FFS Forum: Crack Growth Assessment at Pits

A Recent Trend

I don’t know if it qualifies as a trend, but I’ve certainly noticed something in the last couple of years that didn’t come up at all in the previous 40 years of my career. More often, I’m seeing people perform a detailed stress analysis of individual pits to support subsequent fracture mechanics (FM) cycle-life calculations. That is, they are concerned about the possibility of cracking at the base of pits and want to perform FM flaw growth calculations on those hypothesized flaws to establish a predicted number of cycles-to-failure.

FM flaw growth calculations on hypothesized flaws aren’t new – that’s been a common practice throughout my career. But trying to get detailed numerical solutions (i.e., FEA) to characterize the through-wall stress distribution under the pit does seem to be a relatively recent trend. To be sure, this is in part due to more capable finite element (FE) software and faster desktop computing speeds.

My thoughts on this topic are not so much focused on how to do it, but on whether it is worth the trouble. Or, to put it another way, just because we can do it, should we do it?

The issue usually comes about when pitting is discovered in a critical system and there is concern that cracking at the base of the pits could result in flaw growth and potentially a catastrophic crack-growth failure. Pits do function as stress risers, so the through-wall stress distribution below the pits can be significantly different from the nominal distribution without pits.

The approach I’ve most commonly seen is to construct an FE model of one (or at most, several) “representative” and idealized pits. Many simplifying assumptions are made to narrow the scope of the assessment, such as an assumed hemispherical pit of a given radius. The typical process is shown in Figure 1.

What does API 579 have to say on the topic?

There are several places to look in the API 579 Fitness-for-Service Standard (API 579) for guidance on cracking related to pits.

Part 5, Local Metal Loss

Part 5 includes consideration of “grooves” from erosion or corrosion. As the groove root radius gets smaller, the groove becomes more “crack-like.” This Part provides guidance on when the groove should be considered a crack. However, once the groove is determined to be crack-like, the user is sent to Part 9 to implement an FM-based flaw assessment.

For our problem, we’re starting out with the assumption that there is a crack under the pit. So, Part 5 doesn’t really provide any assistance.