4. โ Structural assets
โ Dependent on site
knowledge/history
โ Analysis done externally
โ Results input into PCMS
โ Process assets
โ Dependent on data
โ Calculation based
โ Results generated within
PCMS
QUALITATIVE VS. QUANTITATIVE ANALYSIS
6. Consequence Risk Drivers
โข Release Quantity
โข Process Constituents
โข Lost Production
Probability Risk Drivers
โข Design Conditions
โข Age of Equipment
โข Susceptibility to Damage
โRisk drivers are items affecting probability and/or
consequence such that it constitutes a significant portion
of the risk.โ
RISK DRIVERS
7. โExtent to which an Event is likely to occur within the time frame under
consideration. The mathematical definition of probability is a real
number in the scale 0 to 1โ
Also referred to as likelihood or frequency in the industry.
PROBABILITY
8. โOccurrence of a particular set of circumstances. The event may be
certain or uncertain, singular or multiple.โ
Also described as a loss of containment due to a failure.
EVENT
9. โA process that induces micro and/or macro material changes over time
that are harmful to the material condition or mechanical properties.โ
DAMAGE MECHANISM
Damage Mechanism Categories
โข Internal Loss of Thickness
โข External Loss of Thickness
โข Environmentally Assisted Cracking
โข Mechanical and Metallurgical Failure
โข General
โข Local
Damage Types
10. Probability is calculated for each susceptible damage mechanism in
PCMS.
PROBABILITY CALCULATIONS
๐๐๐๐๐๐๐๐๐๐๐ =
๐๐๐ โ ๐๐
๐๐ง๐ฌ๐ฉ๐๐๐ญ๐ข๐จ๐ง ๐๐ซ๐๐๐ข๐ญ
Where:
Gff = Generic Failure Frequency
DF = Damage Factor
11. โA probability of failure developed for specific component types
based on a large population of component data that does not
include the effects of specific damage mechanisms.โ
GENERIC FAILURE FREQUENCY
Gff tables found in API 581- Part 2, Table 4.1
Equipment type Gff (failures/yr)
Heat Exchanger 3.06E-05
Pipe 3.06E-05
Tank 650 Tank Bottom 7.20E-04
Tank 650 Tank Shell 1.00E-04
Pressure Vessel 3.06E-05
12. โAn adjustment factor applied to the generic failure frequency to
account for damage mechanismsโ that may be present as a function of
time in service for a specific asset.
For thinning mechanisms:
DAMAGE FACTOR
Where:
A = Age
R= Rate of Corrosion
T= Original Thickness
๐๐ โ
๐ โ ๐
๐
AR/T
Thickness Loss Damage
Factor
AR/T
Thickness Loss Damage
Factor
< 0.08 1 0.25 to 0.30 650
0.08 to 0.10 2 0.30 to 0.35 750
0.10 to 0.12 6 0.35 to 0.40 900
0.12 to 0.14 20 0.40 to 0.45 1050
0.14 to 0.16 90 0.45 to 0.50 1200
0.16 to 0.18 250 0.50 to 0.55 1350
0.18 to 0.20 400 0.55 to 0.60 1500
0.20 to 0.25 520 >0.60 1900
AR/T tables found in API 581- Part 2, Table 5.11
13. Determination of the effectiveness of the inspection in identifying and
quantifying the type and extent of damage per damage mechanism. As
per API effectiveness ranges from Highly Effective (A) to Ineffective (E).
Mosaic uses A, B, & C.
INSPECTION EFFECTIVENESS
API 581 Figure 4.3
Failure Mode Standard (C) Medium (B) High (A)
Internal
Thickness Loss
10 100 1000
External
Thickness Loss
10 100 1000
Environmentally
Assisted Cracking
3 10 30
Mechanical &
Metallurgical
Failure
3 10 30
Inspection Effectiveness Table
14. INSPECTION EFFECTIVENESS
Mosaic MI Documents, Table 2: Guidelines for Assigning Inspection Effectiveness โ General Internal Corrosion
Inspection
Category
Inspection
Effectiveness
Category
Intrusive Inspection Non-Intrusive Inspection
A
Highly Effective
>50% visual examination of the surface area
(internals removed as required) with follow-
up by UT, RT or pit gauge as required
Record spot UT measurements at likely
location(s) with min of 1 CML per
fitting1/section2
B Usually Effective
>20% visual examination of the surface area
(internals removed as required) with follow-
up by UT, RT or pit gauge as required
Record spot UT measurements at likely
location(s) with min of 0.5 CML per
fitting1/section2
C Fairly Effective
<20% visual examination of the surface area
(internals removed as required) with follow-
up by UT, RT or pit gauge as required
Record spot UT measurements at likely
location(s) with min of 0.25 CML per
fitting1/section2
No Credit Ineffective
Less than โCโ effectiveness, no inspection
or ineffective inspection technique used
Less than โCโ effectiveness, no
inspection or ineffective inspection
technique used
16. โAn outcome of an Eventโฆmay be one or more
consequencesโฆconsequences are always negative for safety aspectsโ.
Environmental and Economic consequences are also always negative.
CONSEQUENCE
Remember: An Event is an occurrence of a set of
circumstances, also a loss of containment due to a failure.
17. Consequence can be calculated as either total or worst case. For Mosaic:
CONSEQUENCE CALCULATION
๐๐จ๐ง๐ฌ๐๐ช๐ฎ๐๐ง๐๐ = ๐๐๐ฑ [๐๐๐๐ฅ๐ญ๐ก & ๐๐๐๐๐ญ๐ฒ, ๐๐ฆ๐๐ ๐, ๐๐ง๐ฏ๐ข๐ซ๐จ๐ง๐ฆ๐๐ง๐ญ๐๐ฅ, ๐๐๐จ๐ง๐จ๐ฆ๐ข๐]
18. Consequence associated with product releases affecting personnel (site and
community). Image consequence often goes hand in hand with Health & Safety.
HEALTH & SAFETY
H&S Consequence Drivers:
โข Flammable Events
โข Toxic Events
โข Vapour Cloud Explosions (VCE)
๐๐&๐ $ = ๐๐๐ฑ ๐๐&๐
๐ ๐ฅ๐๐ฆ
, ๐๐&๐
๐๐จ๐ฑ
, ๐๐&๐
๐๐๐
21. HEALTH & SAFETY: VCE
๐๐&๐
๐๐๐
= ๐๐ฎ๐ฆ ๐๐&๐
๐๐๐,๐ฅ๐จ๐ฐ
, ๐๐&๐
๐๐๐,๐ฆ๐๐๐ข๐ฎ๐ฆ
, ๐๐&๐
๐๐๐,๐ก๐ข๐ ๐ก
,
VCE Rating VCE Description
High Flammable & Liquid & Boiling Point < 50 ยฐF
Medium Flammable & Liquid & Operating Temp > Boiling Point
Low Not flammable or Vapor or Liquid & Operating Temp < Boiling Point
Release Quantity (tons) VCE Potential
Min Max Low Medium High
20 9,999,999 $ - $ 5,000,000 $ 10,000,000
10 20 $ - $ 500,000 $ 5,000,000
1 10 $ - $ 50,000 $ 500,000
0 1 $ - $ 5,000 $ 50,000
22. Consequence associated with product releases resulting in
contamination of soil, ground water and/or open water and air.
ENVIRONMENTAL
Environmental Consequences:
โข Clean up of Environment
โข Regulatory Citations and Fines
23. ENVIRONMENTAL CALCULATIONS
๐๐๐ง๐ฏ๐ข๐ซ $ = ๐๐ฎ๐ฆ ๐๐๐ง๐ฏ๐ข๐ซ
๐ฅ๐จ๐ฐ
$ , ๐๐๐ง๐ฏ๐ข๐ซ
๐ฆ๐๐๐ข๐ฎ๐ฆ
$ , ๐๐๐ง๐ฏ๐ข๐ซ
๐ฆ๐๐๐ข๐ฎ๐ฆ/๐ก๐ข๐ ๐ก
$ , ๐๐๐ง๐ฏ๐ข๐ซ
๐ก๐ข๐ ๐ก
$
Rating Description
High Harmful and Toxic
Medium/High Hydrocarbons
Medium Harmful but not Toxic (Most Alkanes)
Low Not Harmful to the Environment
Release Quantity
(tons)
Environmental Impact Rating
Min Max Low Medium Medium/High High
50
99,999,99
9
$ 50,000 $ 5,000,000 $ 5,000,000 $ 10,000,000
25 50 $ 5,000 $ 500,000 $ 5,000,000 $ 5,000,000
5 25 $ 5,000 $ 500,000 $ 500,000 $ 5,000,000
0.75 5 $ - $ 50,000 $ 500,000 $ 500,000
0.05 0.75 $ - $ 50,000 $ 50,000 $ 500,000
0 0.05 $ - $ 5,000 $ 50,000 $ 50,000
24. Economic Consequence is the result of business interruption.
ECONOMIC
Economic Consequences:
โข Lost Production
โข Repair Costs
โข Downtime of Associated Units
๐๐๐๐จ๐ง
๐ฅ๐จ๐ฌ๐ญ ๐ฉ๐ซ๐จ๐
$ = ๐๐๐ซ๐ ๐ข๐ง
$
๐๐๐ฒ
โ ๐๐๐ % โ ๐๐๐ ๐๐๐ฒ๐ฌ
26. Critical Asset Information Value COF POF
Critical Asset Information Value COF POF
Semi-Quantitative Risk Based Inspection Workshop
Risk = x
Probability Consequence
Material of Construction
Process
Operating Temp.
Operating Press.
Flow Velocity
Damage Mode
Date Installed
Last Inspection
Original Thickness
Corrosion Rate
Plant Rate Reduction
Plant Margin
Repair Time
C.S.
H2SO4 98%
oF 80
90 psi
10 ft/s
Gen. Thin.
02/2006
N/A
0.5 inches
.007 in/yr
50%
0.5 MM/day
6 days
What is the Risk?
Secondary Containment
Release Quantity
No
10 bbl
High, Medium, or Low
35. Inspection
Category
Inspection
Effectiveness
Category
Intrusive Inspection Non-Intrusive Inspection
A
Highly Effective
>50% visual examination of the surface area
(internals removed as required) with follow-
up by UT, RT or pit gauge as required
Record spot UT measurements at likely
location(s) with min of 1 CML per
fitting1/section2
B Usually Effective
>20% visual examination of the surface area
(internals removed as required) with follow-
up by UT, RT or pit gauge as required
Record spot UT measurements at likely
location(s) with min of 0.5 CML per
fitting1/section2
C Fairly Effective
<20% visual examination of the surface area
(internals removed as required) with follow-
up by UT, RT or pit gauge as required
Record spot UT measurements at likely
location(s) with min of 0.25 CML per
fitting1/section2
No Credit Ineffective
Less than โCโ effectiveness, no inspection or
ineffective inspection technique used
Less than โCโ effectiveness, no inspection
or ineffective inspection technique used
ADDING INSPECTION CREDIT
36. Critical Asset Information Value COF POF
Critical Asset Information Value COF POF
Risk = x
COF Calculations POF Calculations
=
Factor 1 Category
COF =
x x =
( )( )( )
POF = x AR
t
=
( ) ( )
x
( )
POF = POF =
1 2
POF = x = POF = x =
AR
t
=
( ) x ( )
( )
=
AR
t
=
( ) x ( )
( )
=
Scenario 1 (Date: ) Scenario 2 (Date: )
=
Factor 2 Category
=
Factor 3 Category
=
Factor 4 Category
Risk =
1 Risk =
2
Probability Consequence
Material of Construction
Process
Operating Temp.
Operating Press.
Flow Velocity
Damage Mode
Date Installed
Last Inspection
Original Thickness
Corrosion Rate
Plant Rate Reduction
Plant Margin
Repair Time
C.S.
H2SO4 98%
oF 80
90 psi
10 ft/s
Gen. Thin.
02/2006
N/A
0.5 inches
.007 in/yr
50%
0.5 MM/day
6 days
Health/Safety
Environment
Community
Business
B
C
C
B
$ .5MM 50% 6
Margin Reduction Rep. Time
$1.5MM
Damage
Factor
Generic Failure
Frequency
Inspection
Credit
Age Rate
Org Thk
02/2014 02/2020
8 .007
.5
3.06x10-5
0.0001836
2
B2
14 .007
0.5
0.196
400/1
3.06x10-5
0.01224
4
B4
B
0.112
6/1
Secondary Containment
Release Quantity
No
10 bbl
2.2 Inspection Level:
DF: ( ) / ( )
Inspection Required
By (Date)
Inspection
Effectiveness
Risk =
POF =
2.1 Inspection Level:
DF: ( ) / ( )
Risk =
POF =
B3
3
C
400 10
Semi-Quantitative Risk Based Inspection Workshop
40. Critical Asset Information Value COF POF
Critical Asset Information Value COF POF
Risk = x
COF Calculations POF Calculations
=
Factor 1 Category
COF =
x x =
( )( )( )
POF = x AR
t
=
( ) ( )
x
( )
POF = POF =
1 2
POF = x = POF = x =
AR
t
=
( ) x ( )
( )
=
AR
t
=
( ) x ( )
( )
=
Scenario 1 (Date: ) Scenario 2 (Date: )
=
Factor 2 Category
=
Factor 3 Category
=
Factor 4 Category
Risk =
1 Risk =
2
Probability Consequence
Material of Construction
Process
Operating Temp.
Operating Press.
Flow Velocity
Damage Mode
Date Installed
Last Inspection
Original Thickness
Corrosion Rate
Plant Rate Reduction
Plant Margin
Repair Time
C.S.
H2SO4 98%
oF 80
90 psi
10 ft/s
Gen. Thin.
02/2006
N/A
0.5 inches
.007 in/yr
50%
0.5 MM/day
6 days
Health/Safety
Environment
Community
Business
B
C
C
B
$ .5MM 50% 6
Margin Reduction Rep. Time
$1.5MM
Damage
Factor
Generic Failure
Frequency
Inspection
Credit
Age Rate
Org Thk
02/2014 02/2020
8 .007
.5
3.06x10-5
0.0001836
2
B2
14 .007
0.5
0.196
400/1
3.06x10-5 0.01224
4
B4
B
02/2020
B
0.112
6/1
Secondary Containment
Release Quantity
No
10 bbl
Inspection Required
By (Date)
Inspection
Effectiveness
2.1 Inspection Level:
DF: ( ) / ( )
Risk =
POF =
2.2 Inspection Level:
DF: ( ) / ( )
POF =
B3
3
400 10
B2
2
400 100
Risk =
B
Semi-Quantitative Risk Based Inspection Workshop
41. Calculate Risk & Inspection Due Date
Inspection Plan
Update PCMS
PROGRAM MAINTANENCE
Management of Change
Loss of Containment
Process Hazard Analysis
Risk Based
Inspection
Training
Who here is familiar with RBI?
This presentation=Recap+Extra
For many years we have followed the industry standard of assessing based on fixed interval- does this make sense?
Using API 580 and 581 codes and decades of industry data we have developed a quantifiable approach and a scientific method that makes sense.
Let me give you an example: Your maintenance manual tells you to change your oil every 5000 km or so. Now (Josh) here is a daredevil and subjects his car to rough terrain and every weekend. (Brian) on the other hand cradles his car to sleep- rarely drives it and maintains it to a T. Do you think both cars should have their oil changed at the same interval?
This is where RBI comes in. Next Slide.
Risk=COF*POF
Risk takes into account the situational differences we were talking about earlier.
Letโs get back to our car example here.
X drives his car at a 150 mph in rainy conditions; Y drives his car at 80kmh on a sunny day.
Probability: Who is more likely to get in an accident?
Consequence: Both get in an accident- which accident has worse repercussions?
This training will help you discern between what factors affect your consequence and probability and how to lower the risk associated with these.
Before we address RBI, it is important to distinguish between Qualitative and Quantitative Analysis.
Qualitative Analysis is performed on structural assets resulting in a fixed interval that is obtained through site knowledge and overall results are uploaded into PCMS.
Quantitative Analysis on the other hand is heavily dependent on data and calculations. It is performed on quantifiable process assets and the calculations are performed and results generated in PCMS.
In our RBI presentation we will primarily be focusing on Quantitative Risk.
As stated before, Risk always equals Consequence times Likelihood
To drive this point home here is the Mosaic Corporate Risk Matrix. One can see how the combination of consequence and likelihood is used to obtain a risk rank.
One more thing to point out here is are the blue lines. These are our โrisk thresholdsโ. Beyond this point our asset is past due for inspection.
Risk drivers are things that have a significant impact on your probability or consequence such that it affects your risk.
A few drivers are mentioned here (discuss these). We will discuss them in further in our exercise later.
Letโs talk a little about Probability.
API 580 defines it asโฆ..(definition above)
Where an event isโฆ(next slide)
As described by API 580โฆ(definition above)
For our purposes, an event can be described as a loss of containment due to a failure
How does this failure occur? Due to damage mechanisms(next slide for DM).
The codes describe a Damage Mechanism as (definition above).
General and Local Damage types are only for thinning mechanisms.
During our corrosion study we will identify DMs that you see at Faustina. These DMs will be used to calculate the probability portion of risk using the following calculation.
What do we mean by each susceptible mechanism? Different Damage Factor calculations exist for different types of Damage Mechanisms.
API 581 Part 2 deals with all of these and all these calculations are accounted for in PCMS.
PCMS performs probability calculations for each Damage Mechanism and uses the worst case for the overall risk calculation.
API 581 defines Gff asโฆ(definition above)
Gff: Industrial data collected over years
Probability of failure based on asset type- how many tanks/vessels/heat exchangers failed in a year over several decades.
Remember Gff does not account for specific damage mechanisms. This is where Damage Factor comes in (next slide)
API 581 defines DF as..(definition above)
While API 581 part 2 addresses DF for several different Damage Mechanisms we will only discuss thinning mechanisms for the sake of brevity in this training.
AR/T is the calculation of % wall loss.
AT/T translated to damage factor.
Age: Measure of time wrt corrosion rate used; Current age of asset
Rate: Either modeled or calculated from TMLs. Mosaic uses CAR=Total Circuit Loss/Total Circuit Years
Thickness: Suggested Thickness- CA+Tmin from manufacturerโs drawings= original thickness- API 581 Part 2, 5.5.3
How can we reduce the inevitable effect that DF has on our probability? By performing effective inspections!- next slide.
While damage factor increases your probability, inspection effectiveness gives you credit for performing inspections and reducing uncertainty.
(Inspection Effectiveness helps mitigate the negative effects of Damage Factor).
It is a simple divisor that reduces risk.
An ineffective inspection would result in a divisor of 1, resulting in no mitigation of risk.
How do I know what grade my inspection receives? (Next slide)
Mosaic MI documents has inspection effectiveness tables for all DMs
According to API 580 Consequence is..(described above).
Recap โeventโ
For our purposes we will be using the worst Consequence/highest value
Flammable Event: Occurs when both a leak and ignition occur through an ignition or auto ignition source. Flash fire, Pool fire.
Toxic Event: Can cause effects at greater distances and do not require another event such as ignition. Acute toxic risks with immediate danger to personnel.
VCE: Combustion of a sudden release of a large flammable vapour cloud.
First determine the flammability rating based on table 1
Determine Release Quantity for each flammable constituent, low, medium & high from table 2
Sum up the dollar amount.
All calculations performed by PCMS
First determine the toxicity rating based on table 1
Determine Release Quantity for each toxic constituent from table 2
Sum up the dollar amount.
All calculations performed by PCMS
First determine the VCE rating based on table 1
Determine Release Quantity for each vapor cloud constituent constituent from table 2
Sum up the dollar amount.
All calculations performed by PCMS
First determine the Environmental Impact rating based on table 1
Determine Release Quantity for each constituent from table 2
Sum up the dollar amount.
All calculations performed by PCMS
Lost production=Margin($/day)*Reduced Repair Rate(%)*Estimated Repair Time(days) (calculated by PCMS)
Repair Cost: Replacement/Repair of Damages Assets. Manually entered into PCMS
Downtime: Manually entered into PCMS
Any questions regarding RBI Theory?
No- Great! Letโs move on to a problem that uses all the principles above!
Reiterate Risk=POF*COF
Ask someone in the room to volunteer to stand next to this vessel for $1MM.
Ask them what they need to know to stand next to itโฆ they should say material, operating temp etc.
Show them all the asset information
Ask the room what they think the risk of this asset it- Low, Medium or High?
People should hopefully give you a wide array of answersโฆ take this opportunity to jump into the risk matric (next slide)
Why do we have a corporate risk matrix?
For uniformity and standardization. Introduce the risk matrix.
Blue lines= risk threshold.
If risk threshold is crossed asset is overdue for inspection!
Ask the audience if anyone knows which factors affect COF vs. POF