ASME B31.1-2024 Safety Valve Reaction Forces — FAQ (Nonmandatory Appendix II)

Question 1

What does B31.1 actually require for safety valve discharge loads?

Accepted Answer

Three mandatory things and one nonmandatory method. ¶101.5.5 requires the piping to be designed, arranged and supported to withstand the reaction forces from pressure and momentum effects during normal operation and anticipated transients — a valve lift is exactly that transient. ¶122.6 requires pressure relief piping to be supported to sustain the reaction forces, and ¶122.6.2 adds the arrangement rules: no stop valve between the device and the point of discharge, a discharge line whose sectional area is not less than the full area of the valve outlets feeding it, a run as short and straight as possible and arranged to avoid undue stress on the valve, drainage of both the discharge line and the space above the valve seat, and a discharge that does not impinge on other piping or equipment and is directed away from platforms and areas used by personnel. ¶104.8.2 then checks the stress from the reaction as an occasional load against k·Sh. The method — how to compute the force, the moment and the dynamic amplification — is Nonmandatory Appendix II, which ¶107.8.4 points to. Nonmandatory means the method is guidance you may replace with a better one; the requirement to design for the load is not optional.

Question 2

How is the open-discharge reaction force calculated?

Accepted Answer

Appendix II-2.3.1.1 gives the steady-state reaction at the discharge elbow exit as the momentum term plus the pressure term: mass flow rate times exit velocity over gc, plus the exit static pressure minus atmospheric pressure times the exit flow area. The mass flow rate is the relieving capacity stamped on the valve multiplied by 1.11 — the stamped capacity is a certified minimum, and the actual valve flows more. The exit velocity and pressure come from the II-2.2.1 method: an energy balance on the stagnation enthalpy at the valve inlet using the steam-condition constants a and b from Table II-2.2.1-1, which has separate rows for wet steam below 90 % quality, saturated steam to 1,000 psia and superheated steam to 2,000 psia, then a Fanno-line correction from Figure II-2.2.1-2 entered with the friction term f·Σ(L/D) of the elbow and any pipe between the valve and the exit, using a specific heat ratio near 1.3 for superheated and 1.1 for saturated steam. The open-discharge rules assume the compact geometry of Figure II-1.2-2: the horizontal distance Lo from the inlet-pipe centerline to the discharge exit is limited to four discharge-pipe diameters and the rise m of the discharge elbow to six, and II-1.2 puts the burden on the designer to confirm the rules still apply outside those limits. The PiperPSV reaction card runs this chain with the constants embedded and cited by row.

Question 3

What is the dynamic load factor, and can I just use 1.1?

Accepted Answer

The DLF is the ratio of the peak dynamic response to the response the same load would produce if applied statically, and it exists because the valve pops rather than eases open: II-3.5.1.3 defines it that way and notes that for a single-degree-of-freedom system under a single ramp load the value lies between one and two, depending on the ratio of the valve opening time to the natural period of the installation. The appendix gives a period estimate for the case where the run pipe is rigidly supported, so the valve, its inlet pipe and attachments can be idealized as a cantilevered mass, then reads the DLF from Figure II-3.5.1.3-2 against the opening-time ratio. The floor is explicit: the DLF shall never be taken less than 1.1, and any less conservative value than the figure gives must be justified by calculation or test. So 1.1 is not a default; it is the best case, reached only by a stiff installation with a slow-opening valve. A tall inlet riser carrying a heavy valve has a long period, the opening time becomes short by comparison, and the DLF climbs toward 2. That is also why the II-3.5.1.5 time-history analysis is described as the more accurate route: the appendix method is a bounded approximation, and the closed-discharge case cannot use it at all.

Question 4

What moment do I check at the branch connection under the valve?

Accepted Answer

Not less than the product DLF · F1 · Lo — the reaction force after dynamic amplification, times the moment arm from the discharge elbow to the point being analyzed — which II-3.5.1.1 sets as the floor for the reaction moment used in the II-4.2 stress computation at the branch connection below the valve. The same paragraph requires the reaction and moment effects on the header, the supports and any vessel nozzle to be considered for each valve and for combinations of valves blowing, and II-3.5.1.2 explains why the combinations matter: in a multi-valve header each valve lifts at a different time, not every valve is needed in every transient, and several force combinations can exist. The practical mitigation the appendix suggests is to vary the discharge directions of adjacent valves so their reactions partly oppose rather than add. The moment at the branch goes into eq. (16) of Figure 104.8-1 with the sustained-plus-occasional indices, and because the reaction is a dynamic load whose sign the simple method does not deliver, ¶104.8.4(b) applies: combine it with the sustained moments in the most conservative way. Do not forget the header itself — the moment arm to the nearest header support can be longer than Lo, and that is where the k·Sh case most often governs.

Question 5

What are the stress criteria for the installation?

Accepted Answer

Appendix II-4.1 restates the ¶104.8 checks for every critical point of the installation in three lines: longitudinal pressure stress plus the sustained-load stress must not exceed Sh; that sum plus the occasional-load stress — the valve reaction is the occasional load — must not exceed k·Sh; and the pressure stress plus sustained stress plus the thermal expansion stress must not exceed SA + Sh, which is the liberal-allowance form of the displacement check. The k is the ¶104.8.2 duration factor: 1.2 where the events total no more than one hour at a time and 80 hours a year, 1.15 for up to eight hours at a time and 800 hours a year — and a valve that lifts a few times a year fits the first tier, while a valve that chatters through every startup may not. II-3.3 allows the deadweight moment to be assumed as 1,500 times the section modulus in inch-pounds where the supports meet ¶121, but requires it to be calculated wherever stresses exceed 90 % of the allowable in the eq. (15) and (16) checks. The thermal side has its own trap under II-3.2 and II-2.1: the installation sees several operating modes, including the discharge piping hot after a lift, and the design condition must envelope all of them.

Question 6

What changes for a closed discharge system?

Accepted Answer

Almost everything about the analysis, and the appendix says so plainly. Under steady flow the forces in a closed system are self-equilibrated and do not create significant bending moments except at the point of final discharge (II-2.3.2), so the open-discharge force equation does not describe the load that matters. What matters is the transient: when the valve opens into a long, air-filled discharge line, the pressure wave steepens as it travels and can become a shock before the exit, and the first milliseconds after lift produce momentary unbalanced forces on every straight run between changes of direction. II-2.2.2 therefore recommends a discharge-pipe design pressure of at least twice the steady-state operating pressure, II-2.3.2 requires the designer to compute the magnitude of the transient loads and evaluate their effects, and II-3.5.2 states that closed systems do not lend themselves to simplified techniques and that a time-history analysis may be required. There is a thermal problem as well: II-2.1.2 warns that thermal expansion and back pressure in a closed discharge can be high enough to make the valve malfunction or leak at the valve or flange, so the expansion analysis must cover the discharge piping hot after a lift. If a water seal is used below the seat, II-3.5.3 adds the slug transient of the seal water clearing ahead of the steam on the first cycle.

Question 7

Does the vent stack need its own check?

Accepted Answer

Yes, on two counts. The vent pipe carries reaction forces of its own — Figure II-2.2.1-3 shows the force at the vent entrance and the force at the vent exit, computed by the same momentum-plus-pressure method — and II-2.3.1.2 requires the vent anchor and restraint system to carry the moments from both forces and the unbalanced vertical and horizontal components, with beveled exits discouraged because they add a lateral force. The second check is blowback. The vent must be sized so that no steam blows back out of the vent entrance around the discharge elbow: the momentum leaving the elbow must exceed the momentum at the vent inlet so that air is educted into the vent rather than steam being pushed out of it, and the appendix asks for a margin, not equality, because indoor installations rely on that eduction to clear the steam vented from the valve body. The pressures and velocities in that inequality are the II-2.2.1 values, and the vent inlet pressure computed there is the highest pressure the stack sees and the basis for its design. The overlap of the elbow into the vent must also be long enough that the opening reaction and thermal movement cannot pull it out — the figure calls that out explicitly.

Question 8

Do I combine the relief reaction with earthquake or wind?

Accepted Answer

B31.1 does not give a combination rule for a relief event coincident with an earthquake — it says only that wind and earthquake need not be considered concurrent with each other (¶101.5.2, ¶101.5.3). Appendix II-2.4 lists earthquake and piping-system vibration among the other mechanical loads to be considered, and II-3.4 places the decision squarely on the design specification: it should state whether the system is designed for earthquake, the magnitude, the plant conditions under which it is assumed to occur, and whether the analysis is equivalent-static or dynamic; absent a specification, the designer decides and documents. The engineering reading most owners adopt is that a design-basis earthquake and a full-capacity relief event are independent rare events not combined at full value, while a relief event and normal operating loads always are — but that is an owner's criterion, not a code sentence, and it belongs in the stress design basis where a reviewer can find it. What the Code is unambiguous about is the multi-valve interaction load of II-2.4(a): more than one valve opening on the same header is a combination you must evaluate, not a coincidence you may neglect.

Paragraph	What it sets
¶101.5.5	Piping shall be designed, arranged and supported to withstand reaction forces from pressure and momentum effects during normal operation and anticipated transients.
¶122.6, ¶122.6.2	Pressure relief piping supported to sustain reaction forces; no stop valve between device and discharge point; discharge area not less than the valve outlet; drainage; discharge directed away from personnel.
¶107.8.4	Points to Nonmandatory Appendix II for the design of safety valve installations.
App. II-2.3.1.1	Open-discharge reaction force at the elbow exit: momentum plus pressure term, with mass flow at 1.11 × stamped capacity; a dynamic load factor is to be applied.
App. II-2.2.1, Table II-2.2.1-1	The exit pressure and velocity method with the steam-condition constants a and b and the Fanno-line figure for f·Σ(L/D).
App. II-3.5.1	Reaction moments; the branch-connection moment never less than DLF · F₁ · L_o; multiple-valve combinations; DLF never less than 1.1.
App. II-2.2.2, II-2.3.2, II-3.5.2	Closed-discharge systems: design pressure at least 2 × steady state; momentary unbalanced forces; time-history analysis may be required.
App. II-4.1	Stress criteria for every critical point of the installation: sustained, sustained-plus-occasional against k·S_h, and the expansion combination.