Fractionator Relief Loads

Table of Contents

1. Purpose

This procedure describes the guidelines to be followed for determining Fractionator relief loads. These guidelines supplement the information contained in Tool Documentation ". The procedure requires inclusion of external fire relief rate contributions from heat exchangers associated with the fractionator, except in circumstances that are defined. The procedure also gives guidance on use of P9.8, , HYSYS and UniSim process simulators, to generate relief flash data for a fractionator.

2. Fired Heater Reboiler

In general, assume the relief duty available from a Fired Heater (radiant or convection coil) is:

• 100% of normal heater duty, if the reboiler circulating pump and heater firing continue [note 1], or

• Residual duty, equal to 40% of normal heater duty, if the reboiler circulating pump or heater firing stops [note 2].

The residual duty accounts for the significant amount of heat that will radiate to the process coil from the heater firebox and heater lining. The above guideline for residual duty is conservative whatever refractory material (ceramic fiber, castable, brick, etc.) is ultimately used for the heater lining.

[Note 1] If an SIS (reference Procedure ”) with the necessary safety integrity level (SIL) will shut down heater firing on detection of high pressure in the fractionator, use residual duty to determine the relief load.

[Note 2] It is assumed that, if the reboiler circulating pump stops, the low-flow heater protection instrumentation will then shut down heater firing.

2.1 Relief Loads - General Electrical Power Failure

During a refinery-wide general electrical power failure (GPF), a motor-driven reboiler circulating pump will stop whatever its voltage. The resulting low-flow shutdown will stop the reboiler heater from firing. Therefore, if the operating and spare pumps are both motor-driven, use residual duty to determine relief load when evaluating the GPF case.

If either the operating or spare pump is turbine-driven, assume that the turbine-driven pump is on-line at the time of the GPF. Use 100% of normal heater duty to determine the relief loads because we assume that steam will be maintained for a sufficiently long time that the relief valve lifts. If a high-high pressure SIS [note 1] is provided to stop heater firing, you can use the residual duty to determine relief load.

2.2 Relief Loads - Partial Electrical Power Failure

The Partial Electrical Power Failure (PPF) case assumes fin fans, reflux pumps, etc. draw power from the failed system, but that the reboiler circulation pump draws power from a parallel, energized system, e.g. motors may draw power from an “A” or “B” system, and one bus fails [note 3]. Therefore, use 100% of normal heater duty to determine relief loads when evaluating a PPF case. If a high-high pressure SIS [note 1] is provided to stop heater firing, you can use the residual duty to determine relief load.

2.3 Relief Loads - Other Failure Cases

When evaluating other failure cases (loss of reflux pumps, air condenser failure, etc.) which do not involve fired heater reboiler circulation pumps, use 100% of normal heater duty to determine relief loads. If a high-high pressure SIS [note 1] is provided to stop heater firing, you can use the residual duty to determine relief load.

[Note 3] The partial power failure case with a fired heater reboiler is often defined as a transformer failure (“Low Voltage Electrical Power Lost”) because the circulation pump is often drawing power from a higher voltage level than the condenser fans and reflux pump. If the affected equipment is all on the same voltage level, the “A/B” type of failure should be considered.

3. Partial Electrical Power Failure Analysis

3.1 Reporting the Partial Power Failure Case

A partial power failure case shall be reported on the 807 specification in either of these specific circumstances:

• whenever the required relief area is greater than the General Electrical Power Failure case (especially when the latter is zero) or

• whenever the partial power failure scenario causes relief from more than one PRV, even if the rates do not occur simultaneously (e.g. Xylene Column overhead integrated with Parex Unit column reboilers).

Otherwise, when a partial power failure case is identical to the General Power Failure case or results in a smaller orifice area, it should not be reported, especially when space is limited for reporting all other relevant cases.

Specification notes should be used to describe the basis for the partial power failure case if the details will not be apparent to the customer. Example: “Relief load for Low Voltage Electrical Power Lost case is based on reboiler circulation continuing, stopping of feed, and flooding of overhead condensing system.”

3.2 Feed Stops or Continues?

Most partial power failure cases for fractionators involve loss of reflux and/or overhead heat removal while the heat input continues. The status of the feed during the failure case will affect the heat and weight balance at relief. For any feed which is supplied from a motor-driven pump, analyze both “Feed Stops” and “Feed Continues” scenarios, and report the worse case. This is an important step because the motor may be on a circuit or feeder separate from the reflux pump. For pressured feeds, consider both states unless the hydraulics at relief (including static head) would prevent continuation of the feed.

4. Residual Credit for Air-Cooled Overhead Condenser in a Power Failure

Failure of the fans to an air-cooled condenser may still allow partial condensing duty through natural convection. When the reflux flow is also lost (e.g. reflux pump power is lost), the receiver and condenser will flood eventually unless the heat input from the reboiler stops.

During a General Power Failure, credit for residual condensing duty should be taken only when the total normal reboiling duty is provided by a fired heater and the circulation pumps are motor-driven. In all other cases, reboiler heat duty is continuous which will result in flooding of the overhead system and cancellation of any credit.

5. Design Tool Spreadsheet

Use the Design Tool spreadsheet titled ” to determine Fractionator relief loads, except as indicated below.

Fractionator Configurations

Refer to TD-807-03 section 3 for the typical configurations which cannot be directly analyzed with TZ-807-03. For such configurations (for example, product fractionators) not covered by the tool consult with your PRV Network Member or the PRV Specialist if you are unsure how to perform the analysis.

Relief Cases not covered by the Tool

Two cases which are not analyzed by TZ-807-03 are heat exchanger tube failure and inlet control valve fails open. Guidance for tube failure is provided in IPE-TM-807-04 and T-807-07. Refer to IPE-TM-807-09 for the control valve fails open case.

6. Fractionator - External Fire

The Design Tool spreadsheet TZ-807-03, “Fractionator Relief Load Calcs” was written with inputs and calculations in place to estimate the relief contributions generated by heat exchangers in the . These inputs and calculations should be used together with the preliminary exchanger sizing information from the heat exchanger Tool ”.

If the fractionator does not fall within the scope of TZ-807-03, follow the procedures in ” to calculate the fire case relief load contributions for heat exchangers.

6.1 Equipment included in the fractionator fire zone

Per API standard, all equipment located within an 80 ft circle is considered in the same fire zone. As such, the fractionator, overhead receiver, condenser, reboiler, and associated piping are usually in the same zone and their respective fire case loads are additive. Vessels or exchangers located at least 25 feet above grade are considered outside the fire zone and do not contribute to the fire case load.

6.2 Inclusion of Exchangers in Fractionator External Fire Relief Load

If heat exchanger(s) are:

• within the effective fire height (reference Procedure ), and
• associated with a fractionator [note 4], and
• NOT air-cooled [note 5]

then the estimated relief load generated by them should be included in the External Fire relief load for that fractionator.

[Note 4] For a revamp ONLY, if the plot plan shows that an exchanger is outside the fractionator fire zone (see IPE-TM-807-05 for definition), it should be excluded.

[Note 5] Air-cooled exchangers (condensers and coolers) are assumed to be above the effective fire height, and therefore are excluded.

a. Exchanger Contribution to the Fire Case Load

Calculate the heat input to each exchanger independently using the wetted surface for each exchanger shell. If multiple shells are required for one service, determine heat input separately for each shell. Add 15% to the surface area to account for heat input to the exchanger piping.

The material vaporized in the exchanger must be consistent with the normal material in that exchanger, e.g. feed material should be vaporized in the feed exchanger. Each fire case load will have a different composition and temperature and these streams are mixed in the fractionator to obtain the total fire case load.

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 and kettle type exchangers.

If there is not a free path between an exchanger and the fractionator, do not include it in the fire case calculation. Since these exchangers may be isolated from the fractionator PRV, they require a separate PRV for the fire case. The equations used to calculate the wetted surface area for various exchanger types are given in .

(1) Shell & Tube (Straight or U-tube) or Multi-Tube Exchangers

If the process fluid is on the shell side, the wetted surface area is a cylinder based on the shell diameter and length. (For U-tube exchangers, the surface area of the shell cover is added to the cylinder area.)

If the process fluid is on the tube side, base the wetted surface area on the exchanger head(s), i.e. channel plus channel cover. Do not add the tube bundle to the surface area of the exchanger head because Inflection Point Engineering assumes that it does not contribute significantly to the tube side fire case load.

(2) Kettle Type Exchanger (Reboiler or Steam Generator)

If process fluid is on the shell side, the wetted surface area is a cylinder section based on the kettle weir height, i.e. the liquid height needed to submerge the tube bundle.

If process fluid is on tube side, limit the wetted surface area to the exchanger head. Do not add the tube bundle to the surface area of the exchanger head because Inflection Point Engineering assumes that it does not contribute significantly to the tube side fire case load.

(3) Feed-Bottoms Exchanger

See “Shell & Tube Exchangers” above to determine the wetted surface area for this type exchanger. If the process fluid on each side has a free path to the fractionator, they contribute to the fractionator fire case load. If not, a separate PRV may be required to protect the exchanger, and its fire case load is not additive. For example, when bottoms material is pumped from the fractionator to the feed-bottoms exchanger, a check valve prevents fire case vapor generated in the exchanger from back flowing to the fractionator.

(4) Vertical Thermosiphon Reboiler

The fractionator bottoms material is always on the tube side for a vertical Thermosiphon reboiler. Limit the wetted surface area to the exchanger head(s). Do not add the tube bundle to the surface area of the exchanger head because Inflection Point Engineering assumes that it does not contribute significantly to the tube side fire case load. Depending on tube length and exchanger elevation, the top head may be out of the fire zone, i.e. above 25 feet.

(5) Stab-In Reboiler

Since the exchanger bundle for a stab-in reboiler is enclosed by the fractionator, a stab-in reboiler has little impact on the fire case calculation. The exchanger flange (shell side) and the channel and its cover (tube side) are exposed to the fire and they are added to the wetted surface area. However, the stab-in reboiler may be elevated higher than 25 feet and be completely out of the fire zone.

(6) Air-Cooled Condensers and Air-Cooled Product Rundown Coolers

Air-cooled exchangers have an exceptionally large surface area exposed to the external fire and are capable of generating an excessive fire case load. As such, design air-cooled condensers to be free-draining to minimize the wetted surface area. If the air-cooled exchanger cannot be made free draining, e.g. condensers with a sub-cooling section, sidedraw condensers, or product rundown coolers, elevate the exchanger out of the fire zone, i.e. at least 25 feet above grade.

(7) Double Pipe Exchangers

A double pipe exchanger is generally installed horizontally, four or five feet above grade and is considered to be in the same fire zone as the fractionator. If the process fluid is liquid, or if it is partially vaporized, assume that the entire surface area of the exchanger is wetted.

If feed is in the outer pipe, base the wetted surface area on a cylinder having a length equal to the number of tubes times the tube length and a diameter equal to the outer pipe diameter.

If feed is in the inner pipe, Inflection Point Engineering assumes that no wetted surface area is exposed and that it does not significantly contribute to the fractionator fire case load.

Per API Standard 521, the exponent in the heat input equation is 1.0 for a double pipe exchanger instead of that normally used, i.e. 0.82.

(8) Vent Condenser

Since the vent condenser is free draining to the overhead receiver, it does not contribute to the fractionator fire case.

b. Single or Multiple PRVs

If a clear path does not exist between the individual equipment and the fractionator PRV, provide a separate PRV per ASME code.

If the condenser rundown line enters the receiver at the bottom or below the liquid level, or if the condenser is mounted at grade, consider adding a separate PRV on the receiver. See .

6.3 ‘Forcing’ a Receiver External Fire Relief Load

Occasionally a fractionator’s receiver has a dedicated PRV and is elevated above the effective fire height. Refer to IPE-TM-807-05, section 5.1 and 5.5 for the minimum wetted surface area.

7. Vapor and Liquid Phase Specific Enthalpies

If P9.8 is used as the process simulator, flash the process stream to XX wt fraction [note 6], and subtract the specific enthalpy of the liquid phase from the specific enthalpy of the vapor phase to obtain the latent heat of vaporization. If the latent heat is less than 50 BTU/lb, use 50.

If , HYSYS or UniSim is used, the latent heat cannot be obtained by subtracting the specific enthalpies. These simulators incorporate heats of formation in their enthalpy calculations, making it difficult to isolate the latent heat from the heats of formation.

If or UniSim was 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.

If HYSYS was used for the process simulator, additional flash calculations are required to obtain the thermodynamic data required for relief conditions.

1. Duplicate the stream with the required composition.

2. Flash the stream to XX mol fraction [see note 6] at accumulated pressure.

3. Separate the vapor and liquid phases into two streams.

4. Route the liquid stream through a heater unit operation.

5. Specify the heater to vaporize 0.001 (mol fraction) of the liquid, forcing the temperature out to be the same as the temperature in. (Since the stream composition, vapor fraction, and temperature are specified, there must be one degree of freedom for the flash calculation. In this case, the pressure at the heater outlet is left unspecified, and will be calculated by the simulator.) Important: the outlet temperature of the heater should be set by copying and pasting the inlet value so that the temperatures are exactly equal. Alternatively, a SET operation can be specified to assign the outlet temperature. Manually rounding off the inlet value and entering it as the heater outlet temperature can lead to inaccurate results.

6. Determine the latent heat of vaporization by dividing the calculated heater duty by the mass flow of vapor leaving the heater. If latent heat of vaporization is less than 50 BTU/lb, use 50. If the calculated latent heat for a hydrocarbon mixture is greater than 250 BTU/ lb, use 250 unless the hydrocarbon stream contains more than 5 wt% water. For those cases, use calculated latent heat.

[Note 6] For fractionator feed material * XX = 0.05

For bubble-point material ** XX = 0.001

For dew-point material *** XX = 0.999

For fire case liquids **** XX = 0.30

* If the fractionator feed contains more than 5 mol% (vapor) at normal operating conditions, set XX equal to the normal mol% vapor and flash the feed at accumulated pressure. If 0.05 results in excessive temperature or flash failure, consult PRV Specialist.

** Fractionator bottoms, side draw and upper reboiler streams.

*** Fractionator overhead vapor.

**** If the flash at XX = 0.30 fails or results in a vapor T greater than 795°F, consult TEC-807-05.

The Fractionator tool uses vapor and liquid SPECIFIC enthalpies to construct the heat and mass balance at relief conditions. In the “Relief Flash Data” input panel(s), directly enter the specific enthalpy obtained from HYSYS for one phase, and use the latent heat to determine the specific enthalpy of the other phase, as indicated below.

1. For the fractionator feed(s), net bottom product, sidedraw liquid, etc., directly use the LIQUID phase specific enthalpies obtained from HYSYS in step 5. Calculate the VAPOR phase specific enthalpy by adding the latent heat determined in step 6 to the liquid specific enthalpy of the corresponding stream.

2. For the overhead vapor stream, directly enter the VAPOR specific enthalpy obtained in step 5. Determine the LIQUID specific enthalpy by subtracting the latent heat determined in step 6 from the vapor specific enthalpy.

3. Since the Fractionator tool calculates the latent heat of vaporization using the specific enthalpy data, verify that that the calculated latent heat values in the tool are the same as those determined in step 6.