It is commonly agreed by all involved in the petrochemical, power and pipeline industries that the consistent ability to detect damage mechanisms is very important. Yet, as a general rule, we do not give enough weight to the actual probability of detecting damage, which we refer to as probability of detection or POD. We commonly assume that the POD is all based on the technology utilized or the experience of the technician doing the inspection. These are important factors, but in the real world they do not tell entire story.
To examine why, let’s explore the elements which most impact whether or not flaws will be detected. It’s a combination of the following:
- The inspection plan and the amount of real estate covered
- The technology and equipment you are using, including calibration
- The process and procedures used to carry out the inspection and their consistency
- The testing environment, including weather and other hazards such as weak scaffolding
- The training and ability of the technicians performing the tests, including physical condition, experience with a particular damage mechanism, mental state, etc.
Each one of these factors has a tolerance, and the overall probability of actually finding damage mechanisms or degradation modes is a multiple of these. This mixture is not often considered when gauging the detection capabilities of an NDT method, and this leads to increased risk exposure.
As an example, the probability of finding general wall thinning could be expressed as being:
the maximum POD (100%) x the inspection plan (.90 – .60) x the capability of the thickness gage (99.5%) x procedure quality (.95 – .70) x environment (.90 – .60) x human factor (.90 – .50)
When you multiply them all the way across you can get a wide variation in POD, also known as the quality of the inspection. Of course, POD is not the only thing that is important — working safely and promptly delivering precise and actionable information is also important. However, the accuracy and reliability of the data gathered is the whole reason you’re there in the first place.
If you do the math on your inspection coverage, you’ll stop wondering why you still get surprised by leaks. Let’s take a look at the two extreme scenarios based on multiplying out the earlier formula.
Best case: 100% x .9 x .995 x .95 x .9 x .9 = 68.5%
Worst case: 100% x .6 x .995 x .7 x .6 x .5 = 12.5 %
Keep in mind these are examples and I don’t stand by these actual numbers, but I present them to illustrate that we have a huge amount of potential variability. As an industry, we leave a lot to chance in regards to our planning and the expectations we have for the people gathering the data. What this results in is unnecessary risk exposure and too often missed or inaccurately assessed damage
What can be done about it? By paying attention to the factors above and optimizing where we can (such as better trained technicians or a more thorough inspection plan) we can mitigate these risks and improve the effectiveness of maintenance and inspection programs across the energy and chemical industries. In the long run, this will result in less downtime and safer operations. More about each will be discussed in future posts.