I am looking at an overfill case for a product separator on a refinery hydrotreating unit. The separator receives a two phase mixture from the reactor effluent trim cooler. It separates the feed into three phases; vapor, liquid hydrocarbon, and sour water. The vapor product flows to an amine contactor to remove H2S, then to a KO drum and finally to the recycle compressor. In the event the hydrocarbon product valve fails closed, the separator will fill with liquid hydrocarbon, then the amine contactor will fill with hydrocarbon liquid and then the compressor suction KO drum will fill with liquid. This should not result in the separator exceeding its MAWP. High-High level in the KO drum should shutdown the compressor (2oo3 voting level transmitters). The compressor trip also shuts down the charge pump. Even if the charge pumps fails to shutdown the PSV on the separator is adequate with the compressor down.
However to be conservative, one can consider a contingency were the separator PSV needs to relieve all the vapor and liquid hydrocarbon feeding the separator. I sized a PSV using the HEM method (Annex C API 520) with an inlet line that resulted in an inlet pressure drop of less than 3% of set pressure. However when I looked at the flow regimes in the inlet line the horizontal sections are in plug flow and vertical sections are in annular flow. One of the assumptions for the HEM method is that the vapor and liquid phases must be in mechanical equilibrium i.e. moving at the same velocity. This is clearly not the case with the inlet line under consideration. Making the line bigger likely aggravates the slug flow. I also have concerns that the vertical section of piping will transition to plug flow as the PSV cycles closed.
Is it appropriate to design the PSV using the HEM method and also ignore the flow regimes in the inlet piping with the possibility of a liquid slug hitting the PSV?