure Score
Index Sum vs. Fail Prob Index Score Scenario 1 Scenario 2 Probability Index of Failure Score Score Probability of Failure Score Third Party Damage 60 0.4 90 0.1 Corrosion 70 0.3 10 0.9 Design 80 0.2 90 0.1 Operations 70 0.3 90 0.1 280 76.5% 280 92.7%
Now plan for centerpiece of public scrutiny plan for legal challenges support integrity verification schedule determine appropriate reaction to risk anticipate desire for anchoring the numbers few computer limitations
Enhanced Modeling Approach
Probability of Failure Exposure Mitigation Resistance
Definitions Exposure: liklihood of an active failure mechanism reaching the pipe when no mitigation applied Mitigation measure: prevents or reduces likelihood or intensity of the exposure reaching the pipe Resistance: ability to resist failure given presence of exposure/threat
Information Use--Exposure, Mitigation, or Resistance? pipe wall thickness air patrol frequency soil resistivity coating type CP P-S voltage reading date of pipe manufacture operating procedures nearby traffic type and volume nearby AC power lines (2) ILI date and type pressure test psig maintenance pigging surge relief valve casing pipe flowrate depth cover training SMYS one-call system type SCADA geotech study pipe wall lamination wrinkle bend
Absolute Risk Values Frequency of consequence Temporally Spatially Incidents per mile-year fatalities per mile-year dollars per km-decade conseq prob
Dependent vs Independent Interactions AND Gates CP failure AND coating failure = failure of mitigation CP effectiveness = P/S reading AND P/S distance AND P/S age OR gates PoF = PoF1 OR PoF2 OR PoF3 Corr control = coating effectiveness OR CP effectiveness
Combination of Likely Events 0.8 AND 0.8 AND 0.8 AND 0.8 AND 0.8 = 0.8 x 0.8 x 0.8 x 0.8 x 0.8 = 0.3 actually unlikely 0.8 OR 0.8 OR 0.8 = [1-(1-0.8) x (1-0.8) x (1-0.8)] = 0.992 very likely
Failure Probabilities Overall Pf is Prob Failure by [(Thd Pty) OR (Corr) OR (GeoHaz) ] Ps = 1 - Pf Overall Ps is Prob Surviving [(Thd Pty) AND (Corr) AND (GeoHaz).] So Pf overall = 1-[(1-Pf thdpty ) x (1-Pf corr ) x (1-Pf geohaz ) x (1-Pf incops )]
Final PoF PoF overall = PoF thdpty + PoF TTF + PoF theftsab + PoF incops + PoF geohazard PoF overall = 1-[(1-PoF thdpty ) x (1-PoF TTF ) x (1-PoF theftsab ) x (1-PoF incops ) x (1-PoF geohazard )] Guess PoF if 1%, 4%, 2%, 2%, 0% Calc:
Probability of Failure Exposure Mitigation Resistance
Estimating Threat Exposure Events per mile-year for time independent / random mechanism third party incorrect operations weather & land movements equipment failures MPY for degradation mechanisms ext corr int corr SCC / fatigue
Failure Rates Failures/yr Years to Fail Approximate Rule Thumb 1,000,000 0.000001 Continuous failures 100,000 0.00001 fails ~10 times per hour 10,000 0.0001 fails ~1 times per hour 1,000 0.001 fails ~3 times per day 100 0.01 fails ~2 times per week 10 0.1 fails ~1 times per month 1 1 fails ~1 times per year 0.1 10 fails ~1 per 10 years 0.01 100 fails ~1 per 100 years 0.001 1,000 fails ~1 per 1000 years 0.0001 10,000 fails ~1 per 10,000 years 0.00001 100,000 fails ~1 per 100,000 years 0.000001 1,000,000 One in a million chance of failure 0.0000000001 1,000,000,000 Effectively, it never fails
Time Dependent Mech PoF time-dep = f (TTF) where TTF = time to failure TTF = (available pipe wall) / [(wall loss rate) x (1 mitigation effectiveness)]
Advantages of New Exposure Estimates Estimates can often be validated over time Estimate values from several causes are directly additive. E.g. falling objects, landslide, subsidence, etc, each with their own frequency of occurrence can be added together Estimates are in a form that consider segment-length effects and supports PoF estimates in absolute terms Avoids need to standardize qualitative measures such as high, medium, low avoids interpretation and erosion of definitions over time and when different assessors become involved. Can directly incorporate pertinent company and industry historical data. Forces SME to provide more considered values. It is more difficult to present a number such as 1 hit every 2 years
Measuring Mitigation Strong, single measure or Accumulation of lesser measures Mitigation % = 1 - (remaining threat) Remaining Threat = (remnant from mit1) AND (remnant from mit2) AND (remnant from mit3)
Measuring Mitigation Mitigation % = 1-[(1-mit1) x (1-mit2) x (1-mit3) ] In words: mitigation % = 1 - (remaining threat) remaining threat = (remnant from mit1) AND (remnant from mit2) AND (remnant from mit3) What is cumulative mitigation benefit from 3 measures that independently produce effectiveness of 60%, 60%, and 50%?
Coating-CP Interaction PoD of coating OR gate CP effectiveness
Calibrating Coating Fail Rate % bare from CP current demand. or calibrate coating fail rate using DOT stats coat defect rate dia A, sqft/linear ft A, sqft/mi fail/mil-yr % corr fail coating fail fctr fails/per sq ft/yr 24 6.3 33,175 0.001 0.3 100 9.0E-07 0.003 fails per yr per 33K sq ft 12 3.1 16,588 0.001 0.3 100 1.8E-06 10 2.6 13,823 0.001 0.3 100 2.2E-06 8 2.1 11,058 0.001 0.3 100 2.7E-06 6 1.6 8,294 0.001 0.3 100 3.6E-06 assume 100x as many coating failures as corr failures
PoD for Coating Coating defect rate Probability of Defect in Segment, per year Score per sq ft L = 1 ft L = 10 ft L = 100 ft L = 1000 ft L = 5280 ft excellent 5.0E-07 0.00% 0.00% 0.02% 0.16% 0.83% good 2.5E-06 0.00% 0.02% 0.18% 1.75% 8.91% fair 3.2E-04 2.46% 22.1% 91.8% 100% 100% poor 4.0E-02 11.8% 71.4% 100% 100% 100% absent 1.0E+07 100% 100% 100% 100% 100% 12 diameter PoD = 1 - EXP[-[surface area, ft 2 )x(failure rate per ft 2 )]
Corrosion Control Effectiveness Coating Condition excellent Coating Defect Type Prob of Defect Type per sq ft none 99.9% hole 0.1% shielding 0.0% Is CP fully eff? Prob of CP protecting pipe Scenario Prob Y 0.9 99.99% 0 N 0.1 0.01% 0 Y 0.9 0.09% 0 N 0.1 0.01% 16 Y 0.1 0.0% 0 N 0.9 0.0% 16 Resultant MPY Final probability of 16 mpy damage rate per sq ft 0.01% 16 Final probability of 0 mpy damage rate per sq ft 99.99% 0
Damage vs Failure Probability of Damage (PoD) = f (exposure, mitigation) Probability of Failure (PoF) = f (PoD, resistance) Exposure Mitigation Resistance PoD PoF
Estimating Resistance pipe spec (original) historical issues low toughness hard spots seam type manufacturing pipe spec (current) ILI measurements calcs from pressure test visual inspections effect of estimated degradations required pipe strength normal internal pressure normal external loadings
Best Estimate of Pipe Wall Today Measurement error Degradation Since Meas 2007 Estimate Press Test 1992 (inferred) +/- 5% 8 mpy x 15 yrs = 120 mils ILI 2005 +/- 15% 8 mpy x 2 yrs = 16 mils
Best Estimate of Pipe Wall Today Best Est Today Press Test 1 ILI 1 Bell Hole 1 Press Test 2 Bell Hole 2 ILI 2 NOP
ILI Capability Matrix Defect Type Max Surviving Defect Inspection Validation (Pig- External Internal Axial Circum Dent/ Ovality/ Type Digs) Protocol Corrosio Corrosio Crack Crack Gouge Buckling Lamin Metal Loss Crack Aggressive 5 5 100 10 20 50 50 5 100 MFL high Routine 10 10 100 50 50 50 50 10 100 resolution Min 15 15 100 100 50 50 50 15 100 Aggressive 10 10 100 50 50 50 50 10 100 MFL std Routine 15 15 100 100 50 50 50 15 100 resolution Min 20 20 100 100 50 50 50 20 100 Ultrasound 5 5 100 100 20 20 5 5 100 TFI 20 20 5 10 50 50 50 20 5 EMAT 50 50 10 10 50 50 10 50 10 Ultrasound shear wave crack tool 50 50 10 10 50 50 10 50 10 Caliper, sizing, gauging, inertial 100 100 100 100 5 5 100 100 100 Press test 5 5 5 5 2 2 2 5 5
NOP & Integrity + Integrity info -Rupture potential -Higher stress -Fatigue
TTF to PoF PoF Time
TTF to PoF PoF Time
PoF: TTF & TTF99 PoF PoF=100% PoF=1% time TTF99
Examples TTF = 0.160 / [(16 mpy) x (1-0.9)] = 100 years TTF99 = 0.160 / (16 mpy) = 10 years PoF => lognormal or other =>0.001% TTF = 0.016 / [(16 mpy) x (1-0.9)] = 10 years TTF99 = 0.016 / (16 mpy) = 1 year PoF = 1/TTF = 1%
Approach Advantages more intuitive better models reality distinguishes between unmitigated exposure to a threat, mitigation effectiveness, and system resistance--better risk management decisions eliminates need for unrealistic and expensive re-weighting of variables for new technologies or other changes flexibility to present results in either absolute terms or relative terms more audit friendly