Reliability Engineering Snapshot TM

But somehow Scotty always managed to get that last little bit out of those trusty old engines of the Enterprise. Unfortunately for me, I don't have script writers in my world, I'm stuck. I live in a welded world of stainless steel, mainly 316L (low carbon). On any given day chlorides are in the 650 to 700 PPM region. Depending upon the particular service, temperatures range from 100 C all the way up to 500 C. I also live in a world outside of everybody's test curves, or at least, in regions of the curves that everyone avoids. When it comes to metallurgical problems such as intergranular corrosion, sensitization, oxidation, carbide precipitation, low temperature sensitization (LTS), crevice corrosion, or chloride stress corrosion cracking (CSCC), I have to ask myself this question "What worm hole did I slip into?" Where's Scotty when you really need him.

It's a real barn burner when it comes to the battle over the question of CSCC. The American Society of Metals (ASM) has a lot of good literature on that subject and on stainless steels in general. One of their articles starts right out and says that there are A LOT of viewpoints regarding CSCC. The literature will tell you that CSCC is transgranular in stainless steels, unless it has been sensitized, then it is intergranular. However, none of the literature conveniently fits this particular problem to my satisfaction.

I'm going to offer up my own observations as follows. In this part of the world most of the cracking that takes place is in welded regions. These regions are usually protected, such as bands welded from the outside to join two pieces together. The two sections are never seal welded from the inside. Thus, the environment can get into the region between the exterior band and the interior sections. Other regions are exterior patches, and today's discussion, flange gasket surfaces such as the one shown in the picture to the left. Cracks usually take 8 to 10 years to initiate (i.e. detected by visual inspection). Crack propagation is intergranular within the welded region (picture left). There is never any cracking in non welded regions, just uniform corrosion and or oxidation (picture lower right). Whenever EdX is used to determine the composition of the stuff at the crack tip there is always 5,000 to 20,000 PPM of chlorides present. The type of cracking pictured here only happens where there is a stress condition, mechanical or thermal. Cracking never occurs in stressless regions.

Let's look at this problem the way Sherlock Holmes might have looked at it. When you eliminate all the causes that couldn't have done it, then the one remaining, no matter how illogical it might seem, is the culprit. First off, intergranular corrosion. That would mean that this should be happening everywhere and not in just one location. This type of cracking isn't happening elsewhere on the flange or in the nozzle vicinity. Scratch that one. Second candidate, sensitization from welding. Good possibility, but none of the other welds that are exposed to the same environment are cracking or corroding for that matter. Scratch that one. Third candidate, oxidation. Oxidation (for want of a better term) is evident on the nozzle area (picture lower right). It is granular in appearance to the eye. The cracking in the flange weld region does not have the same appearance. None of its surface is missing. Scratch that one. Fourth candidate, carbide precipitation. If it were the cause then the problem should be seen everywhere and not in just the welded regions. Scratch that one. Fifth candidate, low temperature sensitization (LTS). Good choice, especially for the heat affected zone (HAZ) of the weld, but again, it isn't happening everywhere, or on other welds that are in the open. Sixth candidate, crevice corrosion. Now this really is a good choice. It would explain why the cracking is occurring in just the flange region and not all of the open regions. What kills this choice is the fact that the entire flange is a crevice, and the cracking did not leave the HAZ. This holds true in every other crevice location that this occurs. Now we get down to the last choice, chloride stress corrosion cracking (CSCC). Maybe it isn't the last choice, but for this article it is.

Everywhere cracking has occurred it has always been in a region that sees a higher amount of stress, by several orders of magnitude. Look at the top left picture that shows the flange and nozzle. Imagine, if you will, that on the inside of that flange temperatures are in the 350 C region, while on the outside, the flange and its cover are immersed in water. Ah, makes yours eyes water just thinking about the magnitude of the thermal stresses on that flange weld. The nozzle is trying to grow radially outwards while the flange restricts it. "But Rick," you say, "the weld is in compression, cracks don't initiate or propagate in compression, gotcha ya!" NO, WRONG. Look at the picture to the right. The flange surface at the ID is a built up fillet weld. The flange sits on top of the nozzle. The nozzle ID is smaller than the flange ID. The built up fillet weld is actually in tension. Those cracks will have as much as 5,000 PPM of chlorides at the tip. But why chlorides, why not phosphorous or sodium or potassium, which are the other elements in this service? I'll leave that up to the scientists.

So why not be politically safe and just call it stress corrosion cracking, why stick my neck out and call it chloride stress corrosion cracking? Because something, and I say something, is aggravating the stress. There are welds near this nozzle that see just as high a stress but are not failing. Those welds are out in the open where nothing is protected and chlorides cannot accumulate.