Proposed API 581 Inspection Plan Optimization Example

In a previous post, Proposed API 581 Inspection Plan Optimization, I presented an optimization to API 581⁽¹⁾. In this post I work out an example to illustrate its point. I created an example that gives us probability of failure (POF), consequence of failure (COF), and risk using an example already published by Trinity Bridge in 2011 ⁽²⁾ as a guide. This example's assumptions are:

• The data quality is good
• The RBI Analysis Date was 05/01/2008
• The Planned Date was 05/01/2028
• The asset's In Service Date was 01/01/1972
• Two damage mechanisms exist: internal thinning and internal cracking
• The only inspection performed was for thinning on 04/04/2004, API 581 Level B effectiveness
• No inspections past the RBI Analysis Date were assumed for typical inspection planning purposes as this example shows the risk evaluated "as is" and projected for 20 years
• The Risk Target was assumed = 40 ft²/yr
• The COF (C_f) was simplified to COF Personnel, ft², only
• POF was generated for three damage factor cases with F_MS = 1 and gff_total = 3.06E-05
  • POF for combined damage factor, P_f(t) = gff_total ⋅ D_f(t) ⋅ F_MS
  • POF for thinning damage factor, P_f(Thinning) = gff_total ⋅ D_f(Thinning) ⋅ F_MS
  • POF for cracking damage factor, P_f(Cracking) = gff_total ⋅ D_f(Cracking) ⋅ F_MS
• Risk was calculated three ways
  • With POF for combined damage factor, R(t) = P_f(t) ⋅ C_f
  • With POF for thinning damage factor, R(Thinning) = P_f(Thinning) ⋅ C_f
  • With POF for cracking damage factor, R(Cracking) = P_f(Cracking) ⋅ C_f

These three risk values were projected over a 20 year plan period and compared to the risk target of 40 ft²/yr. We will not address the specific steps to calculate these risk values in this example, but below is a presentation in tabular format of the results.

Table (1): Risk Values Projected Over 20 Years

The table above only presents the relevant columns and values that are needed for illustration. The yellow highlight shows:

• The calculated risk value (ft²/yr) for each risk case when it met or past the Risk Target.
• The time (yr) for each risk case to meet or pass the Risk Target.

The results show for each risk case:

• The combined damage factor case exceeded the Risk Target almost immediately, on or before 05/01/2009. The time to target for this case was one year or less.
• The individual thinning damage factor case exceeded the risk target just shortly after the combined, on or before 05/01/2010. The time to target for this case was two years or less.
• The individual cracking damage factor case did not exceed the risk target until just on or before 05/01/2024. The time to target for this case was 16 years or less.

So here we are in 2008, we have run the risk analysis to determine where we are and how to plan the inspections for the future. Here are the outcomes:

• Using the traditional 581 planning approach with the 20 year period, it is likely that both the cracking and thinning inspections would be due on or before 05/01/2009 to get past the Plan Date of 05/01/2028.
• Using the proposed individual damage mechanism risk approach, the thinning inspection would be due on or before 05/01/2010, and the cracking inspection would be due on or before 05/01/2024 to get past the Plan Date of 05/01/2028.

There is obvious value here and it allows the Owner/User to be more flexible in their inspection planning process. However, there are literalists who believe that if they follow the methodology exactly, they will not be penalized if an incident occurs. Experience tells us that this is a false assumption. As API 581 is open to interpretation, for the literalist if it is not written down, it is precluded from being done. This is a narrow interpretation with which many of you will disagree.

My personal belief is that API recommended practices, such as 581, should be interpreted by the Owner/User. I also believe that if alternatives are not provided in these same recommended practices, governments and regulatory agencies will take them literally. This will invalidate the Owner/User argument that if it is not explicitly precluded in the recommended practice, then it can be done. Finally, this will stop the Owner/User from interpreting these recommended practices to their benefit. Do any of you believe that there is only one way of "doing things"?

Let's continue the conversation! I am interested in your feedback on this proposal. Please comment below or contact me directly to start the conversation.

REFERENCES

1. API RP 581, Risk-Based Inspection Methodology, Third Edition, April 2016. API Publishing Services, 1220 L Street, N. W., Washington, D. C. 20005
2. Lynne Kaley, "API RP 581 Risk-based Inspection Technology Demonstrating the Technology Through a Worked Example Problem”, Inspectioneering Journal, March/April 2009.

Comment on this article