ILI vs DA—The Risk View

A rather common issue these days is decision making involving comparisons between In Line Inspection (ILI) vs Direct Assessment (DA, and, in particular, External Corrosion DA or ECDA).  You may need to research these techniques, if not already familiar to you, since we will assume the reader is fairly familiar with both.

Both are acceptable ways to assess integrity when complying with aspects of US regulatory Integrity Management (IMP).  However, they add value—that is, they reduce risk—in two quite different ways.

Comparisons between the two are challenging.  Not because the issues are overly complex but rather because there are many nuances to consider.  In this article, we’ll use a very simple example to examine just how each contributes to risk management: ie, the ways in which each reduces risk.   This examination will illustrate several key points:

  1. How good risk assessment can add clarity to otherwise confusing issues.
  2. Why this particular comparison is not intuitive to many.
  3. Yet another example of why we must measure PoF via three independent measurements of exposure, mitigation, and resistance (see previous articles).

Spoiler alert:  here’s a preview of the conclusion.  One of the two provides mostly information on mitigation while the other provides mostly information on resistance.  This will be key to understanding why the two play different roles in risk reduction.

But on to our examination. To reduce the number of possible confounding issues, let’s theorize a rather simple scenario.  Two segments of nearly identical pipelines exist.  Both have current and future estimates of unmitigated corrosion rates, mitigation effectiveness, and possible loss of strength due to corrosion damages.  One is assessed by ILI and no external metal loss anomalies are found.  The other is assessed by ECDA, where two overline surveys followed by confirmatory excavations, find no anomalies.  So, both assessments have been performed and neither has found any external metal loss at all. Let’s assume both are perfect in their ability to detect metal loss—not realistic, but avoids a confounding aspect in our comparison.

How does a good risk assessment use this knowledge of ‘inspected with no findings of deficiencies’?  It first recognizes what has been learned.  The ILI has shown that there are no weaknesses exist that were caused by external metal loss.  The risk assessment had conservatively assumed that some corrosion metal loss had been occurring.  Since none was found, the ‘clock has been reset’. The assumed metal loss did not actually occur, and so the original strength (wall thickness) is used in risk calculations.  However, we do not know if future damage is imminent—perhaps corrosion is just beginning in some locations.  The ILI also provides some indirect evidence regarding exposure and mitigation.  Let’s discount that for now.

The ECDA has shown that the mitigation—cathodic protection (CP) and coating—are performing as intended at all points.  However, we do not know if damages have occurred in the past—perhaps recent upgrades to a CP system conceal past deficiencies.  As with ILI, we have also some indirect evidence regarding other PoF issues.  Afterall, an underlying premise with DA is that the findings from more likely damage locations provides assurances for the less likely locations.  For instance, an ECDA finding of no deficiencies, coupled with knowledge that mitigation systems have not changed, perhaps offers some assurance that past damages have probably not occurred.  Let’s also discount that for now and look at some sample numbers for these two assessed segments.

year 1 unmit 16 16
mit 98.5% 99.2%
res 0.22 0.15
PoF 0.0011 0.0009
year 3 ann red in mit 0.01 0.01
yrs 3 3
mit 95.5% 96.2%
res 0.217 0.148
PoF 0.003 0.004


*go to to see details behind these calculations

These sample values include corrosion rate, mitigation effectiveness, changes in mitigation effectiveness per year (ie, rate of coincident gaps in CP and coating), and assumed changes in wall thickness prior to the assessments.  The results are interesting.  The PoF comparison for year one after the two assessments shows that the ECDA-assessed segment has slightly lower PoF.  In year 3 however, the ILI-assessed segment shows lower PoF.

As some might suspect, yes, this example was contrived so that the two techniques effectively swapped positions in risk-reduction-effectiveness over time.  But that is not necessarily unrealistic.  There are so many nuances of scenario that the possibilities are endless.  Recall, for instance, that we ignored indirect evidence that was possibly available and also the efficiency with which anomalies could be detected with each technique. Either technique can be more beneficial depending on many factors, including the time horizon for which the risk assessment reports.

Here is the key point of this discussion.  ECDA is a forward-looking technique since it mostly yields information on mitigation:  how well protected is the system.  ILI and pressure testing are backwards-looking techniques:  what damages have already occurred.  Both are important.  Neither tells us everything.  Both contribute to risk reduction.  But they do so in different ways.  Good risk assessment allows us to clearly see the role of each and make better decisions.