Looking at Figure 2, the process scheme for the alkylation plant, we can see that the location of the failed elbow is in the discharge of the propane recycle pump back to the depropanizer. HF is expected in the overhead of the depropanizer which is the feed source for the pump. The overhead drum also serves as a means to decant free HF from the propane before is returned to the depropanizer as reflux. So one would expect a finite amount of soluble HF in the propane stream under normal operations.
The piping system is stated to have been installed in 1972 and was without modification or replacement of components over a period of more than 46 years. That suggests metal loss was extremely small and the rate of corrosion was quite low for much of the life of the piping system.
The metallurgy of the failed elbow is uniquely different from what would be expected to be the case for a piping item conforming to the specification for the material. Perhaps this was a contributing factor.
One possible cause of the localized thinning may have been carryover of HF into the propane with the high HF fluid held up in a non-flowing section of the pump discharge piping. Be interesting to see if the HAZOP of the piping system contemplated the possibility of "more than" HF in the non-flowing region of the piping system.
Al Center
Professor of Practice
Robert Frederick Smith School of Chemical and Biomolecular Engineering
Cornell University
Ithaca, NY 14853