Relief Valve Guidelines for External Fire Case

1. Purpose

This procedure provides guidelines for calculating the fire case relief rate. Topics discussed are: the effective fire height, wetted surface area, insulation credit, heat input, and methods for determining the latent heat of vaporization and the initial rate of liquid release from a PRV located on a normally liquid-full vessel.

2. Table of Contents

1. Purpose 1

2. Table of Contents 1

3. General 1

4. Guidelines for Effective Fire Height 2

4.1 Effective Fire Height in API Standard 521 (Process Equipment) 2

4.2 Effective Fire Height in API Standard 2000 2

4.3 NFPA Code (upon customer request in BEDQ) 3

5. Total Wetted Surface Area 3

5.1 Horizontal Vessels 3

5.2 Non-Trayed Vertical Vessels 3

5.3 Vapor-Filled Vessels 4

5.4 Fractionating Columns 5

5.5 Overhead Receivers 5

5.6 Liquid Full Vessels 5

5.7 Shell-and-Tube Heat Exchangers 7

5.8 Vessel Containing a Low Vapor Pressure Liquid 7

6. Environmental Factor (Fire-proof Insulation Credit) 8

6.1 Customer Request (in PRV section of BEDQ) 8

6.2 HF Alkylation and Detergent Alkylate Process Units 8

6.3 Revamps 9

7. Heat Absorption Equations 9

8. Multiple Equipment Items protected by a Single PRV Service 10

8.1 Load Calculation 10

8.2 Fire Zone 10

9. Latent Heat of Vaporization & Relief Properties 11

9.1 Use of a Process Simulator 11

9.2 Use of API 521 Figure A-1 12

Attachment 1 Work Process for Liquid Discharge through a PRV during an External Fire 14

Attachment 2 Wetted Surface Area of an Exchanger 16

Figure A.1 API Recommended Practice 521 18

3. General

This procedure is based on American Petroleum Institute (API) Standard 521, API Standard 2000, and National Fire Protection Association (NFPA) Code. Inflection Point Engineering normal practice for fire protection is to follow these API standards if applicable, unless instructed by the customer to use NFPA guidelines.

The design equations used to protect equipment exposed to an external fire are based on full-scale tests conducted by the API and various other organizations. The tests determined that the amount of energy transferred to a vessel during an external fire is a function of the wetted surface, the vessel size, the type of vessel insulation and its thickness. Heat transfer equations were developed using this test data, but were later modified by the API to be more representative of refinery operation. The API and NFPA approaches to vessel protection during an external fire are summarized below.

The methods and equations represented in this procedure have been embodied in Tool .

4. Guidelines for Effective Fire Height

The effective fire height provides an upper limit to the wetted surface area used to calculate heat input during an external fire. It is referenced to any surface upon which a substantial spill of flammable material can accumulate, whether that surface is at grade or on a platform. Although radiant heat transfer is not limited by height, the effective fire height reflects engineering judgment regarding the severity of a fire event for which relief protection is considered reasonable. This approach also represents the consensus view of the member companies of the API.

4.1 Effective Fire Height in API Standard 521 (Process Equipment)

When determining the relieving rate caused by an external fire, apply a 25 ft fire height limit to calculate the wetted surface for a vessel unless special circumstances apply, as discussed in the rest of this section and in Section 5.5. The fire height limit assumes the following conditions apply:

• Flammable liquid does not accumulate under the protected equipment because the grade is sloped
• Fire fighting equipment and trained personnel are readily available to limit the extent of the fire
• Protected equipment is located within the process area

If one or more of these conditions are not met, use API Standard 2000, as discussed in paragraph 4.2.

4.2 Effective Fire Height in API Standard 2000

Apply a 30 ft fire height limit to wetted surface for:

• Storage tanks
• Equipment located remotely from the process area
• Equipment located in an area that is curbed to contain flammable liquid spills

Since NFPA uses the heat input equation developed by the API for Standard 2000, select the NFPA option in Tool and enter a 30 ft effective fire height.

Note: The fire height limit given in API Standard 2000 does not apply to HF Alkylation Units. Although the acid section in this unit is curbed, adequate drainage is provided to limit pooling of flammable liquid.

4.3 NFPA Code (upon customer request in BEDQ)

Apply the NFPA Code with a 35 ft height limit to calculate the wetted surface when the customer requests Inflection Point Engineering to design according to NFPA Code.

5. Total Wetted Surface Area

Wetted surface area applies only to that portion of the vessel wall that contacts internal liquid and is below the effective fire height as defined by the API and NFPA. Although the entire vessel is subjected to radiant and convective heat transfer during a fire case, only the vessel wall contacting liquid will transfer significant heat to the contents. Calculate the wetted surface area of the equipment and increase the area to account for associated level instruments and piping. Note: Inflection Point Engineering default is to increase the wetted area by 15%.

5.1 Horizontal Vessels

If the liquid level is controlled, base the wetted surface area on the normal operating liquid level. If the level is not controlled, assume the liquid level is at 80% of the float or gauge glass range. If the vessel has a water boot, add the wetted surface area of the water boot to the main vessel’s wetted surface area and assume that the entire contents are hydrocarbon.

If the horizontal vessel can be isolated from other equipment, the minimum wetted surface area will correspond to a liquid level equal to 30% of the level determined by the previous paragraph. This provision applies specifically to vessels elevated near or entirely above the effective fire height, including elevated overhead receivers (section 5.5).

5.2 Non-Trayed Vertical Vessels

If any portion of the bottom head is below the effective fire height, include the entire head in the wetted surface area calculation. For 2:1 elliptical heads, the head extends 1/4 of the vessel diameter below the vessel tangent line. If the vessel is liquid full and the top head is located below the fire height, include the top head as well.

If the liquid level is controlled, base the wetted surface area on the normal operating liquid level. If the liquid level is not controlled, assume the liquid level is 80% of the float or gauge glass range.

5.3 Vapor-Filled Vessels

If a liquid level is not anticipated in a vessel during normal operation, but a liquid level recorder or gauge glass is provided, e.g. a compressor knockout drum, assume the liquid level is 80% of the maximum float or gauge glass level. In a separate calculation, determine vapor expansion due to external fire assuming there is no liquid in the vessel. Use Tool to determine the vapor expansion rate. Report the relieving rate requiring the largest orifice area on the 807 Project Specification.

NOTE: T-807-05 requires the operating pressure as an input. To be conservative, use the lowest expected operating pressure.

For vessels not normally containing a liquid, but having a level recorder or gauge glass, use one of the following alternatives to characterize the liquid to be vaporized during the fire case.

• If there is a vapor/liquid separator immediately upstream of the knockout drum, consider using the composition of liquid in the separator.
• Alternatively, cool the vapor contents in the vessel to the lowest ambient (winterizing) temperature to see if the liquid dew point is reached. If so, use the composition of the dew point liquid.

If a liquid phase is not created at the lowest ambient temperature, use T-807-05 to determine the vapor expansion rate. For example, condensation typically will not occur in a compressor suction drum processing PSA grade hydrogen, even at low temperatures.

Note that a vapor-filled vessel may rupture due to over-temperature during an external fire, even though its relief valve limits over-pressure. Verify that vessel pressure can be reduced using normal process instrumentation. If not, and especially if the vessel is large and/or operated at high pressure, consider using one or more of the design options listed below to prevent loss of containment.

• Install fire-proof insulation on the vessel to increase the time required for over-temperature.
• Provide a depressuring system to reduce the stress on the vessel wall.
• Elevate the vessel above the fire height limit, i.e. remove it from the fire zone.

If the project team decides it is necessary, discuss the risk of over temperature and the various design options with the customer. (If none of the above options are acceptable, the customer may decide instead to install a water deluge system to cool the vessel wall.)

5.4 Fractionating Columns

Reference Procedure "

5.5 Overhead Receivers

If a relief valve is located on the overhead vapor line and it protects both the fractionator and overhead receiver, determine wetted surface area and heat input separately. Add the vapor rate from both the fractionator and the overhead receiver to obtain the total fire case load.

If the overhead receiver can be isolated from the fractionator PRV, install a separate PRV on the receiver. Determine the wetted area according to section 5.1.

5.6 Liquid Full Vessels

If a vessel is liquid full, calculate the wetted surface for that part of the vessel located below the fire height limit. Determine heat input to the vessel, calculate a vaporization rate, and base the required orifice area on relieving vapor only, i.e. assume there is complete vapor-liquid disengagement.

Initially, however, liquid will pass through the relief valve due to thermal expansion, vapor displacement of liquid, and/or incomplete vapor/liquid disengagement. Although the liquid displacement rate is not used to size the PRV, the estimated initial liquid flow rate through the PRV and the total amount of liquid both impact flare system design. Rate the relief valve for liquid service using its effective orifice area and the liquid sizing equation. (For liquid rating methodology, see .) In addition, determine the total quantity of liquid displaced into the relief system. Report both the initial liquid flow rate and the total quantity of liquid displaced in a note on the PRV specification.

NOTE: The rules of this section assume that the PRV is located at the top, or near the top, of the protected equipment. If there is normally a process liquid holdup in a connected vessel above the PRV, the all-vapor basis for sizing the PRV may not be valid. If the PRV cannot be located near the highest point in the system, sizing for liquid relief may be required. Consult the PRV Specialist for guidance.

a. Total Quantity Displaced from a Liquid-Filled, Vertical Vessel

Assume that all liquid within 2 vertical feet of the relief nozzle on the vessel will be carried over with the vapor relief stream into the relief header.

If the PRV is located on the top head, the volume of liquid carried over is equal to the top head plus the cylindrical section below the top tangent needed to provide the 2 ft dimension.

Vtotal = Vhead + Vcyl

• i.e. Vtotal , ft3 = 0.132 (D)3 + (/4)(D)2 * (2 - D/4)

where D = vessel diameter, ft.

For example, the total liquid volume displaced through a PRV installed on the top of a 6 ft diameter vessel is:

Vtotal = 0.132 (6)3 + (/4)(6)2 * (2 - 6/4) = 42.64 ft3

For vertical vessels with a diameter greater than 8 ft and with a PRV mounted on the top head, the liquid carryover volume equals the volume of the top head, i.e. Vhead = 0.132(D3). This automatically provides at least two feet of vertical height.

If the PRV is located on the side of a liquid-filled vertical vessel, the total amount of displaced liquid includes the vessel head plus the cylindrical volume needed to provide 2 ft vertical distance below the relief nozzle:

Vtotal , ft3 = 0.132 D3 + (/4)(D)2 * (hnoz +2)

Note: hnoz is the distance in feet between the top tangent line and the vessel nozzle used for the PRV connection.

b. Total Quantity Displaced from a Liquid-Filled, Horizontal Vessel

Assume that all liquid within 2 ft of the relief valve inlet will be carried into the relief header by the fire case vapor. If the PRV is located at the top of the vessel or on the process outlet line, the volume of liquid carried over is equal to the vessel tangent length times the cross sectional area of a circular segment having a vertical height of 2 ft. As discussed above, report this total quantity in a note on the PRV specification.

NOTE: The two-foot basis is very conservative for horizontal vessels with diameter less than four feet. For such vessels, the liquid to be displaced is half the vessel volume.

c. Displaced Hydrocarbon Liquid Flashes in Relief Header

If light hydrocarbon liquid is displaced through the relief valve, determine if flashing will occur in the discharge piping from the relief valve. Assume that the liquid is at normal operating temperature upstream of the valve, and that it flashes adiabatically to relief header superimposed back pressure. If flashing does occur, review the metallurgy requirements based on the adiabatic temperature. Report the vapor and liquid flow rates in the header and total quantity of displaced liquid remaining after the flash in a note on the PRV specification. See .

5.7 Shell-and-Tube Heat Exchangers

The protection philosophy for exchangers is covered in Procedure ". Refer to this procedure to determine whether the fire case will apply to an exchanger service.

The wetted surface area for exchangers is dependent on the exchanger type, the fluid phase distribution, whether or not it is free draining, and its elevation.

• Most shell & tube exchangers are located near grade, which places them within the fire zone. If the exchanger is partially below the effective fire height, estimate the portion of the exposed wetted area that is below the limiting height.

• If all-liquid or two-phase material normally flows through an exchanger, assume that the shell wall (or channel wall) is completely wetted. Exceptions include: free draining overhead condensers (zero wetted area) and kettle-type exchangers (area is based on the normal level).

When the fire case load for exposed heat exchangers is included in the relief rate calculations, the formulas of should be used.

5.8 Vessel Containing a Low Vapor Pressure Liquid

The wetted area of a vessel containing a low vapor pressure liquid is determined per sections 5.1 – 5.7. If the bubble point temperature of the liquid is above 795 F (424 C) at accumulated pressure, Inflection Point Engineering assumes that thermal cracking will become significant. Relief properties and latent heat of vaporization are determined as described in Section 8.2.

When the bubble point temperature is above 795 F (424 C), the vessel could still rupture due to over-temperature during an external fire, even though a PRV is provided for over-pressure protection. In this case, verify that vessel pressure can be reduced using the normal process instrumentation, or consider adding protective measures as described in Section 5.3 for a vapor-filled vessel.

6. Environmental Factor (Fire-proof Insulation Credit)

Generally, Inflection Point Engineering does not take an insulation credit to reduce heat input to a vessel during a fire case. The API Standard warns that water from the fire fighting equipment may dislodge the insulation and render it ineffective. Since Inflection Point Engineering does not specify how to install the insulation, we assume it will be dislodged during the fire. Although this assumption is conservative, the cost of the flare header and/or the relief valve is not significantly affected.

There are circumstances where an insulation credit is permissible as discussed below. In such cases, include the following note in the relief valve specification as applicable:

“Relief rate for external fire takes credit for fire-proof insulation. Contractor to confirm that insulation type and its installation meet the requirements of API Std 521.”

6.1 Customer Request (in PRV section of BEDQ)

A customer may instruct Inflection Point Engineering to take credit for fire-proof insulation. The type of installation and its thickness determines the amount of credit that may be taken. Unless otherwise instructed by the customer, use Table 1 to determine the appropriate environmental factor for use in the heat input equation.

Table 1.

Fire-Proof Insulation Credit (1)

Insulation Thickness	Environment Factor
1 inch	0.3 (2)
2 inches	0.15
3 inches	0.10
4 inches and higher	0.075 (3)

(1) Insulation must be one of the types listed in Table 7 of API Std 521 (2007). Foam type insulation is not acceptable.

(2) Use an environmental factor = 0.30 when the insulation thickness is unknown.

(3) Do not use an environmental factor lower than 0.075.

6.2 HF Alkylation and Detergent Alkylate Process Units

Take an insulation credit for HF Alkylation and Detergent Alkylate Process Units because the fire relief load sets the design of the neutralization section. For these units, include a special note in “Project Specification 907 – Insulation” requiring specific types of sheathing and/or banding of the insulation. Consult the Technology Specialist for applicable notes to include with the relief valve specification.

6.3 Revamps

If necessary, take a fire-proof insulation credit in revamp situations in order to reuse an existing relief valve and/or vessel nozzle. Use Table 1 to determine the appropriate environmental factor.

7. Heat Absorption Equations

Inflection Point Engineering uses heat absorption equations developed by the American Petroleum Institute, unless the customer requests that the more conservative NFPA code be followed. If the customer specifies an alternate, non-API fire height in the BEDQ, Inflection Point Engineering will use that value as an override, provided it gives a larger relieving rate than the API equations. See the above discussion regarding effective fire heights to be used with the different equations.

Note: The design code determines which heat input equation to use.

• API Std 521 assumes that flammable liquid will not pool under the vessel. Heat input based on this recommended practice is given by:

Q = 21,000 F A0.82

• API Std 2000 assumes that the equipment is located in a diked or curbed area, or that it is remote from the process area. Heat input based on this standard is given by:

Q = 34,500 F A0.82

• NFPA code heat input equation is the same as for API Std 2000, i.e.:

Q = 34,500 F A0.82

The variables in all three design equations are defined as:

Q = heat absorption, Btu/h

F = environmental factor = 1.000 for bare metal surface (not insulated)

A = wetted surface area, square feet

Note: The area exposure exponent, 0.82 was empirically derived using multiple fire test data. It “discounts” the wetted surface area for large vessels because they are not likely to be completely engulfed by flames. [For piping and e.g. double-pipe heat exchangers that we may reasonably expect to be totally engulfed by a fire, this area exposure exponent becomes 1.00].

8. Multiple Equipment Items protected by a Single PRV Service

8.1 Load Calculation

Calculate the fire-case vapor load from each vessel and exchanger separately and then add them together. Liquid in various equipment items does not have the same composition, and the vapor rate and its properties must be consistent with that liquid composition. Calculate the molecular weight of the vapor exactly, based on the MW of each contributor to the total fire vapor rate. Calculate the relief vapor temperature, Z, and Cp/Cv as the weighted sum of the values for each contributor, weighted according to the mass vapor rate.

8.2 Fire Zone

For Schedule A projects, Inflection Point Engineering does not consider the separation between equipment when determining which items are exposed to fire simultaneously. All items with a free path to a common PRV service are included in the fire case.

For projects requiring consideration of fire zones (e.g. revamps or other customer request), the following guidance is provided.

Inflection Point Engineering follows API (Recommended Practice 520 / Standard 521 and Standard 2000) for the design and specification of Pressure Relief Devices. In the absence of any instruction from the customer, Inflection Point Engineering would ensure the design was conservative by using API’s upper limit of 5,000 sq. feet (460 sq. meters) for the fire zone. This area allows for less than adequate drainage or fire-fighting facilities.

API does not recommend any particular shape of the fire zone. However, it is often thought to be a "fire circle". For an area of 5,000 sq. feet this fire circle will have a diameter about 80 feet (24.4 meters). If, when moving a fire zone around on a plot plan, an equipment item has at least the majority of its perimeter inside the zone, it should be 100% included (provided it is not above the effective fire height).

ONLY on customer instruction will Inflection Point Engineering use a fire zone area of 2,500 sq. feet (230 sq. meters), which is equivalent to a fire circle diameter of 56.4 feet (17.2 meters).

(The lower limit for the area of a fire zone could be justified if there are both:

• adequate liquid drainage in the area such that liquid cannot collect and spread across a wide area, and
• adequate fire fighting facilities such that the fire could be fought effectively and brought quickly under control.)

9. Latent Heat of Vaporization & Relief Properties

The relief load tools calculate the relieving rate using values of latent heat of vaporization. Throughout the duration of every relief event, liquid compositions and latent heats of vaporization are typically changing continuously. To avoid a transient analysis, the latent heats of vaporization are determined using ‘point’ vaporization values for the bulk liquid streams at accumulated pressure. Latent heats can be obtained from a process simulator (this is the preferred Inflection Point Engineering approach if it works) or from

Vessels normally containing both hydrocarbon and free water phases should be evaluated by assuming the vaporizing liquid is hydrocarbon. Select a stream whose composition reflects the majority of the liquid present (normally the bottoms stream).

9.1 Use of a Process Simulator

For multi-component liquids, the latent heat of vaporization is best determined using a process simulator, i.e. P9.8, UniSim, , or HYSYS. Although the commercial simulators are internally consistent, they determine the specific enthalpy of liquid and vapor differently from P9.8. These simulators incorporate heats of formation in their enthalpy calculations, making it difficult to isolate the latent heat effects from the heats of formation. Accordingly, different thermodynamic operations are required for each simulator to obtain consistent latent heat values.

• If P9.8 is used as the process simulator, vaporize 0.30 wt fraction of the liquid, and subtract the specific enthalpy of the liquid phase from the specific enthalpy of the vapor phase to obtain the latent heat of vaporization.
• When Aspen or UniSim is used for the process simulator, use tool ” to determine the relieving temperature, vapor compressibility, vapor and liquid specific enthalpies, Cp/Cv ratio, and latent heat of vaporization.
• When HYSYS is used for the process simulator, additional flash calculations are required to obtain the thermodynamic data required for relief conditions. Use the procedure given in TEC-807-01, section 7 to calculate the latent heat of vaporization and vapor properties.

Regardless of which process simulator is used, if the process liquid has an extremely high boiling point, 30% vaporization may return a temperature over 795°F (424°C). In such cases, determine the bubble-point temperature of the liquid at accumulated pressure. If the bubble point is less than or equal to 795°F, use the bubble point temperature and calculate the latent heat at this condition. If the bubble point is still above 795°F, use Figure A.1 to obtain the latent heat.

Similarly, if the process simulator fails to converge, or if the relief flash returns a latent heat of vaporization less than 50 BTU/lb, use a latent heat value of 50 BTU/lb in the normal vaporization model to represent the expansion rate of the dense fluid, i.e. fluid above its critical point. Treat the relieving fluid as vapor at the critical temperature. Use Z = 0.7 and k = 1.0.

9.2 Use of API 521 Figure A-1

[Note: a copy of Figure A.1 from API RP 521, is attached. If it does not print out clearly, change to a PostScript printer. The same figure appears in API 521 5th edition (2007).]

When there is simulation data that results in a temperature greater than 795°F (424 C), Inflection Point Engineering assumes that thermal cracking will occur. Use Figure A.1 to determine latent heat and properties as follows:

• Plot the intersection of the simulation temperature and liquid molecular weight. (If necessary, extrapolate a line of constant MW out to the simulation temperature.)
• Extend a line of constant pressure from the intersection back to the 795°F temperature line. Find the MW curve which passes through this point (extrapolating if necessary). Use this MW for the cracked material. Note that the pressure may not correspond to the fluid vapor pressure.

When there is no simulation data, use Figure A.1 as follows:

• Plot the intersection of accumulated pressure and liquid molecular weight. If the intersection falls within the figure (40 MW 280, Pacc 10 psia, latent heat 50 Btu/lb), the latent heat and temperature can be read directly from Figure A.1. [Although the molecular weight data shown in this figure is based on normal paraffins, this method provides a conservative estimate of the latent heat of vaporization for hydrocarbon mixtures.]
• If the intersection of accumulated pressure and molecular weight lies above the 50 Btu/lb latent heat curve but below 795°F, use 50 Btu/lb for the latent heat and use the temperature from the intersection.

• If the intersection of accumulated pressure and molecular weight lies above 795°F, Inflection Point Engineering assumes that thermal cracking will be significant. In that case, the intersection of accumulated pressure and 795 °F determines the MW of the cracked material and the latent heat value. (If necessary, extrapolate a line of constant MW through the intersection of 795°F and the relieving [accumulated] pressure).
• The values of MW and latent heat at 795°F required in this step can be calculated from the following equations as an alternative to the graphical interpolation.

Compressibility Factor and k = Cp/Cv

• When Figure A.1 is used, conservative values of Z = 1 and k = 1.0 should be used if no better information is available.

Attachment 1 Work Process for Liquid Discharge through a PRV during an External Fire

If the PRV protects a liquid-full vessel during an external fire, rate the valve for liquid flow assuming the PRV is full open at 121% overpressure, and that the liquid does NOT flash. [The non-flashing assumption assures using the maximum mass flow for design of the Knockout (KO) Drum.] Once the liquid passes through the PRV, it will flash adiabatically at the relief header pressure. Use a process simulator to determine the degree of flashing, and publish the results in a note on the 807 Project Specification. (Generalized notes are given below.) The Contractor will use this information for design of the relief piping & supports, and the knockout drum.

Procedure

1. Create a dummy ‘External Fire’, non-flashing liquid case in the 807 Sizing & Selection Tool, separate from the true External Fire, vapor sizing case.

2. Input the normal operating liquid temperature, specific gravity, and viscosity, along with a starting guess flow of 100,000 lb/h for the liquid rate (vapor rate = 0).

3. Go to the orifice sizing screen under Main Menu button 3, and note the required orifice area for the dummy liquid case and the total installed area for the PRV service (note: do not change the selected orifice size or number of online PRV’s).

4. Calculate the total flow rate as the starting guess flow (100,000 lb/h) multiplied by the total ratio of installed orifice area divided by the required orifice area for the dummy non-flashing liquid case.

5. Use a process simulator to create a stream having the flow resulting from Step 4 and the composition of the normal vessel liquid. Initialize this stream at normal vessel operating temperature and pressure and flash it adiabatically to the SUPERIMPOSED back-pressure. [This is the lowest pressure in the system, and combined with normal operating temperature, will result in the lowest potential auto-chill temperature.]

6. Check the auto-chill temperature against the lower limit (-20°F) for carbon steel. If the flash temperature is less than -20°F, specify 316 SS (or other low temperature alloy) for the PRV body and bonnet. If the PRV tailpipe is specified, it should be impact-tested carbon steel between -20°F and -50°F, and stainless steel below -50°F.

7. Eliminate the dummy liquid sizing case: change its cause to ‘None’. It is normally not necessary to delete the data for the dummy sizing case.

8. If less than 0.1 wt% flashing occurs at the superimposed back pressure, insert project specific data in the note below and publish it in the 807 Project Specification.

PRV installed on liquid-filled vessel. Liquid initially displaced at [??? vols/time]. Total amount of liquid displaced during external fire case is [??? vol units].

9. If 0.1 wt% flashing or more occurs at superimposed back pressure, insert project specific data in the note below and publish it in the 807 Project Specification.

PRV installed on liquid-filled vessel. Liquid initially displaced at [??? Vols/time] and will partially flash through the PRV. The discharge from the PRV will be at [?? ° temp unit] and will consist of [??? Mass/times] of vapor with a molecular weight of [???], and [??? Vols/time] of liquid with a specific gravity of [???]. Total amount of liquid displaced during an external fire is [??? Vol units].

Attachment 2 Wetted Surface Area of an Exchanger

Heat input to an exchanger during an external fire must be considered per ASME code. However, Inflection Point Engineering does not know the exact dimensions for the exchangers at the basic engineering stage, and we estimate their size based on the exchanger sizing tool, TZ-400-01 "Exchanger Calcs.xls". Although Inflection Point Engineering issues preliminary fire case loads, the detailed design contractor is ultimately responsible for all relief loads.

The surface area of an exchanger is estimated using the tube length and bundle diameter obtained from the exchanger calc tool. Assumptions built into the wetted surface area calculations are:

• surface area of the exchanger is completely wetted by the process liquid and/or two-phase mixture, except as noted for a kettle type exchanger.
• shell diameter is same as the tube bundle diameter.
• shell length is same as tube length, plus head if applicable.
• exchanger head includes the channel and channel cover.
• channel length is a simple function of the bundle diameter
• channel cover is 2:1 elliptical head regardless of TEMA type.

Definitions:

L tube = nominal tube length, ft.

D shell = nominal tube bundle diameter, ft.

The wetted surface area of the shell is determined by:

SA shell = D shell * L tube , plus head if applicable, as noted below.

The wetted surface area of one exchanger head (channel and cover) is given below.

SA head = D shell * L head + 1.089 * (D shell)2

The length of an exchanger head is considered to be a function of the shell diameter.

L head = 0.5 ft + 0.4 * D shell

The wetted surface areas for the shell side and tube side of an exchanger are summarized in Table A1.

Table A1. Wetted Surface Area for Various Exchanger Types[1]	Table A1. Wetted Surface Area for Various Exchanger Types[1]	Table A1. Wetted Surface Area for Various Exchanger Types[1]
Exchanger Type	Shell side	Tube side
S&T [straight tubes or multi-tube]	SA shell	2 * SA head
S&T ['U' tubes][2]	SA shell + SA head	SA head
Kettle [3]		0.5 * SA head
Stab-In [4]	SA stab in = D shell * 0.5 D shell	SA head
Double-Pipe [5,6]	SA D pipe = D pipe * L tube * N	Not applicable
Packinox CFE [7]		Not applicable
Vertical CFE [8]	SA head (bottom only)	Not applicable

Notes:

1. Equations for surface area are for a single exchanger unit. If multiple shells are required to reduce bundle weight or to handle a temperature cross, calculate the wetted surface area based on one shell. Use the heat input equation to determine heat input to that one shell, and then multiply by the number of shells.

2. When U-tubes are specified, the shell side surface area must include one head. The shell side nozzle is located beyond the tube bend, increasing the total surface area of the shell.

3. Obtain the kettle diameter by dividing the tube bundle diameter by 0.6. Since the liquid level in a kettle is typically maintained 2 inches above the tube bundle, we estimate that only 62.5% of the kettle perimeter is wetted. In addition, we estimate that the product sump contributes one-half the surface area of the kettle head to the total wetted surface area of the kettle. (Surface area of head is based on a kettle diameter = Dbundle / 0.6)

4. The surface area for a stab-in reboiler is limited to the reboiler nozzle. We estimate the length of this nozzle is one-half the tube bundle diameter.

5. Inflection Point Engineering does not require fire relief protection for double-pipe exchangers. The formulas can be used if the customer requests protection.

6. Obtain the number of tube sections (N) required for the exchanger duty. Determine the surface area of one tube section based on the nominal OD of the outer pipe, and multiply that area by the number of tube sections.

7. Packinox combined feed-effluent exchangers are typically used in Isomar, Pacol, Platforming, and Tatoray units. The basis for wetted area is a continuation of feed flow for a few minutes before the operator can stop the flow. With recycle gas stopped, feed can spill over into the bottom head. The vaporization is based on wetted area equal to the bottom head plus 20% of the tangent length. Obtain the shell diameter and tangent length from the CFE optimization study.

Assume the wetted length of the shell is 20% of the tangent length. Apply the effective fire height limit after calculating this length. Use the final elevation, if known,. If not known, use 10 feet.

8. Vertical CFE’s can be used in the same units as the Packinox. Although the shell side normally has no liquid level, there is liquid hold-up from this side in the bottom head. Use the total surface area of the bottom head to account for potential drainage and other uncertainties. For the tube side, the only exposed surface is the top head, which Inflection Point Engineering ignores. Obtain the dimensions from TZ-400-01 for the project, or consult an Exchanger Specialist.

Notes: The adjacent molecular weight labels near the horizontal axis for 150, 180 and 220 should be 150, 160 and 180, respectively.

The correct value for the far left latent heat curve is 150 Btu/lb.