Distillation Column PSV Sizing

Let me also mention two footnotes to this discussion. (1) Before doing anything, check to be sure that the reboiler can actually overpressure the column, even without the condenser running. It not too uncommon to find that the reboiler will pinche out before the process side is at relieving pressure. In other words, the saturation temp of the steam, at supply pressure, might not be high enough to boil the process fluid at the relieving pressure. (2) You can usually make a significant reduction in the heat input for this scenario by doing a simple calculation to reduce the reboiler's heat input, based on the fact that the reboiler dT during relief is lower than the dT used in its design. That, of course, means that the reboiler will be transferring heat at a lower rate. It's a simple calc. Just calculate a "corrected" LMTD using the relieving temp as the cold side outlet temp. Then multiply the reboiler Q by the ratio of the "corrected" Q over the uncorrected Q. This assumes that the U value is constant, and that's a good assumption

Loss of reflux triggers a domino effect of other events which ultimately lead to the overpressure. The specifics of those other events can't be generalized for all distillation columns. You'll need to assess this based on the specific design and control details for this particular column. Once you've laid out that sequence of events, following loss of reflux for this particular column, the answer to your question will be clearly revealed to you. BTW, I'm assume this scenarion is based on loss of reflux flow while the column is running normally. Run this scenario through your mind and lay out the sequence of events following (and caused by) loss of reflux. Here's a typical sequence of events for this scenario: 1.Loss of reflux pump (Note that this could actually be 2 independent failures if there is a spare pump that's programmed to auto-start. If so, consider whether or not this is even a credible scenario.) 2.Without the reflux pumps, the reflux drum will overfill and back up into the condenser, stopping the condenser from operating. 3.The column then pressures up once the condenser stops, but whether or not this happens, and how quickly it happens is a function of the control scheme for the column. Commonly, the steam flow to the reboiler is programmed to maintain a constant temperature at some location in the column. If the reflux pumps and the condenser are out of operation, then the temperature profile in the column will start to rise. If the column uses the control scheme mentioned above, the reboiler controls will respond by cutting back on steam flow. Unless there's a min-flow setting for the steam (usually there's not) then the steam valve might eventually close completely. a.Be aware that even if the reboiler controls respond by shutting off the steam, you could still reach the relief pressure due to the residual heat in the reboiler and/or slow response of the steam valve controller. A conservative assumption would be that the reboiler continues to operate at some low level of heat input. 4.Review this sequence with your client and revise if as necessary for this particular column design and control scheme. 5.After laying out the specific details for this particular scenario, as described above, the relieving flow is simply the resulting amount of boil-up from the reboiler. That said, I don't often run the calc's for this particular scenario, because there's another

scenario on my list that's nearly identical in its effect, and it's equally as bad or worse than this one. I'm talking about a start-up scenario in which the column is being started while someone has failed to properly line up water flow to the condenser. Let me also mention two footnotes to this discussion. (1) Before doing anything, check to be sure that the reboiler can actually overpressure the column, even without the condenser running. It not too uncommon to find that the reboiler will pinche out before the process side is at relieving pressure. In other words, the saturation temp of the steam, at supply pressure, might not be high enough to boil the process fluid at the relieving pressure. (2) You can usually make a significant reduction in the heat input for this scenario by doing a simple calculation to reduce the reboiler's heat input, based on the fact that the reboiler dT during relief is lower than the dT used in its design. That, of course, means that the reboiler will be transferring heat at a lower rate. It's a simple calc. Just calculate a "corrected" LMTD using the relieving temp as the cold side outlet temp. Then multiply the reboiler Q by the ratio of the "corrected" Q over the uncorrected Q. This assumes that the U value is constant, and that's a good assumption.

Loss of reflux triggers a domino effect of other events which ultimately lead to the overpressure. The specifics of those other events can't be generalized for all distillation columns. You'll need to assess this based on the specific design and control details for this particular column. Once you've laid out that sequence of events, following loss of reflux for this particular column, the answer to your question will be clearly revealed to you. BTW, I'm assume this scenarion is based on loss of reflux flow while the column is running normally. Run this scenario through your mind and lay out the sequence of events following (and caused by) loss of reflux. Here's a typical sequence of events for this scenario: 1.Loss of reflux pump (Note that this could actually be 2 independent failures if there is a spare pump that's programmed to auto-start. If so, consider whether or not this is even a credible scenario.) 2.Without the reflux pumps, the reflux drum will overfill and back up into the condenser, stopping the condenser from operating. 3.The column then pressures up once the condenser stops, but whether or not this happens, and how quickly it happens is a function of the control scheme for the column. Commonly, the steam flow to the reboiler is programmed to maintain a constant temperature at some location in the column. If the reflux pumps and the condenser are out of operation, then the temperature profile in the column will start to rise. If the column uses the control scheme mentioned above, the reboiler controls will respond by cutting back on steam flow. Unless there's a min-flow setting for the steam (usually there's not) then the steam valve might eventually close completely. a.Be aware that even if the reboiler controls respond by shutting off the steam, you could still reach the relief pressure due to the residual heat in the reboiler and/or slow response of the steam valve controller. A conservative assumption would be that the reboiler continues to operate at some low level of heat input. 4.Review this sequence with your client and revise if as necessary for this particular column design and control scheme. 5.After laying out the specific details for this particular scenario, as described above, the

relieving flow is simply the resulting amount of boil-up from the reboiler. That said, I don't often run the calc's for this particular scenario, because there's another scenario on my list that's nearly identical in its effect, and it's equally as bad or worse than this one. I'm talking about a start-up scenario in which the column is being started while someone has failed to properly line up water flow to the condenser. Let me also mention two footnotes to this discussion. (1) Before doing anything, check to be sure that the reboiler can actually overpressure the column, even without the condenser running. It not too uncommon to find that the reboiler will pinche out before the process side is at relieving pressure. In other words, the saturation temp of the steam, at supply pressure, might not be high enough to boil the process fluid at the relieving pressure. (2) You can usually make a significant reduction in the heat input for this scenario by doing a simple calculation to reduce the reboiler's heat input, based on the fact that the reboiler dT during relief is lower than the dT used in its design. That, of course, means that the reboiler will be transferring heat at a lower rate. It's a simple calc. Just calculate a "corrected" LMTD using the relieving temp as the cold side outlet temp. Then multiply the reboiler Q by the ratio of the "corrected" Q over the uncorrected Q. This assumes that the U value is constant, and that's a good assumption. Three quick remarks: 3) Favourable instrument response is not taken into account per API STD 521, assume control valves stay in their normal position. Be very carefull with taking credit for a temperature pinch (see below) and in my opinion a clean U should be used. For more information about (hand) calculation of the relief load of a tower: "Distillation column relief loads part 1 & 2", P.L. Nezami, Hydrocarbon Processing Magazine, April 2008 "Tower pressure relief calculation", S. Rahimi Mofrad, Hydrocarbon Processing Magazine, September 2008 "Avoid common mistakes in sizing distillation safety valves", M. Bradford, D.G. Durrett, Chemical Engineering Magazine, July 1984 "How to size relief valves", A. Mukerji, Chemical Engineering Magazine, June 1980