Sec. VIII Div. 1 Separator Sump Attachment

[dtn6770]

I’ve been asked to review the design of a 500 psig MAWP separator having a NPS 12 liquid “sump” that is set on (abutted to) the exterior surface of the main, horizontal, 30” diameter shell. Liquids are communicated to the sump through two 2” diameter holes that are drilled through the main shell in the area ‘covered’ by the sump. The sump-to-shell connection is T-joint made by a single, external groove weld (sump material) topped with a fillet weld.

This is a far cry from a traditional sump that is treated / designed essentially as a nozzle, Code wise. I’ve known this fabricator to creatively minimize their costs and this design fits their MO. I have non-Code reasons to reject the design but would like to include Code reasons as well but, with limited research, so far I’m not having much luck. The design is a variation of a nozzle and a communicating chamber, Category D joint all the way but nothing explicit in the Code that I’ve seen.

Comments or suggestions are appreciated.

 

[jtseng123]

My opinion: as long as the weld design (full penetration with outside fillet (and inside fillet if possible)), and proper NDE, it is good to go. The shell plate with holes inside the 12" nozzle becomes non-pressure part.
Also, fabricator has to run nozzle reinforcement per code to see if pad is needed.

My question: is the (2)-2" hole big enough for process ? why not drill one big hole ?

 

[dtn6770]

Why not drill one big hole indeed. Their strategy is to avoid reinforcement by using two small holes that the Code doesn't require to be reinforced. A larger hole would start to resemble a real nozzle and kick in all the requirements thereof.

I haven't found a Code basis for evaluating the strength of that type of joint and it doesn't lend itself to RT. Part of my process will be to ask for their calculations.

I'm not thrilled about the non-venting vapor space that will exist at the top (sides) of the sump.

Thanks for the comments.