Re-rating a Two-Phase Separator After a Process Change
A vertical two-phase separator built in 1998 is being debottlenecked. Nameplate data: SA-516 Gr 70 shell, inside diameter 1800 mm, nominal shell thickness 19.05 mm, seamless course with joint efficiency E = 1.0, corrosion allowance 3.0 mm, design 1.6 MPa at 120 C, MDMT -29 C. All nozzles carry Class 150 RF A105 flanges, and the relief valve is set at 1.55 MPa. A UT survey last month gave a minimum measured shell thickness of 16.4 mm in the liquid zone; the same survey five years earlier read 16.6 mm.
Process wants to operate the vessel at 2.2 MPa and 260 C after a catalyst change. The turnaround window is 10 days. The operations manager has asked whether the datasheet and nameplate can simply be revised, since 'the shell has plenty of metal left and the UT shows almost no corrosion.' A management-of-change package and a relief-valve re-set are already drafted, and the valve vendor has quoted a two-week lead time for a new spring.
1. Determine whether the shell can be re-rated to 2.2 MPa at 260 C. Show the governing calculation, state the maximum pressure the existing shell can carry at that temperature, and list the assumptions you would have to verify before relying on the answer.
Read the worked answer
Use the ASME VIII Div. 1 UG-27(c)(1) circumferential formula: t = P R / (S E - 0.6 P) + CA, with R = 900 mm and S = 138 MPa for SA-516 Gr 70 at 260 C (Section II Part D).
Required: t = 2.2 x 900 / (138 - 1.32) = 14.5 mm, plus 3.0 mm corrosion allowance = 17.5 mm. Available measured thickness is 16.4 mm, so the 2.2 MPa re-rate fails on the shell alone.
Reverse the formula with the pressure-carrying metal actually present (16.4 - 3.0 = 13.4 mm): P = S E t / (R + 0.6 t) = 138 x 13.4 / (900 + 8.0) = 2.04 MPa. Round down to 2.0 MPa to keep a margin, and only if every other item checks out.
Assumptions to verify: that the UT minimum is a grid survey and not one isolated pit; that the next inspection interval does not erode the 3 mm allowance; that mill certificates support S at 260 C and the impact-test exemption at -29 C for the actual thickness; that weld and heat-affected zones were scanned; that external pressure, nozzle reinforcement, support loads and the relief case are re-checked at the new condition. A single minimum reading is not a re-rate basis on its own.
2. The operations manager argues the flanges are irrelevant because the pressure rises only 0.6 MPa above the current value. Evaluate that argument.
Read the worked answer
It is wrong, and it is the decisive point. ASME B16.5 ratings are pressure-temperature dependent, not pressure-only. A Class 150 A105 flange is rated about 1.7 MPa at 120 C but falls to roughly 1.1 to 1.2 MPa at 260 C as the allowable stress of the material drops with temperature. The proposed condition of 2.2 MPa at 260 C therefore exceeds the flange rating by close to 90 percent, not by a comfortable margin.
The vessel MAWP is set by the weakest pressure-containing element, so the governing limit here is the flanges, not the shell. Even the 2.0 MPa the shell could tolerate is well above the flange capability at 260 C. Gaskets and bolting also need re-checking: a sheet gasket is unlikely to be suitable, spiral-wound or ring-joint with appropriate bolting would be needed, and bolt relaxation at 260 C affects the joint integrity.
The practical conclusion is that no re-rate to 2.2 MPa at 260 C is possible without replacing the nozzles with Class 300 flanges and re-checking nozzle loads, or without limiting the operating temperature. Any relief-valve set pressure must also be based on the lowest MAWP of any element, not on the shell value.
3. Set out the path you would recommend to the project, including the regulatory and documentation route and what you would refuse to do.
Read the worked answer
Recommend a two-stage response. In the short term, issue an engineering change that caps the operating envelope at the governing MAWP of the weakest element, which is the flange rating at the proposed temperature. The process cannot be released at 2.2 MPa and 260 C. The relief-valve set pressure, the management-of-change file, the P&IDs, the datasheet and the inspection plan all follow from that lower number.
If 2.2 MPa at 260 C is genuinely required, nozzle replacement with Class 300 flanges and matching gaskets and bolting is a turnaround-plus scope, not a 10-day scope; the plan should say so with a realistic duration and a fabrication slot for the spools. All elements are then recalculated: shell, heads, nozzles, external pressure, supports and the relief case.
On the regulatory route, this is an alteration under the NBIC, so it needs an R-stamp holder, revised calculations, a new nameplate, relief-valve re-certification to the new set pressure, and notification to the jurisdiction if the vessel is registered. I would refuse to issue a revised datasheet or a nameplate revision that implies the shell has been re-rated without those steps, and I would record the refusal and the basis in the management-of-change file.