Depressuration Report

Depressurisation Report

NO-HLD-10-AET2-001021

Rev.

Status

Date

Revision memo

00

IFR

25.05.12

ISSUED FOR REVIEW

Issued by GK/SLT

Checked by MØ

Approved by NOS

This document has been generated by an Electronic Document Management System. When printed it is considered as a for information only copy. The controlled copy is the screen version and it is the holder's responsibility that he/she holds the latest valid version.

Document number

DEPRESSURISATION REPORT Document Type : REP System / Subsystem : Contractor document number : 4100H11.002

NA

Discipline :

NO-HLD-10-AET2-001021

PRO

Revision 00 Rev Date :

Status

IFR

25.05.12 Page 2 of 62

TABLE OF CONTENT

1. 2. 3. 4.

Summary ............................................................................................................................................... 3 Introduction ........................................................................................................................................... 3 Abbreviations ........................................................................................................................................ 4 Depressurisation Criteria....................................................................................................................... 5 4.1. Process Sectionalisation........................................................................................................... 5 4.2. Calculation Design Basis .......................................................................................................... 5 4.3. Initial condition .......................................................................................................................... 6 4.3.1. Calculation results – fire case ............................................................................................... 6 4.3.2. Equipment and piping integrity during fire............................................................................. 7 5. Depressurisation methodology.............................................................................................................. 7 5.1. General ..................................................................................................................................... 7 5.1.1. Pipe....................................................................................................................................... 7 5.1.2. Equipment............................................................................................................................. 7 5.2. Fire Case .................................................................................................................................. 7 5.3. Cold case – minimum design temperature ............................................................................... 8 5.4. Manual Depressurisation .......................................................................................................... 9 6. Depressurisation and Cold temperature result...................................................................................... 9 6.1. HP Depressurisation ................................................................................................................. 9 6.1.1. HP3-20-RO0010 1st Stage Separator .................................................................................. 9 6.1.2. HP4-20-RO0024 2nd Stage Separator ............................................................................... 10 6.1.3. HP6- 35-RO1025 Gas Export Compressor 1...................................................................... 11 6.1.4. HP20- 35-RO1225 Gas Export Compressor 2.................................................................... 12 6.1.5. HP8-31-RO0015 1st Stage Recompression ....................................................................... 13 6.1.6. HP9-31-RO0036 2nd Stage Recompression...................................................................... 14 6.1.7. HP10-31-RO0056 3rd Stage Recompression..................................................................... 15 6.1.8. HP11-31-RO0074 4th Stage Recompression..................................................................... 15 6.2. LP Derpressurisation .............................................................................................................. 16 6.2.1. LP1-20-RO0042 3rd Stage Separator ................................................................................ 16 6.3. Maintenance Depressurisation ............................................................................................... 17 6.3.1. 10-RO-1222 Brent gas well (Typical) To HP....................................................................... 18 7. References .......................................................................................................................................... 19 APPENDIX A - DEPRESSURISATION LOAD SCHEDULE....................................................................... 20 Depressurisation Load Schedule ............................................................................................................ 20 Manual Depressurisation Load Schedule ............................................................................................... 21 Summary of Low temperature calculations............................................................................................. 22 APPENDIX B - NEW*S Documentation ..................................................................................................... 23

This document is the property of COMPANY. It must not be stored, reproduced or disclosed to others without written authorisation from the Company. DEPRESSURISATION REPORT-REV00.DOC

Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


Status

IFR

25.05.12 Page 3 of 62

1. SUMMARY The purpose of the depressurisation system is to reduce the pressure in hydrocarbon systems. Depressurisation can be planned operations, for example before maintenance, or it can be at an emergency in case of fire, to ensure that the system will not be subject to stresses above design at extreme temperatures. In addition, the depressurisation shall reduce the hydrocarbon inventory in order to minimize the amount of hydrocarbons fed to a fire in case of a rupture. The main segments modelled in the 3D-model have been simulated in NEW*S. For the Gas Export and the 4th stage Recompression segment further investigation is needed to keep the minimum temperature above -46°C. The blowdown time need to be prolonged or it may be limited by initiating depressurisation at a higher temperature than the minimum ambient of - 10°C. The blowdown rate currently exceeds 10.15 MSm³/d, calculations need to be rerun with longer depressurisation time until 10.15 MSm³/d is not exceeded. Equipment and piping integrity during fire will be part of the detail engineering work.

2.

INTRODUCTION

The main objective of this report is to describe the design of the Hild depressurisation system with respect to: x

Design rates and capacities to the flare systems.

x

Sizes and sizing basis for major blowdown devices.

x

Sizes of depressurisation lines in the flare system.

x

Uncertainties and areas for further work.

The design is in accordance with all applicable guidelines, standards and company requirements. Where this is not the case, Company will be notified for review and approval of the deviation. Reference is made to relevant standards and philosophies listed in section 10 and to the following design documents: NO-HLD-10-AET2-001106

Safety Analysis Flow Diagram, Oil Separation

NO-HLD-10-AET2-001107

Safety Analysis Flow Diagram, Oil Export

NO-HLD-10-AET2-001108

Safety Analysis Flow Diagram, Gas Recompression

NO-HLD-10-AET2-001109

Safety Analysis Flow Diagram, Gas Treatment And Export

NO-HLD-10-AET2-001110

Safety Analysis Flow Diagram, Flare, Closed & Open Drain

NO-HLD-10-AET2-001112

Safety Analysis Flow Diagram, Fuel Gas System

NO-HLD-10-AET2-001126

Utility Flow Diagram, Flare System


Document number


3.

NA

Discipline :

NO-HLD-10-AET2-001021

PRO


Status 25.05.12 Page 4 of 62

ABBREVIATIONS

BDV

Blowdown valve

ESDV

Emergency Shutdown Valve

F

Fire

FPSO

Floating Production Storage Offloading (Vessel)

HP

High Pressure

ISO

International Standard Organisation

KO Drum

Knock Out Drum

LAH

Level Alarm High

LAL

Level Alarm Low

LSHH

Level Switch High High

LP

Low Pressure

LSLL

Level Switch High High

NEW*S

Fluid properties and process simulation program from Bubblepoint AS

PAH

Pressure Alarm High

PDMS

Plant Design Management System from Aveva.

PFP

Passive Fire Protection

PSHH

Pressure Switch High High

PSLL

Pressure Switch Low Low

PSS

Process Shutdown System

PSV

Pressure Safety Valve (Relief valve)

P&ID

Piping and Instrument Diagram

RO

Restriction Orifice

SD

Safety Shut Down

SOP

Settle-out pressure

UFD

Utility Flow diagram


IFR

Document number


4.

NA

Discipline :

NO-HLD-10-AET2-001021

PRO


Status

IFR

25.05.12 Page 5 of 62

DEPRESSURISATION CRITERIA

The ISO 23251 /3/ suggests a depressurisation down to 6.9 barg or 50% of design pressure, which ever is lower, within 15 minutes. According to NORSOK S-001 /4/ the use of passive fire protection shall be minimized by rapid depressurisation. Simultaneous depressurisation is the preferred method for evacuation of hydrocarbons. It is considered as the safest depressurisation method due to less complex configuration than for staged depressurisation. Low temperature depressurisation calculations needs to be performed in order to validate the selected minimum design temperature.

4.1.

Process Sectionalisation

The process is divided into depressurisation sections by sectionalisation valves (SD or ESD valves). Reference is made to GS EP SAF 261 /5/, chapter 5.1.3 defining the criteria for whether a BDV is required. Ref. Table 4-1.

Table 4-1 Criteria for BDV.

Piping or Vessel

BDV required That cannot be isolated No That can be isolated, but cannot No be exposed to fire. That can be isolated and can be exposed to fire: x Flammable gas x P > 7 barg and PVgas > 100 bar.m3 x Liquefied HC x Mgas or Mliq > 2 tonnes of C4 and more volatile x Liquid HC x No x Two-phase x P > 7 barg and PVgas > 100 bar.m3 x Toxic gases x As required for protection of personnel

The depressurisation facilities consisting of an actuated block valve (BDV), orifice plate, pipe expander and a manual isolation valve (locked open, full bore) in that order, all in upstream system pipe spec.

4.2.

Calculation Design Basis

The depressurisation calculations will be performed in the NEW*S software Version 3.30, 2011 (From Bubblepoint AS). The software has been validated by Imperial College on the Skarv BP FPSO project and gives satisfactorily results compared with the BLOWDOWN software (Imperial College). Further it should be noted that the methodology utilised in the NEW*S software is acknowledged by the API 521 committee (Reference is made to API 521 /3/ chapter 5.15.2.3. In this section there is made a reference to a more rigorous method. Reference [141] effectively describes the thermodynamic background for the NEW*S software presented at the GPA conference in 1993). Reference is made to Appendix B This document is the property of COMPANY. It must not be stored, reproduced or disclosed to others without written authorisation from the Company. DEPRESSURISATION REPORT-REV00.DOC

Document number


4.3.

NA

Discipline :

NO-HLD-10-AET2-001021

PRO


Status

IFR

25.05.12 Page 6 of 62

Initial condition

No credit is taken for fire-fighting equipment. The flash is assumed to be isentropic and compensated by heat from the wall heated by the external fire load. Initial condition and design criteria for calculating depressurisation loads for sections exposed to fire are as follows: x

Temperature, ref. Table 4-2.

x

Pressure, ref. Table 4-2.

x

Vessel liquid level, ref. Table 4-2.

x

Heat input is 100 kW/m2

x

Backpressure is 1 bara

x

Combined piping and equipment volumes, wetted areas and metal mass to be used.

For Emergency depressurisation AET will follow the Hild Field Development Basis of Design /6/ section 15.2.15 quoting that the platform shall be designed in accordance with GS EP SAF 261 /5/. AET suggest using PSHH or settle-out pressure based on PSHH at suction and discharge side of compressor and normal operating temperature. Ref GS EP SAF 261 section 5.2.2 quoting "The initial pressure to be considered shall be the maximum operating pressure, which will normally correspond to the PSHH." This will derogate from GS EP EPC 103 /7/, section 13.2 quoting "The initial pressure will be the system design pressure/safety relief valve set pressure, except for the compressor systems for which the settle out pressure will be considered."

Table 4-2 Initial conditions, depressurisation

Case Purpose

Fire? Initial temperature Initial pressure

Initial liquid level

4.3.1.

Emergency depressurisation Maximum total depressurisation flowrate to flare Sizing of depressurisation orifice and depressurisation piping Yes Normal operating PAHH / Settle out pressure

LAHH/LL to be checked

Planned depressurisation Low temperature Total depressurisation Minimum design flowrate expected upon temperature case planned shutdown

No Normal operating

No Minimum ambient

Normal operating pressure Based on constant / Settle out pressure volume flash from PSV set pressure to minimum ambient temperature Normal operating LAHH/LL to be checked

Calculation results – fire case

Reference is made to Depressurisation Load Schedule attached in Appendix A, which gives a total blowdown rate of 10.15 MSm3/d to the HP Flare and approx 0.16 MSm3/d to the LP Flare.


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


Status

IFR

25.05.12 Page 7 of 62

According to GS-EP-SAF 261 /5/ it is a good principal to fire insulate liquid containing vessel and piping to reduce boiling and thereby avoid longer depressurisation time. This principle is foreseen to result in fire insulation on all piping, pressure vessels, filters, electrical heaters and shell and tube heat exchangers containing liquid. Passive fire protection is foreseen required on the 2nd Stage Separator and the 3rd Stage Separator. The fluids in the 2nd Stage Separator and 3rd Stage Separator contain significant quantities of water. As the segments are depressurised, the vapour phase will not be saturated with water, which subsequently will result in a significant mass flux of water to the vapour phase as the temperature increases and the pressure drops. These vessels are recommended to be fire insulated to prevent a steam explosion.

4.3.2.

Equipment and piping integrity during fire

HOLD - will be part of detail engineering work.

5.

DEPRESSURISATION METHODOLOGY 5.1.

General

x

Sensitivity checks on depressurisation rate have been performed on different years for the 1st Stage Separator section, worst case is HILD-Gas_Alone-2018-07 will be used for Fire Case

x

NEW*S will be used for calculations outlined in section 5.2.

x

NEW*S will be used for low temperature calculations outlined in section 5.3.

5.1.1.

Pipe

The PDMS 3D model will be used to determine pipe work volumes within each segment (limited by sectionalisation valves). The inner diameter is determined using “Piping And Valve Material Specification” ref /1/. In addition, 15% of the calculated pipe work volume is added to include a margin in this early stage of the project. Volumes will be rechecked as part of detail engineering work.

5.1.2.

Equipment

The equipment volume is based on the Master Equipment List /8/. The end head of all vessels are assumed to be elliptical. An additional 10% of the total calculated volume is added to include a margin. Volumes will be updated when supplier information is available.

5.2.

Fire Case

The purpose of the fire case is to calculate the orifice area and to design the upstream and downstream pipe dimensions of the blowdown line.

Start conditions:


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


Status

IFR

25.05.12 Page 8 of 62

x

Normal operating temperature

x

PSHH or settle-out pressure based on PSHH at suction and discharge side of compressor.

x

High high liquid level in separators

x

High high liquid level in scrubbers

x

Credit for insulation only if PFP, i.e equipment and pipes provided with insulation for other purpose are regarded as uninsulated.

x

The heat input is 100 kW/m² (Global fire, ref /4/)

x

Add all calculated blowdown rates (Simultaneous opening of the BDV for all fire zones) and find the total blowdown rate,

x

If the blowdown rate exceeds 10.15 MSm³/d, calculations to be rerun with longer depressurisation time until 10.15 MSm³/d is not exceeded.

x

As a first estimation, the pressure drop requirement is to reach 6.9 barg within 14 minutes. A time delay of 1 minute is assumed in order to allow sectionalisation valves to close and blowdown valve to open. A total depressurisation time will then be 15 min, ref GS EP SAF 261 /6/. This default depressurisation time, is for vessel wall thickness of 25 mm; for thinner walls, the depressurisation time shall be reduced i.e. 3 minutes for each 5 mm. For thicker walls, the depressurisation time cannot be longer than 15 minutes unless a specific study is validated by Total.

The depressurisation time of the compressor segments and the TEG Contactor segment needs to be revised when supplier information with respect to maximum pressure drop per time [bar/min] has been clarified. Currently it is assumed that these segments will withstand the gradients resulting from an API blowdown.

5.3.

Cold case – minimum design temperature

The purpose of the Cold Case is to determine the minimum operating temperature of the blowdown segment and the minimum operating temperature downstream the orifice. The low temperature calculation is performed on the main vessel in a process segment, ignoring all piping. This is regarded to be a conservative approach as the piping holds much more heat capacity in the steel relative to heat capacity of the process fluid, and would contribute to increase the minimum design temperature if included in the calculations. The low temperature calculations performed in NEW*S give separate minimum temperatures in both the fluid and the pipe/vessel wall for both the liquid and vapour phase. The lowest steel temperature reported is applied as the basis for the minimum design temperature and material selection for a segment. However, it is evaluated that the depressurisation nozzles on the separators and the connected depressurisation lines will locally reach a significantly lower temperature due to the good heat transfer caused by a high flow rate throughout the depressurisation. Therefore, the minimum gas temperature is applied as the basis for material selection and minimum design temperature for these parts of the segments whenever it affects the material selection. Low temperature calculations will be performed for high and low pressure segments. Reference is made to Depressurisation Load Schedule attached in Appendix A for cold temperature results. Low temperature calculations are performed with the orifice size calculated in fire case and initial conditions are given below

Start conditions: x

The initial pressure is based on constant volume flash from settle out PSHH or PSV set pressure and normal operating temperature to an ambient temperature of -10 °C, as stated by Company /6/.


Document number


NA

x

Pipe work metal and volume excluded.

x

Use LSLL in vessels. LSHH to be checked as well.

x

Use the orifice size calculated in Fire Case

5.4.

Discipline :

NO-HLD-10-AET2-001021

PRO


Status

IFR

25.05.12 Page 9 of 62

Manual Depressurisation

Equipment that may be isolated and maintained during operation is equipped with manual depressurisation line to Flare. These depressurisation lines are equipped with a restriction orifice that shall be sized to keep temperatures and flow rates within the design limitations of the equipment. Reference is made to the Depressurisation Load Schedule attached in Appendix A, a separate sheet gives the overview of maintenance blowdown with rates and corresponding cold temperatures.

6.

DEPRESSURISATION AND COLD TEMPERATURE RESULT

The input data collected from the spreadsheet Hild Equipment volumes.xls is included in a Unit Defined Command Procedure file (UDC-file) – for calculation of pressure profile (time dependent pressure) during depressurisation. An UDC-file is used by NEW*S to read input data to defined units in NEW*S and is actually a collection of most of the commands needed to run a depressurisation simulation. In the UDC-file, all lengths are in meter and all times are in minute. The simulation gives, besides pressure in the segment with time, blowdown rates with time, time to reach backpressure from flare and remaining volume and mass of fluid in the segment with time. Based on experience from previous project the Coefficient of Discharge used for all restriction orifice plates with critical flow have been CD= 0.83932. All the UDC and results obtained with the process simulation tool NEW*S, will be found in the folder \depressurisation simulations\ The folder name will indicate the corresponding RO.

6.1.

HP Depressurisation

6.1.1. HP3-20-RO0010 1st Stage Separator 20-RO0010 is installed on the 1st Stage Separator, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001210. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. As input for the calculation stream 20010 from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

85

Total vessel volume, m3

63.62

Initial flow rate, kg/h

60103.2

14 minute to reach 7.9 bara Calculated orifice diameter, mm

38.74


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO

Revision Status 00 Rev Date : 25.05.12 Page 10 of 62

IFR

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\HP3-20RO0010 All relevant data are included in NEW*S UDC HP3-20-RO0010.udc.txt and simulation result is printed in HP3-20RO0010.doc Low temperature calculation is performed in NEW*S on the 1st Stage Separator, 10-VZ2001, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000

Operating pressure, bara

62.5

Calculated minimum temperatures: Vapor (nozzle temperature),C

-51.80

Liquid,

-22.53

C

Minimum metal temperatures: Vessel (vapour space),

C

-22.29

Vessel (liquid space), C -21.32 Minimum downstream temperature in the flare system at t= 2.45 is

-71.22

Reference is made to NEW*S simulations stored in: depressurisation simulations\Cold temperatures\ HP3_cl-1sep All relevant data are included in NEW*S UDC HP3_cl-LL.udc.txt and result is printed in HP3_cl-LL.doc The minimum design temperature of -46°C in the segment and a cold sleeve in the depressurisation nozzle on the Separator with minimum design of -60°C is sufficient.

6.1.2. HP4-20-RO0024 2nd Stage Separator 20-RO0024 is installed on the gas outlet of the 2nd Stage Separator, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001211. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. As input for the calculation stream 20020 from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

32


184.53


46236


56.05

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\HP4-20RO0024 All relevant data are included in NEW*S UDC HP4-20-RO0024.udc.txt and simulation result is printed in HP4-20RO0024.doc


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR

Low temperature calculation is performed in NEW*S on the 2nd Stage Separator, 10-VZ2002, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


21.45


-44,92

Liquid,

-14.17

C


C

-16.21


-49.32

Reference is made to NEW*S simulations stored in: depressurisation simulations\Cold temperatures\ HP4_cl- 2sep All relevant data are included in NEW*S UDC HP4_cl.udc.txt and result is printed in HP4_cl.doc The minimum design temperature of -46°C in the segment and a cold sleeve in the depressurisation nozzle on the Separator with minimum design of -60°C is sufficient.

6.1.3. HP6- 35-RO1025 Gas Export Compressor 1 35-RO1025 is installed on the Gas Export Compressor 1 outlet, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001231. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. As input for the calculation stream 35003A from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

117


33.87


61165,2


29.09

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\HP6-35RO1025 All relevant data are included in NEW*S UDC HP6-35-RO1025.udc.txt and simulation result is printed in HP6-35RO1025.doc Low temperature calculation is performed in NEW*S on the Gas Export Compressor Suction Scrubber 1, 10VZ3501, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


92.9


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR


-51.23

Liquid,

-32.55

C


C

-24.94


-87.96

Reference is made to NEW*S simulations stored in: depressurisation simulations\Cold temperatures\ HP6-ExeA The results need further investigation. To keep the minimum design temperature of -46°C in the segment, the blowdown time need to be prolonged or it may be limited by initiating depressurisation at a higher temperature than the minimum ambient of -10 °C.

6.1.4. HP20- 35-RO1225 Gas Export Compressor 2 35-RO1225 is installed on the Gas Export Compressor 2 outlet, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001233. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. As input for the calculation stream 35003B from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

117


29.91


53874


27.3

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\HP7-35RO1225 All relevant data are included in NEW*S UDC HP7-35-RO1225.udc.txt and simulation result is printed in HP7-35RO1225.doc Low temperature calculation is performed in NEW*S on the Gas Export Compressor Suction Scrubber 2, 10VZ3502, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


92.9


-52.70

Liquid,

-33.35

C

Minimum metal temperatures:


Document number

DEPRESSURISATION REPORT Document Type : REP System / Subsystem : Contractor document number : 4100H11.002 Vessel (vapour space),

C

NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR

-28.0


-88.58

Reference is made to NEW*S simulations stored in: depressurisation simulations\Cold temperatures\ HP7-ExeB The results need further investigation. To keep the minimum design temperature of -46°C in the segment, the blowdown time need to be prolonged or it may be limited by initiating depressurisation at a higher temperature than the minimum ambient of -10 °C.

6.1.5. HP8-31-RO0015 1st Stage Recompression 31-RO0015 is installed on the 1st Stage Recompression outlet, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001221. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. As for thinner walls, the depressurisation time shall be reduced i.e. 3 minutes for each 5 mm. The segment has a wall thickness of 8mm, so as pr governing document /6/ the blowdown time has to be reduced to less than 5 minutes to reach 50% of design pressure. As input for the calculation stream 31008 from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

8.1


12.73


4420,2

4.1 minute to reach 5 bara Calculated orifice diameter, mm

30.8

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\HP8-31RO0015 All relevant data are included in NEW*S UDC HP8-31-RO0015.udc.txt and simulation result is printed in HP8-31RO0015 to 2 bara.doc Low temperature calculation is performed in NEW*S on the 1st stage Suction Scrubber, 10-VZ3101, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


5.65


-29.06

Liquid,

-16.27

C


C

-10.23


-30.68


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR

Reference is made to NEW*S simulations stored in: depressurisation simulations\Cold temperatures\ HP8-1r1 The minimum design temperature of -46°C in the segment is sufficient.

6.1.6. HP9-31-RO0036 2nd Stage Recompression 31-RO0036 is installed on the 2nd Stage Recompression outlet, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001224. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. As for thinner walls, the depressurisation time shall be reduced i.e. 3 minutes for each 5 mm. The segment has a wall thickness of 8mm, so as pr governing document /6/ the blowdown time has to be reduced to less than 5 minutes to reach 50% of design pressure. As input for the calculation stream 31018 from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

10


7.03


4280

4.2 minute to reach 5 bara Calculated orifice diameter, mm

27.3

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\HP9-31RO0036 All relevant data are included in NEW*S UDC HP9-31-RO0036.udc.txt and simulation result is printed in HP9-31RO0036.doc Low temperature calculation is performed in NEW*S on the 1st stage Suction Scrubber, 10-VZ3101, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


6.97


-38.81

Liquid,

-18.04

C


C

-10.25


-117.57



Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR

6.1.7. HP10-31-RO0056 3rd Stage Recompression 31-RO0056 is installed on the 3rd Stage Recompression outlet, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001226. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. As for thinner walls, the depressurisation time shall be reduced i.e. 3 minutes for each 5 mm. The segment has a wall thickness of 8mm, so as pr governing document /6/ the blowdown time has to be reduced to less than 5 minutes to reach 6.9 barg. As input for the calculation stream 31028 from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

23


4.19


4894

4.6 minute to reach 7.9 bara Calculated orifice diameter, mm

19.2

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\ HP10-31RO0056 All relevant data are included in NEW*S UDC HP10-31-RO0056.udc.txt and simulation result is printed in HP10-31RO0056.doc Low temperature calculation is performed in NEW*S on the 3rd stage Suction Scrubber, 10-VZ3103, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


15.9


-29.06

Liquid,

-16.27

C


C

-10.23


-39.68


6.1.8. HP11-31-RO0074 4th Stage Recompression 31-RO0074 is installed on the 4th Stage Recompression outlet, and routed to the HP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001228. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls.


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR

As for thinner walls, the depressurisation time shall be reduced i.e. 3 minutes for each 5 mm. The segment has a wall thickness of 18mm, so as pr governing document /6/ the blowdown time has to be reduced to less than 10 minutes to reach 6.9 barg. As input for the calculation stream 31038 from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

51


29.06


41609

9.4 minute to reach 7.9 bara Calculated orifice diameter, mm

36.9

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\ HP11-31RO0074 All relevant data are included in NEW*S UDC HP11-31-RO0074.udc.txt and simulation result is printed in HP11-31RO0074 doc Low temperature calculation is performed in NEW*S on the 4th stage Suction Scrubber, 10-VZ3104, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


38.3


-43.38

Liquid,

-40.12

C


C

-14.33


-51.55

Reference is made to NEW*S simulations stored in: depressurisation simulations\Cold temperatures\ HP11-4r1 The results need further investigation. To keep the minimum design temperature of -46°C in the segment, the blowdown time need to be prolonged or it may be limited by initiating depressurisation at a higher temperature than the minimum ambient of -10 °C.

6.2.

LP Derpressurisation

6.2.1. LP1-20-RO0042 3rd Stage Separator 20-RO0042 is installed on the 3rd Stage Separator, and routed to the LP Flare header. Reference is made to P&ID NO-HLD-10-AET2-001213. The RO is sized based on volume from the depressurisation simulations\Hild Equipment volumes.xls. The minimum temperature of -46°C in the downstream LP flare system is the restriction when designing the flow orifice for the 3rd Stage Separator. Calculations performed in NEW*S indicate a maximum rate of 10491 kg/h to be within the design criteria of -46°C for the segment.


Document number


NA

Discipline :

NO-HLD-10-AET2-001021 Revision Status 00 Rev Date : 25.05.12 Page 17 of 62

PRO

IFR

As input for the calculation stream 20030 from HILD-Gas_Alone-2018-07 have been used. Else, Operating pressure, bara

4


191.68


10491

Less than 14 minute to reach 1.5 bara Calculated orifice diameter, mm

67.02

References are made to NEW*S simulations for orifice calculation – see: depressurisation simulations\LP1-20RO0042 All relevant data are included in NEW*S UDC LP1-20-RO0042.udc.txt and simulation result is printed in LP1-20RO0042.doc Low temperature calculation is performed in NEW*S on the 3rd Stage Separator, 10-20VZ2003, with an orifice size giving the same pressure profile as the one calculated in the fire case. Initial conditions and result are: Operating temperature, C

-10.00000


6.13


-40.69

Liquid,

-11.42

C


C

-12.39


-41.44

Reference is made to NEW*S simulations stored in: depressurisation simulations\Cold temperatures\ LP1_cl-3sep All relevant data are included in NEW*S UDC LP1_cl.udc.txt and result is printed in LP1_cl.doc

6.3. Maintenance Depressurisation All calculations of maintenance blowdown valves are stored in the attached folder: depressurisation simulation\Maintenance manual valve The folder name will indicate the corresponding FO. For low temperature calculation, reference is made to: depressurisation simulations\Cold temperatures\Maintenance manual valve


Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR

6.3.1. 10-RO-1222 Brent gas well (Typical) To HP 10-RO-1222 is installed on the Brent gas well and routed to the HP Flare header. Reference is made to P&ID NOHLD-10-AET2-001200. The Mach number in the downstream flare system is the restriction when designing the flow orifice on the typical Brent gas well. Calculations performed in Flare net indicate a maximum rate of 1900 kg/h. Input stream used for the calculation is 20010, PSHH=94 bara and normal operating temperature is 66.48°C.


Document number


7.

NA

Discipline :

NO-HLD-10-AET2-001021

PRO


IFR

REFERENCES 1. NO-HLD-10-AET2-100005, Piping And Valve Material Specification 2. Scandpower, Guidelines for the Protection of Pressurised Systems Exposed to Fire – Rev 2, 31.03.04 3. BS EN ISO 23251:2007 4. Norsok S-001- Edition 4, February 2008. 5. GS EP SAF 261, Emergency Shut-Down and Emergency De-pressurisation (ESD & EDP). Rev.02 6. NODOC01-#928856-V8, HILD field Development Basis of Design, Functionality Description & Operating Requirements - Rev 7. 7. GS EP ECP 103, Process Sizing Criteria. Rev.05 8. NO-HLD-00-AET2-000502 Master Equipment List 9. Determination of Temperatures and Flare Rates During Depressurization and Fire. (Sverre Overå, Ellen Stange and Per Salater, Presented at 72nd annual GPA Convention March 15-17, 1993, San Antonio, Texas 10. Size Depressurisation and relief Devices for Pressurised Segments Exposed to fire. (Salater, Overaa, Kjensjord, CEP (AIChe) sept. 2002) 11. NO-HLD-00-GEO-955616, #955616 -Metocean Specification For The Hild Field 12. GS EP SAF 262, Pressure protection relief and hydrocarbon disposal systems. Rev.02 13. NORSOK P-001, Process Design, Ed.5 Sep 2006 14. NORSOK P-100, Process Systems, Rev.03, Feb 2010 15. .NO-HLD-10-AET-1-002000, Material Selection and Corrosion Protection Report 16. P09010-MTA-0044 - METOCEAN REPORT DRAFT FROM ARGOSS 17. NORSOK L-002, Piping system layout, design and structural analysis, edition 3, July 2009



Depressurisation Load Schedule

APPENDIX A - DEPRESSURISATION LOAD SCHEDULE

Document Type : REP System / Subsystem : Contractor document number : 4100H11.002

NA

Discipline :

DEPRESSURISATION REPORT PRO

Document number Revision Status 00 Rev Date : 25.05.12 Page 20 of 62

IFR

NO-HLD-10-AET2-001021


Manual Depressurisation Load Schedule


NA

Discipline :



IFR

NO-HLD-10-AET2-001021


Summary of Low temperature calculations


NA

Discipline :



IFR

NO-HLD-10-AET2-001021

Document number


NA

Discipline :

NO-HLD-10-AET2-001021

PRO


APPENDIX B - NEW*S DOCUMENTATION


IFR

Safety

Size Depressurization and Relief Devices for Pressurized Segments Exposed to Fire Per Salater, Sverre J. Overaa and Elisabeth Kjensjord, Norsk Hydro ASA

Piping and equipment must withstand fires without rupturing. This can be accomplished by properly designing relief and depressurization systems and using passive fire protection, when needed.

T

his article presents the minimum requirements for performing proper depressurization and fire-relief calculations together with a procedure for sizing depressurization and relief systems for pressurized systems exposed to fire. An engineering approach for modeling geometrically complex process segments is detailed. This approach excludes the necessity of describing the total segment geometry in detail. A fire model is described with its required input parameters. The parameters will vary for different fire characteristics.

Minimum requirements for calculations Several simulation tools are available for sizing depressurization orifices and relief valves. Most lack the necessary physical modeling required. The list below summarizes the minimum requirements for the design of depressurization and relief devices for pressurized systems exposed to fire: • rigorous thermodynamics (multicomponent fluid model and use of equation of state) • fire modeling (emissivities, absorptivity, temperature, convection, initial flux, duration and size) • segment (vessel, pipe) material properties, i.e., tensile (rupture) strength, heat capacity, conductivity — all are temperature-dependent This article is extracted from the Norsk Hydro Best Practice on Depressurization and Fire Relief Design. This Best Practice is a major basis for the international guide, “Guidelines for the Design and Protection of Pressure Systems to Withstand Severe Fires” (1) (to be issued shortly) and the Norwegian guide, “Guidelines for Protection of Pressurized Systems Exposed to Fire” (2).

38

www.cepmagazine.org

September 2002

CEP

• heat-transfer coefficients (boiling, radiation, convection and conduction) • mass transfer between the fluid phases (boiling and condensation) • fluid flow, i.e., flow-regime calculations (laminar or turbulent) that are input into heat-transfer and pressure-drop calculations • modeling of the process-segment geometry (system volume, system outer and inner wall areas, weight, wall thickness, liquid and gas volumes) • insulation (thickness and conductivity) • stress or strain, depending upon the rupture criteria used, to which all pipes and equipment are exposed.

The fire Modeling of fires for engineering purposes requires simplifications compared to the more thorough turbulent-combustion models used in computational fluid dynamics (CFD). Nonetheless, large-scale tests (3) have verified that an engineering approach to fire modeling gives wall- and fluid-temperature profiles that are close to measured ones when choosing the appropriate input parameters for Eq. 1. Before proceeding with the fire model, some terminology needs to be defined: A global fire is a large fire that engulfs the entire or a significant part of the process segment. A local fire exposes a small (local) area of the process segment to the fire peak heat-flux. A jet fire is an ignited release of pressurized, flammable fluids. A pool fire

Local fire

Heat Flux

is the combustion of flammable or combustible fluids spilled and retained on a surface. The ventilation- and fuel-controlled fires are related to the stoichiometric ratio of air-to-fuel (Figure 1). Figure 1 is general for both a jet and a pool fire; the difference being a higher flux for the jet fire. For a pool fire, the API fire (4) is illustrated as the lower, dashed line to the right. Note that in the equation API RP 521 uses, increasing the area reduces the flux. The dashed lines represent the average heat flux. However, when studying the total volume of a fire, any point on the continuous curve will be found. A ventilation-controlled fire is to the left of the peak heat-flux in Figure 1 at a stoichiometry of < 1. The fuel-controlled fire is to the right, i.e., the stoichiometry is > 1.

Global fuel-controlled Globalventilationcontrolled

1.0 Stochiometric Ratio

■ Figure 1. The heat flux from a fire and its relation to the stoichiometric ratio.

The fire equation The heat flux absorbed by a segment from a fire, qabsorbed (kW/m2), can be modeled as:

(

)

qabsorbed = κ α segment ε fire T 4 fire − ε segment T 4 segment +

(

h × Tgas − Tsegment

)

Open-fuel-controlled pool fire

(1)

and more soot attaches to the surface. For more on absorptivity and emissivity, see Ref. 5. By combining the suggested highest or lowest typical values into the fire equation, the heat fluxes toward a cold segment are found (Table 4 (6)). Typical heat fluxes measured in large-scale jet fire and pool-fire tests are within the maximum and minimum values in Table 4. Norsok (7) recommends using the initial incident heat fluxes as specified in Table 5.

Sizing procedures The fire-relief and depressurization calculations determine: • size of the relief valves and depressurization orifices • requirements for passive fire protection (PFP) • size of the pipes downstream from the relief and depressurization valves (if any)

The absorbed heat flux will be reduced with increasing segment temperature, and a steady-state segment temperature will be reached when the heat influx from the fire equals the heat outflux from the segment. The view factor, which is not included in Eq. 1, is a Table 1. Typical flame emissivities for global scaling factor for the radiative terms. The view factor is and local fires. ≤ 1.0. It equals 1.0 when the segments that absorb radiation “see” nothing but an optical, thick flame. Calculation of Type of Fire Global Fire Local Fire εfire εfire view factors is difficult and a conservative assumption involves use of a view factor of 1.0, which results in Eq. 1. Ventilation-controlled pool fire 0.6–0.75 0.7–0.9 Fuel-controlled pool fire 0.6–0.75 0.7–0.8 The incident heat flux is calculated by setting αsegment = Jet fire 0.5–0.75 0.6–0.75 1.0 and disregarding the segment emissivity term. The “initial incident heat flux” from a fire is calculated by setting αsegment = 1.0 and Table 2. Typical temperatures and convective heat-transfer Tsegment equal to the normal operating temcoefficients for a global fire. perature (of the cold segment). 2 Type of Fire

Input to the fire equation The different terms in the fire equation are combined to achieve the required initial heat fluxes. Tables 1, 2 and 3 suggest values to be used. The segment absorptivity and emissivity in Eq. 1 are normally equal and depend upon the nature of the surface. Typical values are 0.7–0.9. A value of about 0.8 is typical for oxidized surfaces. The value will change as more

Tfire, °C

Tgas, °C

Ventilation-controlled pool fire 1,000–1,050 Fuel-controlled pool fire 950–1,000 Jet fire 1,000–1,150

850–950 800–900 950–1,050

h, W/m ·K 15–30 15–30 50–125

Table 3. Typical temperatures and convective heat-transfer coefficients for a local fire. Type of Fire Ventilation-controlled pool fire Fuel-controlled pool fire Jet fire

Tfire, °C 1,050–1,125 1,000–1,050 1,000–1,150

CEP

Tgas, °C Equal to Tfire Equal to Tfire Equal to Tfire

September 2002

h, W/m2·K 20–30 20–30 100–150

www.cepmagazine.org

39

Safety

Table 4. Minimum and maximum heat fluxes calculated by Eq. 1 for the suggested values of the input parameters toward a cold segment. Type of Fire

Maximum Incident Flux, kW/m2

Minimum Incident Flux, kW/m2

Maximum Absorbed Flux, kW/m2

Minimum Absorbed Flux, kW/m2

326 367 138 187 158 228

121 187 88 124 101 142

306 347 127 171 145 208

98 160 65 92 75 105

Global jet fire Local jet fire Global fuel-controlled pool fire Local fuel-controlled pool fire Global ventilation-controlled pool fire Local ventilation-controlled pool fire

• optimum location of the system’s Table 5. Initial incident heat fluxes against "cold" segment (7). sectionalization valves • minimum design temperature for Type of Fire Global Fire Local Fire the flare system (if any) and for the (Average load) (Maximum point load) pressurized segment. The minimum design temperature may Pool fire, enclosed area, ventilation-controlled 200 kW/m2 130 kW/m2 influence the materials selection of the Pool fire (crude), open or enclosed area, 100 kW/m2 150 kW/m2 system under evaluation. The design usu- fuel-controlled ally begins by considering carbon steel or Jet fire 250 kW/m2 stainless steel as the material of construction. However, temperature calculations may result in the need to use a different grade of steel, for culations. This is to allow for future tie-ins and expected example, replacing a normal carbon steel with one suited project growth. However, by increasing the design cafor low temperatures. pacity of flare system, less PFP will usually be required. Although a local fire has a higher heat flux than a Depressurization-orifice-sizing procedure global fire, the global fire normally exposes the pressurPrior to running depressurization calculations, the folized segment to the largest flux of heat energy, due to the lowing must be established: larger area encompassed by a global fire. Hence, the • the fire scenarios (jet fire, pool fire, local fire, global global-fire parameters determine the rupture pressure. On fire, etc.); define the initial heat flux, the duration and the the other hand, the local fire has the highest heat flux, so size (extent) of the fire(s) its parameters determine the rupture temperature of the • the criteria for unacceptable rupture, which are usually process segment. one or more of the rupture pressures, the released Valves and flanges are not accounted for in the procedure. flammable/toxic fluid at rupture, and the time to rupture Also, it does not consider the mitigating effects of active fire • the time from the start of a fire until depressurizaprotection (such as a deluge). The sizing criteria are set to tion is initiated avoid an unacceptable rupture that could escalate the fire. • the physical properties — ultimate tensile strength Step 1: Perform an initial estimate of the size of all de(UTS), heat capacity and thermal conductivity — at elevated temperatures (up to 800–1,000°C) for the materials of Nomenclature construction used in the depressurization segments • the depressurization segment geometry (system volh = convection heat-transfer coefficient of air/flame in contact with segment, W/m2·K ume, wall area, weight, etc.). = absorbed heat flux from the fire, W/m2 q Once the above data are assembled, follow the iteraabsorbed T = flame temperature, K fire tive procedure in Figure 2. The goal of this depressuriza= temperature of air/flame in contact with segment, K Tgas tion design is to limit the use of passive fire protection Tsegment = segment temperature (time-dependent), K (PFP) by depressurizing as fast as possible, while remaining within the discharge capacity of the flare sysGreek letters: tem. PFP should be avoided due to the risk of the conseαsegment = segment absorptivity, dimensionless = flame emissivity, dimensionless εfire quences of undetected corrosion under insulation, and εsegment = segment emissivity, dimensionless the additional installation and maintenance costs inκ = Stefan-Boltzmann constant = 5.67 × 10-8 W/m2·K4 curred by PFP systems. σaxial = longitudinal stress, MPa When designing a new plant, it is not recommended to = hoop stress, MPa σhoop consider using the entire flare-system capacity in the cal= equivalent stress (Von Mises), MPa σ Von Mises

40

www.cepmagazine.org

September 2002

CEP

Required information prior to blowdown Iteration - Description of the fire scenarios (type of fire, duration, heat fluxes, size) - Blowdown section geometry (system volume, area, weight, etc.) - Ultimate tensile strength at elevated temperature of materials in the blowdown section - Manual or automatic blowdown, i.e., time delay for start of depressurization - Acceptance criteria for rupture

Step 1: Estimate the size of all orifices and calculate the pressure profile and flare rates for all segments. Use the fire with the largest heat input (kW). No PFP in this initial iteration.

Reduce the size of the orifice.

(Step 1) Is the flare system capacity utilized (when adding all of the simultaneous blowdown rates together)?

No

(Step 1) Evaluate to increase the blowdown rate, preferably for the most hazardous blowdown section.

Step 2: Add insulation if required. Calculate the process segment pressure profile. Use the fire with the largest heat input (kW). Tip: Do several calculations with varying amounts of fire insulation.

Yes

No

In case of any of the "ORs"

(Step 1) Is the blowdown rate less than maximum l-dP/dt l

Yes

Step 3: Calculate the wall-temperature profile for all pipes and equipment. Use the local fire with the highest heat flux (kW/m2).

Step 4: Use the temperature profiles from Step 3 to calculate the rupture pressure for all pipes and equipment. Compare with the pressure profile from Step 2 (Step 1 in the first iteration). Acceptance criteria: - Pipe rupture pressure - Equipment rupture pressure - Released fluid at rupture - Time to rupture - No rupture

Step 5: Are the acceptance criteria for rupture met?

No

Step 6: Decide which pipe/equipment to fire insulate or Increase orifice diameter if available capacity in the flare system, or reduce system volume by relocation of sectionalization valves or increase the flare system capacity or change material quality or increase wall thickness.

Yes

Step 7: Calculate the minimum design temperature (low-temperature design temperature) of the blowdown section and the flare system tail pipe.

■ Figure 2. Follow this sizing procedure to design depressurization orifices.

Is the minimum design temperature acceptable?

No

Start depressurization at a higher temperature (or change material).

Yes The design of this section blowdown orifice and fire insulation requirements is finished.

CEP

September 2002 www.cepmagazine.org

41

Safety

pressurization orifices, using the capacity of the flare system in the calculations. A recommended first estimate is an orifice diameter that takes the pressure below the “unacceptable” rupture pressure within the typical time to rupture. The initial pressure should be the highest normal operating pressure or an equalization pressure (settle-out pressure) for a compression segment. A global fire should be used. The typical time-to-rupture can be set at that interval it takes to reach a 600–800°C wall-temperature, depending upon the wall thickness. A value of 5–10 min is typical for a dry wall exposed to a medium-heat-flux jet-fire, with no depressurization. One way of improving the safety of the plant is increase the rate of depressurization, as the hazardous aspects of the segment increase. The total blowdown rate can be kept unchanged by increasing the depressurization times of the less hazardous segments. A segment containing large amounts of light liquids (e.g., condensate or liquefied petroleum gas (LPG)), those that will result in boiling-liquid expanding-vapor explosions (BLEVEs) are regarded as a particularly hazardous section. In any case, there may be limitations on maximum pressure gradients for items such as compressors or gaskets. Step 2: Add insulation, if required, and simulate the pressure profile during depressurization when exposing the segment to the global fire. For the first iteration only, omit this step and go to Step 3. A global fire will add heat to the fire-exposed area without PFP. The initial pressure in this calculation should be equal to the highest normal operating or settle-out pressure. Credit for insulation should be given only for PFP. Piping and equipment with insulation used for purposes other than for PFP should be regarded as uninsulated. Unrealistic backpressure in the flare system may result in a too-rapid simulated depressurization. The orifice backpressure should be based on the time-dependent simultaneous depressurization rates. If a depressurization segment is 100% fire-insulated, then the integrity of the insulation and supports usually determines the maximum allowable depressurization time, which is typically 30–60 min. Account for the integrity of the insulation by extending the depressurization time for a 100% fire-insulated section. The reduced depressurization rate for this section is used to allow for the increase of the rate from a most-hazardous depressurization-section. A reduced depressurization rate may increase the fire duration, if a leak in this section is the source of the fire. Step 3: Simulate the temperature profile for all piping and equipment in the depressurization segment when exposed to the local peak-heat flux. A jet fire is normally used in these calculations, but the local load for a pool fire should be used if the segment will not be exposed to a jet fire. “All piping” means all pipes with different diameters, pressure classes and/or material qualities. The temperature profile for one particular pipe usually is rather insensitive to pressure changes

42

www.cepmagazine.org

September 2002

CEP

within a segment, i.e., the temperature profiles from the first iteration can be kept throughout the whole iteration procedure. A final update of the temperature profiles must be performed prior to the last iteration. Step 4: Calculate whether or not rupture occurs. Determine the stress or strain that all pipes and equipment are exposed to for the temperatures and pressures seen during the depressurization (Calculated in Steps 1 or 2, and Step 3) and determine whether the segment will rupture. Two failure (rupture) criteria are often used: the maximum stress or maximum strain (% elongation). The maximum stress criterion is usually the UTS. Rupture strain is a matter of definition. Strain calculations require finite-element modeling of the system, which is usually not performed during this step-wise method. Such calculations should be performed for verification purposes during the final design. The suggested approach is to calculate the stress from the internal pressure and add extra stress (margins) when calculating the longitudinal stress. The stresses of importance for a pipe are the hoop stress caused by internal pressure, and the longitudinal stress. The longitudinal stress is the sum of axial stresses due to pressure; the weight of the pipe, valves, fittings, branch pipes, etc; stress due to reaction forces exerted on the pipe by pressure; and stress due to thermal elongation of the pipe. The equivalent stress (von Mises) is the stress to be compared with the temperature-dependent UTS to determine whether rupture occurs. The hoop stress, σhoop, is equal to: σ hoop =

Pressure × Outer dia. 2 × Wall thickness

(2)

The longitudinal stress, σlong, is given by: σlong = 1/2σhoop + x

(3)

The equivalent stress is given by: σ Von _ Mises = σ hoop 2 + σ axial 2 − σ hoop × σ axial

(4)

The term x in Eq. 3 represents all stress except for that set up by the pressure. A piping engineer should be consulted when determining the value of x. It is recommended that the UTS by reduced by 20% or more, depending on the reliability of the UTS data. The 20% is a safety margin. Reduce the wall thickness by accounting for the mill tolerance. It must be assumed that the lower mill tolerance is delivered. Reduce the strength by including the weld factor. Again, a piping engineer should be consulted. Step 5: Check the rupture pressure against the acceptance criteria. If all piping and equipment in the de-

Required information prior to relief-calculation iteration - Description of the fire scenarios (type of fire, duration, heat fluxes, size) - Relief segment geometry (system volume, area, weight, etc.) - Ultimate tensile strength at elevated temperature of materials in the relief section - Acceptance criteria for rupture

Add insulation if required. Calculate the process segment pressure profile. Use the fire with the largest heat input (kW).

Calculate the wall-temperature profile for all pipes and equipment. Use the local fire with the highest heat flux (kW/m2).

Use the temperature profile to calculate the rupture pressure for all pipes and equipment. Compare with the actual pressure.

Acceptance criteria: - Pipe rupture pressure - Equipment rupture pressure - Released flammable fluid at rupture - Time to rupture - No rupture

Are the acceptance criteria for rupture met?

No

Decide which pipe/equipment to fire insulate.

the flare-system tail-pipe. This depressurization calculation should be performed without fire input to the section. All planned types of insulation (not only fire insulation) should be taken into account. The initial temperature should be the minimum ambient temperature or minimum operating temperature, whichever is lower. The initial pressure is calculated from a cooldown of the system down to the start temperature, prior to depressurization. The cooldown calculation should be performed using the trip pressure from the highest shutdown pressure. The minimum temperature in the flare tail-pipe should be calculated with the depressurization segment as the only source to the flare system.

Other considerations Some key ones to note are: • The loss of bolt pre-tensioning due to bolt elongation as a result of increasThe design of this section fire-relief valve and fire-insulation requirements is finished. ing temperature is important when studying flange failure. The piping engineer should be consulted on this. Flanges ■ Figure 3. Sizing procedure for fire-relief valves is similar to that for orifices. are recommended to be fire-insulated. • Lines in the flare system having no flow during a fire depressurization (e.g., downstream pressurization segment meet the acceptance criteria, then pressure-control and pressure-relief valves) are usually the fire insulation is completed. Go to Step 7 for low-temfire-insulated because: they are thin-walled and can heat up perature calculation, otherwise go to Step 6 and add in inrapidly; the depressurization gas flowing in the flare syssulation. Alternatively, go back to Step 1 and increase the tem does not cool these pipes; and the flare system will be size of the orifice or increase the flare system capacity. pressurized to a value near its design pressure, at the same Step 6: Decide which piping/equipment to fire-insutime that the pipe temperature is high. late. If any run of piping or piece of equipment does not meet the acceptance criteria, then add PFP to one or more Fire-relief-valve sizing procedure of these components. It is recommended to add PFP to Sizing of the fire-relief valves (Figure 3) should be the corrosion-resistant pipe with the largest diameter. But, performed with the same minimum requirement as specif there are pipes that are already insulated for reasons ified at the beginning of the article. Also, the procedure other than PFP, these should be fire-insulated first. closely follows that for sizing depressurization orifices, The reasons for choosing the pipe with the largest diwith the exception, of course, that the pressure will inameter are it is the most critical with respect to reaction crease until the relief valve opens. The size of the relief forces and pressure waves when it ruptures, and it will revalve should be such that the minimum relief-rate quire the largest amount of insulation per length. Large equals the liquid boil-off and gas-expansion rates. This pipes are also cheaper to paint and insulate (per unit area) avoids a pressure increase above the set pressure of the than smaller ones. The reason for insulating the corrorelief valve. sion-resistant pipes first is to avoid insulation on metals A fire-relief valve will usually not protect a pressurthat corrode more easily. When partially insulating pipes, ized system against rupture if the gas-filled part of the it is preferable to add the covering on an area where the system is exposed to fire. The fire-relief valve will norpossibility of a fire is largest and where inspection of the mally protect against rupture if the flame is exposed to insulation and pipe can be performed easily. the wetted wall only when the boiling liquid on the inStep 7: Calculate the design low-temperature limitaside keeps the wall temperature at a reasonably low tion of the depressurization segment (this is known as value. For multicomponent mixtures, the temperature the minimum design-temperature calculation) and in

CEP


43

Safety Antisurge Valve

Blowdown Orifice

To the Flare System Blowdown EV

Process Stream Sectionalization Valve EV or XV

PSV = Pressure Safety Valve EV = Emergency Valve XV = Sectionalization Valve

To the Flare System

PSV

Cooler

Compressor

Sectionalization Valve EV or XV Process Stream

The real geometry of any type of process segment (e.g., the above system drawn in continous lines) is transformed into a hypothetical cylindrical vessel (the vessel below in continuous line). The hypothetical vessel is used in calculation of system pressure during depressurization and relief. The diameter and length are increased until the volume equals the volume of the original system (use the dominating diameter of the original system). The liquid level is adjusted to match the liquid volume of the original system, "Add or subtract" wet and dry areas to match the wet and dry area of the original system. Set a wall thickness equal to the dominating wall thickness of the original system and adjust the weight until the weight matches the original sysem weight. Hence, we have a hypothetical segment where the system area and weight do not fit the hypothetical cylinder, but do fit the original system.

■ Figure 4. Modeling

the real process segment as a hypothetical one simplifies sizing procedures.

Blowdown valves open in fire situations.

Process Stream

Scrubber Sectionalization Valve EV or XV

Sectionalization valves close in emergency situations.

PSV

Flare System Blowdown EV

Flare System Blowdown Orifice

Length Di

will increase as the lighter components evaporate, and the wall will eventually reach the rupture temperature, even if it is liquid-filled.

Modeling process segments Here, we present an approach to modeling complex depressurization and relief segments. The method models the complete complex geometry by creating one hypothetical segment that represents the total system volume and heat-transfer areas, and several sub-segments that represent the “real” geometries of the segment. The hypothetical segment is used for calculation of the system pressure during depressurization or relief. The subsegments are used for calculation of the temperature response of each piece of piping and equipment within a process segment. The sub-segments are modeled with the same geometric information that is required for a wall-temperature-response calculation — namely, the sub-segment wall thickness, segment or pipe diameter and inside fluid.

44

www.cepmagazine.org

September 2002

CEP

Hypothetical segment used for system pressure The hypothetical segment is modeled with the real system volume, system outer-area, with and without PFP, system inside-area in contact with the gas, system inside-area in contact with the liquid, and system weight (of piping and equipment). The hypothetical segment is modeled as a cylinder. This shape is specified with a diameter equal to the mostdominating (volume) diameter of any item of piping or equipment in the original segment. The associated wall thickness for this diameter should be used. The length of the cylinder is set such that its volume equals the volume of the original segment. The liquid level is adjusted to obtain the actual liquid volume. The hypothetical cylinder now represents the correct system volume. However, the outer area of this cylinder must be corrected to the realsystem outer area by subtracting or adding gas and liquid area to the segment. Therefore, heat transfer with the surroundings (fire, ambient air) will be modeled over the real

area. The system outside and inside areas may be different for a high-pressure system (typically, 100–150 bar), with large amount of small-diameter piping (usually, 3 in. and below), and should similarly be corrected to the real system inside area. This effect is largest for pipes with a corrosion allowance, that is, those made of carbon steel, generally. The dominating (by weight) metal quality should be specified. Unique geometry used for wall-temperature calculations. The temperature response of a pipe or piece of equipment exposed to fire depends upon the metal wall thickness, metal properties (e.g., heat capacity and conductivity) and the thermal mass of the inside fluid. Each combination of these must be calculated individually to determine the possible different wall-temperature responses of the system. The exact inside and outside diameters (I.D.s and O.D.s) should be used. The I.D. will determine the thermal mass of the inside fluid. The O.D. will enable calculation of the fire-exposed area. Any length can be used, since the sub-segment volume and mass are proportional to the length. The actual thickness, minus a corrosion allowance, should be used. The actual thickness is the nominal wall thickness, minus the allowable tolerance of the pipe thickness (the mill tolerance). The corrosion allowance should be accounted for after reduction by the mill tolerance. Typical values of the mill tolerance are 1.5–3 mm for carbon steel, with the larger value being

Literature Cited 1. Institute of Petroleum, “Guidelines for the Design and Protection of Pressurized Systems to Withstand Severe Fires,” Inst. of Petroleum, U.K. (to be issued shortly). 2. Scandpower Risk Management AS, “Guidelines for Protection of Pressurised Systems Exposed to Fire,” Scandpower AS, Norway (www.scandpower.com/?CatID=1071) (May 13, 2002). 3. Stange, E., et al., “Determination of Temperatures and Flare Rates During Depressurization and Fire,” paper presented at 72nd Annual Gas Processors Association Convention, San Antonio, TX (1993). 4. American Petroleum Institute, “Guide for Pressure-Relieving and Depressurizing Systems,” API Recommended Practice (RP) 521, 4th ed., API, Washington, DC (1997). 5. Incropera, F. P., and D. P. DeWitt, “Fundamentals of Heat and Mass Transfer,” 4th ed., John Wiley, New York (1996). 6. Health and Safety Executive, “Joint Industry Project on Blast and Fire Engineering of Topside Structures,” OTI 92 596/597/598, HSE, U.K.,(1991). 7. Norsok Standard, “Technical Safety,” S-001, Rev. 3, (www.nts.no/ norsok) (2000).

Acknowledgment We would like to thank Erik Odgaard, Jan A. Pappas and Geir Johansen (Norsk Hydro, safety and piping discipline) for their helpful discussions.

more common. The mill tolerance is specified on the pipe data sheet. When this tolerance is not specified by the pipe supplier, consult a piping engineer. The correct metal or alloy composition must be used to obtain the correct material properties (i.e., heat capacity, conductivity and the UTS). If the UTS is not available at elevated temperatures, a preliminary tensile curve can be made based on the UTS at 20°C; this is usually available. When the strength at an elevated temperature is known for a material that is close in physical properties (the same family of materials, such as two different carbon steels), then the percentage difference in the UTS at 20°C is kept at all temperatures. That is, the new tensile strength curve should have the same shape as the known curve. When the tensile strength at elevated temperatures is not known for a material in the same family, then the percentage difference between the UTS and yield at 20°C is reduced linearly between 20 and 1,000°C. The UTS should be released by at least 30% (not 20%) in such an approximation. When the UTS is equal to the yield strength in the above calculation, then the UTS should be set equal to the yield strength at higher temperatures.

Concluding remarks We believe that the calculation of longitudinal stress represents the major challenge when performing depressurization and fire relief design. Modeling of time-dependent fire characteristics (the extent and heat flux) also represents a challenge, since the plant layout is usually unknown during the design stage. Yield and UTS data at temperatures above 400–500°C often do not exist for the materials used in the system. ◆ PER SALATER is a principal engineer in Norsk Hydro ASA (N-0246 Oslo, Norway; Phone: +47 22 53 76 91; Fax: +47 22 53 95 37; E-mail: [email protected]). He has ten years of experience as a process and system engineer for Norsk Hydro’s North Sea oil and gas facilities. His areas of expertise are system design, heat-exchanger thermal design and design of depressurization systems. He has with Sverre Overaa co-patented (PCT/NO99/00123, U.S. Patent 09/673467) a design that eliminates the flare system for oil-and-gas processing facilities and replaces it with a blowdown header connected to storage. He holds an MSc in mechanical engineering from the Norwegian Institute of Technology, Trondheim. SVERRE J. OVERAA is a principal engineer at Norsk Hydro ASA (Phone: +47 22 53 81 00; Fax: +47 22 53 27 25; E-mail: [email protected]). He is a member of the GPA Technical Section F Research Committee. Overaa has 18 years of experience as a process and systems engineer for Norsk Hydro’s North Sea oil-and-gas facilities and is currently head of technical systems for the Ormen Lange project under development by Norsk Hydro. His areas of expertise are simulation, fluid properties and system design. He has with Per Salater co-patented PCT/NO99/00123, U.S. Patent 09/673467. ELISABETH KJENSJORD is a process engineer with Norsk Hydro ASA (Phone: +47 22 53 81 00; Fax: +47 22 53 27 25; E-mail: elisabeth.kjensjord@ hydro.com). She is involved in process design of Norsk Hydro’s oil-and-gas facilities and is presently engaged in the development of the Grane oil field located on the Norwegian Continental Shelf in the southern part of the North Sea. She holds an MSc in chemical engineering from the Norwegian Univ. of Science and Technology.

CEP


45

1

Pipes exposed to medium sized jet fires - rupture conditions and models for predicting time to rupture Author: Per Salater Co-author: Sverre J Overa Norsk Hydro ASA, Oslo, Norway Presented at FABIG, London and Aberdeen, January 2004 and Houston, March 2004 Introduction This article presents data from experiments on pressurized steel pipes exposed to medium sized jet flames together with engineering models for predicting time to rupture. In 1996 and 1997 Norsk Hydro, Norway’s second largest oil company, performed several medium scale jet-fire tests on pressurized pipes. Pipes were pressurized with nitrogen to approximately 85-90% of the design pressure of the pipes and then exposed to a 170 - 190 kW/m2 jet-fire (incident heat flux). The pressure was kept constant during the tests, and internal wall and gas temperatures were measured until the pipes ruptured. Some of the test results are presented and discussed in this article. The tests were performed to verify Norsk Hydro’s engineering methods /1/ for calculation of time to pipe rupture. The Norsk Hydro methods have later been included in the Institute of Petroleum’s “Guidelines for the design and protection of pressurized systems to withstand severe fires” /2/ and the Norwegian “Guideline for Protection of Pressurized Systems Exposed to Fire” /3/. The tests, together with several cold (without fire) depressurization tests of both large process systems and small vessels, have confirmed that Norsk Hydro’s process simulation tool “NEW*S” /4/ is an engineering tool that reproduces experimental depressurization and fire tests. Norsk Hydro has since 1992 used NEW*S in the design of depressurization orifices and fire relief valves on it’s oil and gas installations. NEW*S has also been used to determine the required amount of fire insulation and the minimum design temperatures for pipes, vessels and flare systems. At present NEW*S is used in the design of the Ormen Lange onshore facility, a gas plant (70 MSm3/day) to be built on the west cost of Norway with start up in 2007. NEW*S supports most of the requirements given in reference /2/ and /3/. The jet fire test rig The tests were performed by SINTEF (Trondheim, Norway). The test rig is illustrated in Figure 1. See also Photo 1. The jet-fire was a 6 MW propane flame. The propane flow rate was 0.13 kg/s and the incident heat flux from the jet-fire was between 170 and 190 kW/m2 at its highest and 150 kW/m2 in average for the entire circumference of the axial center of the pipe. In comparison, the local, peak heat flux of the severe jet-fires in reference /2/ and /3/ are 250 kW/m2 for flow rates between 0.1 and 2 kg/s and 350 kW/m2 for flow rates larger than 2 kg/s. The fires in reference /2/ and /3/ are “optical thick” flames, i.e. there is no radiant heat exchange between the pipe (object) and the cold surroundings outside the flame. When comparing the experiments with the calculated temperatures, it can be concluded that an optical thick flame was achieved for the small diameter pipes ( d 8”), but not for the large diameter pipes ( t 10”).

2

~ 6 meters

~ 1 meter

3.5 meters Test pipe

Propane lance

Jet fire

Figure 1:

Piperack The test rig

Photo 1:

The test rig

3

The length of the test pipe sections was 3 meter. The pipe was free to expand on the pipe rack, hence no stress due to thermal expansion was created and the pipe stress was primarily due to the pressure. The jet-fire was an open fire. A windshield wall was placed on one side of the test rig. The jetfire was approximately 1 meter in diameter at the pipe location. The flame length was approximately 6 meters and the distance from the propane lance to the pipe was approximately 3.5 meters. The fire tests were carried out outdoors and the ambient conditions (varying wind and temperature) influenced to some degree the fire heat flux. For all the tests the inside wall temperature was measured at 4 locations over the circumference of the axial center of the pipe inner wall as illustrated in Figure 2. The center of the pipe was also the center of the jet-fire. These 4 temperatures are reported in Figure 3 to 9. A pressure control valve was installed on all the pipes, and the pressure was kept constant ( r 3 bar) during the tests.

3 mete

rs

"Top" Gas top

Pipe and flame centre O

90 "Back"

180O

Flame direction 0O

"Front"

270O Gas bottom "Bottom"

Piperack

Figure 2:

Location of thermocouples. The location of the inner wall thermocouples was identical for all the tests.

For half of the tests several more thermocouples were installed. In these tests the internal gas temperature and inner wall temperatures were measured in the axial direction from the center of the pipe. 6 out of 16 tests are presented in detail in this article.

4

Table 1:

Tests that are presented in detail in this article. Pipe material Pipe design pressure Pressure during test Pipe dimension (barg) (barg) (inch) Test 1 Carbon steel 100 1 *) 10 Test 2 Carbon steel 50 46 4 Test 3 Carbon steel 450 98 8 Test 4 Carbon steel 50 48 10 Test 5 Carbon steel 100 82 10 Test 6 22 Cr Duplex 100 87.5 10 *) Test performed to measure the temperature difference between the inner and outer wall Temperature difference between the inner and outer wall

The temperature difference between the inner and outer wall was measured only in one initial test. This test was run at atmospheric pressure on a 10”, 100 bar carbon steel pipe. The temperature gradient was 3.4oC/mm at its maximum at the location where the highest temperature was measured. At this temperature gradient the radial heat flux in the wall is calculated to 130135 kW/m2. The total heat flux was somewhat higher since there was heat transfer also in the axial direction and in the circumferential direction. The radial temperature gradient is presented in Figure 3. It increases rapidly up to 3oC/mm, then there is a further increase up to 3.4oC/mm before it starts to decline. The increase from 3 to 3.4oC/mm may be due to a change in wind (flame) condition. The temperature difference would have declined towards zero if the test had been prolonged until the wall temperature had stabilized, i.e when the net heat input to the pipe (incident heat from the fire minus re-radiation from the pipe) became zero. As can be seen from Figure 3 the temperature was still increasing when the experiment was terminated. Inner wall temperature

Radial temperatur gradient in wall 800

3.5

700 Temperatur (Celsius)

4

Kelvin/mm

3 2.5 2 1.5 1

600 500 400 300 200 100

0.5

0

0 0.0

1.0

2.0

3.0 Time (minutes)

Figure 3:

4.0

5.0

6.0

0.0

1.0

2.0

3.0

4.0

5.0

Time (minutes)

Differential temperature between inner and outer wall + inner wall temperature, test #1

6.0

5

Temperatures The tests illustrated in Figures 4, 6, 8, 9 and 10, all terminated with a pipe rupture and the time to rupture is at the end of each curve. The presented wall temperatures are measured at the inner wall of the pipe. The outer wall temperature is higher than the presented temperatures, but the difference will, as explained above, approach zero as the temperature increase pr time (dT/dt) approaches zero. Hence, the measured inner wall temperature at rupture is close to the outer wall temperature in most experiments. The calculated temperatures are mean wall temperatures, i.e. average temperature in the radial direction. The calculated temperatures are presented together with the measured temperatures in the figures.

Temperature (C)

4", 50 bar design pressure, Carbon steel Test run at 46 barg 1000 900 800 700 600 500 400 300 200 100 0

Top Calculated

Calculated Wall top Wall front, 100 mm left of centre

Wall bottom Rupture

0

Figure 4:

Back Bottom

2

4

6 8 10 Time (min)

12

Wall back

14

16

Temperature profiles for the 4” carbon steel pipe of 50 bar design pressure operating at 46 barg, test # 2

For all the experiments the highest heat flux and wall temperature was located at position 0o (see Figure 2), i.e. at the side of the pipe facing towards the source of the jet fire. The second highest temperature was in most experiments at the top of the pipe (90o) and the lowest at the bottom of the pipe (270o). In test 4 (see Figure 6) it is probable that the jet-fire was lower compared to the other tests since the bottom temperature is the second highest. The circumferential position of the highest heat flux would have moved if the pipe had been moved closer to the jet flame, hence it cannot be concluded that the highest wall temperature always will be towards the source of the jet fire.

6

Temperature (C)

8", 450 bar design pressure, Carbon steel Test run at 98 barg 1000 900 800 700 600 500 400 300 200 100 0

Calculated

Front

Back

Calculated

Bottom

Wall top Top

Wall front Wall bottom Wall back

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 Time (min)

Figure 5:

Temperature profiles for the 8” carbon steel pipe of 450 bar design pressure operating at 98 barg, test # 3. This pipe did not burst.

The temperature increase is highest at the start of the experiments and then decreases until the pipe ruptures or the flame is shut off. For the small diameter pipe in test # 2 the temperature increase is initially 7.4oC/sek, for test # 4 5oC/sek and for test # 3 1.8oC/sek. The difference is approximately a linear function of the pipe outer area-weight ratio. Table 2 summarizes the measured wall temperature responses. At 6-700oC the Ultimate Tensile Strength (UTS) – the material strength - of typical steel material is largely reduced. Table 2: Summary of measured wall temperature responses Temperature Time to raise the wall temperature for a 175 kW/m2 jet fire [min:sec] 4”, 50 bar, 10”, 50 bar, 10”, 100 bar, 10”, 100 bar, 8”, 450 bar, carbon steel carbon steel 22Cr Duplex carbon steel carbon steel *) 200oC 0:23 0:42 0:54 1:04 2:18 o 400 C 1:02 1:57 2:12 3:00 4:44 600oC 2:10 4:06 4:14 6:22 8:27 o 800 C 5:14 8:00 17:03 *) This was a 190 kW/m2 fire

7

10", 50 bar design pressure, Carbon steel Test run at 48 barg

1000 900

Calculated

Temperature (C)

800

Front Top & back

700 Bottom

600

Calculated Wall top

Gas top

500

Wall front Wall bottom

400 Gas bottom

300

Wall back Gas top Gas bottom

200 Rupture

100 0 0

1

2

3

4

5

6

7

8

9

10 11 12 13

Time (min)

Figure 6:

Temperature profiles for the 10” carbon steel pipe of 50 bar design pressure operating at 48 barg, test # 4

The highest internal gas temperature is measured at the top of the horizontal pipe, the lowest at the bottom (see Figure 6). The temperature difference is as high as 200oC, indicating a large free convection heat transfer inside the pipe. This temperature stratification is seen in all the experiments, and is also what we have measured in our several “cold” blowdown experiments, some of them presented in /5/. The free convection heat transfer term must be included in the heat transfer calculations. The fire model used in the wall temperature profile calculations is: Qnet = Ds HfV Tr4 + h(Tf - Ts(t)) - HsV Ts(t)4

(1)

The first term is the radiation heat from the flame to the pipe, the second term is the heat convection from the flame gases and the last term is the re-radiation from the pipe. Qnet Hs Ds Hf V Tr Tf

net (absorbed) heat transfer to the pipe (W/m2) emissivity of the pipe material (-) absorbtivity of the pipe material (-) emissivity of flame (-) Stefan-Boltzmanns constant = 5.6710-8 (W/m2 K4) radiation temperature of flame (K) flame gas temperature (K)

8

Ts(t) h

surface temperature of the material (K) convective heat transfer coefficient (W/m2K)

This fire model is equal to the fire model in /1/, /2/ and /3/. When comparing calculations with measured temperatures, parameters for equation 1 were chosen to fit a jet-fire with initial incident heat flux of 175 kW/m2 and an initial absorbed heat flux of 145 kW/m2. The parameters used are: x x x x x

Flame radiation temperature 980oC *) Flame gas temperature 980oC *) Flame emisivity 0.7 Pipe emisivity and absorbtivity 0.7 Convective heat transfer coefficient 80 W/m2K *) The fire was premixed for the duplex pipe test. 1020oC is used in the calculation of this test, resulting in a 191 kW/m2 jet-fire (incident heat).

These parameters are within the range of the typical values suggested in /1/ and /2/. In Figure 7 the resulting net (absorbed) heat transfer to the pipe is given as a function of pipe temperature. Note that Equation (1) also can be used to reproduce the open pool fire used in API RP 521. This is discussed in /1/.

Net (absorbed) Heat Flux [kW/m2]

The fire model used to reproduce the experimental medium scale jet fire 175 150 125 100 75 50 25 0 0

100

200

300

400

500

600

700

800

900

1000

Wall temperature [Celsius]

Figure 7:

Net heat flux to pipe using the above fire parameters

As seen in Figures 4, 5, 6, 8 and 9, the calculated wall temperatures with the above chosen parameters are close to the measured temperatures for the small diameter pipes, but too high for the large diameter pipes at temperatures above 600oC. This is partly due to the fact that the fire model assumes an optical thick flame, i.e. there is no radiant heat exchange between the pipe (object) and the cold surroundings outside the flame.. For the 10” pipes, the radiation to the cold surroundings seem to dominate more and more above 600oC. For the small diameter pipe this is not the case to the same degree since the diameter is small compared the diameter of the jet-fire.

9

Another reason is that the flame temperature increases towards the center of the flame. For the 10” pipe the average flame temperature at the wall location was colder than for the 4” test. The highest measured wall temperature was observed on a 1” pipe, the second highest on a 2” pipe.

Temperature (C)

Although the fire temperature at the wall location varied for the different pipe sizes, the same fire temperature is used in all the wall temperature calculations. By changing the fire temperature, a calculated wall temperature closer to the measured temperature will be achieved for the 10” pipes. A constant fire temperature is used here since this normally is the case in engineering calculations.

10", 100 bar design pressure, Carbon steel Test pressure = 82 barg

1000 900 800 700 600 500 400 300 200 100 0

Calculated

Front Back Bottom

Wall top - failed

Gas top

Wall front Wall bottom Wall back Gas bottom

Gas top Rupture

0

Figure 8:

Calculated

2

4

6

8

Gas bottom

10 12 14 16 18 20 22 Time (min)

Temperature profile for the 10” carbon steel pipe of 100 bar design pressure operating at 82 barg, test # 5

10

Temperature (C)

10", 100 bar design pressure, 22Cr Duplex Test run at 87.5 barg 1000 900 800 700 600 500 400 300 200 100 0

Calculated Top Back

Calculated Bottom

Wall top Wall front, 100mm left of center Wall bottom Wall back Rupture

0

1

2

3

4

5

6 7

8

9 10 11 12 13

Time (min) Figure 9:

Temperature profile for the 10” duplex pipe of 100 bar design pressure, operating at 87.5 barg, test # 6

Time to rupture

Photo 2:

During and after test (50 bar, 10”, carbon steel)

The time to rupture for a specified stress level is highly dependant on the pipe wall thickness see equations (2) to (4). The pipe wall thickness is usually not equal to the nominal wall thickness since pipes are specified with a mill tolerance. Carbon steel pipes normally have additional corrosion allowance. The carbon steel pipes in the experiments were specified with 3 mm

11

corrosion allowance. The pipes were new and the entire corrosion allowance was intact. The duplex pipe had no corrosion allowance. The mill tolerance is often specified as r 12.5% of the nominal wall thickness, and the pipe producers are allowed to utilize this tolerance and thus produce a pipe thickness that can be as low as 87.5% of the nominal wall thickness. The lower end of the wall thickness is normally seen for the expensive stainless steel materials, typically -10% since the pipe producers need some margin to –12.5% themselves. For carbon steel the mill tolerance varies between r 12.5%. For all the experiments the pipe ruptured at the location of the highest temperature, i.e. at location 0o (see Figure 2). Measured time to rupture is presented in Table 3. Also presented is calculated time to rupture when using the highest measured wall temperature, varying wall thicknesses and the pipe failure criteria described below. Table 3:

Time to rupture – measured and calculated. The measured temperatures are used in the calculations **).

Pipe material

Carbon steel, 4” 50 bar, Test # 2 Carbon steel, 8” 450 bar, Test # 3 Carbon steel, 10” 50 bar, Test # 4 Carbon steel, 10” 100 bar, Test # 5 22 Cr Duplex, 10” 100 bar, Test # 6

Measured

Calculated – using the nominal wall thickness

Calculated – wall thickness reduced with 10% mill tolerance

(Minutes)

(Minutes)

(Minutes)

Calculated – wall thickness reduced with 3 mm corrosion allowance **) (Minutes)

Calculated – wall thickness reduced with 3 mm corr. allowance and 10% mill tolerance (Minutes)

12.8

12

7.5

3.3

3

>25 *)

>25 *)

>25 *)

>25 *)

>25 *)

10.1

9.2

8.6

7.4

6.6

18.8

18.8

15.2

13.6

12.4

10

8.7

7.9

Not applicable

Not applicable

*) Did not burst and rupture is not calculated. **) The measured temperature profile would have been steeper if the corrosion allowance had been removed. Table 4 presents time to rupture when using the calculated temperature profile for a pipe without corrosion allowance. The time to rupture is calculated as proposed in /3/ where the stress is calculated from equation (4) and compared with the material Ultimate Tensile Strength (UTS). In /3/ the recommended pipe failure criteria is when the equivalent stress is larger than the UTS. ( p OD) V hoop (2) Hoop stress: 2w ( p OD) V axial V ext V displ Longitudinal stress: (3) 4w

Equivalent stress:

V

V hoop 2 V axial 2 V hoop V axial

(4)

12

where

V hoop V axial V ext V displ

is the hoop stress (MPa)

p OD w

is the pressure (MPa) is the outer diameter (meter) is the wall thickness (meter)

is the axial stress (MPa) is the longitudinal stress due to external loads (weight of pipe and valves etc ), (MPa) is the longitudinal stress due to thermal expansion (MPa)

During all the tests V ext has been close to zero since the pipe span was only 3 meter. V displ was zero. Figure 10 illustrates the calculated time to rupture using the specified failure criteria from /3/ and the temperature measured at location 0o (see Figure 2). The UTS is reduced as the temperature is increased. The UTS is taken from /3/ and corrected for the specified UTS from the pipe datasheet. The procedure for correcting the UTS is specified in /3/. The pipe rupture is predicted where the curves cross. As seen in table 3 and the above temperature measurements, a small diameter low-pressure pipe with the corrosion allowance intact can withstand higher temperatures due to greater wall thickness. By using the failure criteria from /3/, the rupture temperature for a 4”, 50 bar pipe is reduced from 920oC to 730oC and the time to rupture is reduced from 12.8 minutes to 3.3 minutes when the wall thickness is reduced with the corrosion allowance. It is therefore important not to include the corrosion allowance when performing fire depressurization and relief design since the corrosion allowance will eventually not be there. Note that the importance of the corrosion allowance is reduced with increasing design pressure (wall thickness). As an example rupture condition was not reached for the 1” and 2”, 50 bar pipes operating at 45 bar – both pipes with the corrosion allowance intact and reaching a wall temperatures above 1000oC at the end of the test. From the curves in Figure 10 one can see that when the UTS curve is steep around the time of rupture, a small error in the calculated pipe stress is of minor importance to the estimated time to rupture. If however, the UTS decreases slowly with time prior to rupture, a small error on the pipe stress causes a large error in the estimated time to rupture. Also, as rupture takes place at high temperature, correct UTS data is more important at high temperature than at low temperatures.

13

Stress in pipe v.s. UTS 4", 50 bar, carbon steel, 46 barg operating pressure

Stress in pipes v.s. UTS 10", 50 bar, carbon steel, 48 barg operating pressure 200

200

175

175

Rupture

Rupture

150

125

Stress (MPa)

Stress (MPa)

150

100 75

125 100 75

50

50 25

25 0 0

2

4

6

8

10

12

14

0 0

Time (minutes)

4

6 8 Time (minutes)

10

12

Stress in pipe v.s. UTS 10", 100 bar, 22Cr Duplex, 87.5 barg operating pressure

Stress in pipe v.s. UTS 10", 100 bar, carbon steel, 82 barg operating pressure

200

200

175

175

Rupture 150

Stress (MPa)

150 Stress (MPa)

2

125 100 75

125 100 75

50

50

25

25 0

0 0

Figure 10:

2

4

6

8 10 12 Time (minutes)

14

16

18

20

0

2

4

6

8

10

Time (minutes)

Calculated rupture compared with actual rupture when using the failure criteria from /3/. Rupture is calculated when the curves are crossing. Actual rupture is at the far right end of the horizontal curves.

Legend for figure 10: - Blue (declining curve) = UTS as a function of measured temperature. - Red (lowest horizontal) = Stress in pipe, corrosion allowance intact, wall thickness is not reduced with the mill tolerance, i.e. the nominal wall thickness. - Yellow (second lowest horizontal, highest horizontal for duplex) = Stress in pipe, corrosion allowance intact, wall thickness reduced with the mill tolerance. - Turquoise (second highest horizontal) = Stress in pipe, wall thickness reduced with the corrosion allowance (3mm), wall thickness is not reduced with the mill tolerance. - Green (highest horizontal) = Stress in pipe, wall thickness reduced with the corrosion allowance (3mm) and the mill tolerance.

12

14

From Figure 10 one can see that the calculated time to rupture is either correct or slightly conservative for the carbon steel pipes for which the pipe stress was most likely given by the red curve (the lowest horizontal curve). For the duplex pipe the yellow pipe stress curve (the highest horizontal) is the most likely curve, and the calculation is still good though more conservative. Revision 2 of the Norwegian “Guideline for protection of Pressurized Systems Exposed to Fire” /3/ will be available in February 2004. The stress equations will be revised in this issue. The revised equations give slightly better calculated results. The guideline is available at no cost from Scandpower Risk Management’s web page, www.scandpower.com/?CatID=1071 The measured and calculated temperatures in Table 3 and Figure 10 are based on pipes with the corrosion allowance intact. If the temperature calculations are performed for the same pipes but with the corrosion allowance removed, the temperature profile will be steeper and the time to rupture shorter, see Table 4. Table 4: Pipe

Time to rupture for pipes without corrosion allowance. The temperature profile is calculated after removing 3 mm wall thickness. Calculated – Calculated – 175 kW/m2 jet-fire 250 kW/m2 jet-fire *)

[Minutes] [Minutes] Carbon steel, 4” 1:50 1:10 50 bar, Test # 2 Carbon steel, 10” 3:30 2:10 50 bar, Test # 4 Carbon steel, 10” 7:40 4:30 100 bar, Test # 5 *) The jet-fire heat flux recommended in reference /2/ and /3/ for an optical thick jet-fire of leak rate between 0.1 and 2 kg/s. From this one can conclude that automatic and rapid depressurization at fire detection is essential to avoid pipe ruptures for thin walled pipes, i.e. pipes with design pressure up to 100 bar. Spurious fire detection must be avoided by a proper design of the fire detection system.

15

Geometry of pipes after rupture

Photo 3:

Typical pipe geometries after rupture. None of the pipes that ruptured created pipe fragments. Left: 50 bar, carbon steel, 10”. Right: 100 bar, carbon steel, 10”.

Table 5: Test

Pipe geometry after rupture (approximate numbers). Bend angle Reduced wall Size of hole thickness at rupture relative to the pipe diameter location (%) o Carbon steel, 4” 14 33 ~ 6xD 50 bar, Test # 2 Carbon steel, 4”, 17o 72 ~ 3xD 100 bar, Test # 7 22 Cr Duplex, 10” 90o 61 ~ 6xD 100 bar, Test # 6

The carbon steel pipe for test 7 is of a more ductile material than test 2. The carbon content is 0.12% and 0.17%, respectively. In test 7 the ductile behaviour was observed as a sudden increase in diameter just before rupture. This was also seen as a reduced system pressure in the last few seconds. For all the tests that ended with a rupture, the size of the hole was more than 2 times the internal diameter of the pipe. Hence, when calculating the amount of gas released through the hole this must include flow from both directions towards the leak hole. A good estimate of the initial leak rate is two times the flow rate through an orifice with a diameter equal to the internal pipe diameter. A repeat test was run for all pipes. The rupture temperature and geometry was in most cases unchanged. None of the pipes that ruptured created pipe fragments. The 10” duplex pipe rupture caused stronger impact on the test rig than the 10” carbon steel pipes.

16

An engineering approach to the calculation of time to rupture

A “to the point” engineering approach to fire relief and depressurization calculation is presented in /1/. For a more comprehensive description of the topic see /2/ or /3/ of which guideline /3/ is more useful in engineering calculations since it gives more recommendations than /2/ and stress equations and material data are also presented. In brief, the engineering procedure is: x Determine the fire scenario (type of fire, heat flux, fire duration and extent of fire) x Determine the system volume, areas and weight x Determine the acceptance criteria for rupture (what is unacceptable, i.e. maximum pressure at rupture, minimum time to rupture, maximum release of hydrocarbons). x Determine the UTS at elevated temperatures for all pipe and equipment materials in the system. x Calculate the pressure profile of the system. Use this pressure profile in the stress calculations. x Calculate the temperature profile for all the pipe dimensions and equipment in the depressurization/relief segment. Use this temperature profile in the calculation of the maximum allowable stress (UTS). The pressure and temperature calculations should be calculated simultaneously to speed up the calculations. The temperature calculations must include the heat transfer to and the thermodynamic (expansion or compression) of the inside fluid x Compare the pipe stress with the maximum allowable stress for all pipe dimensions. The same to be performed for the equipment. x If a rupture is calculated and this is an unacceptable rupture, change the design (add fire insulation, increase size of depressurization orifice, change material, increase system design pressure or combine any of these measures) and update the pressure profile if necessary. Since 1992 Norsk Hydro has used the in-house process simulation tool NEW*S when performing depressurization and relief calculations. NEW*S is also used to determine the fire duration by performing simultaneous depressurization of a segment to the flare system and to the surrounding through a leaking point. NEW*S is also used when determining minimum design temperatures of depressurization segments. The Norsk Hydro philosophy has always been to verify the models used through experiments, preferably controlled, large scale experiments. These experiments have shown the importance of employing correct thermodynamic and heat transfer models when performing depressurization calculations. The experiments and 12-15 year of design experience (offshore and onshore plants) have enabled us to determine what is important and what is less important. A lot of this experience has been made available in /1/, /2/ and /3/. We have seen from our latest offshore and onshore projects that by using these guidelines, an unexperienced but competent process or safety engineer with some piping knowledge, can produce results with good quality if given a simulation tool that can carry out correct pressure and temperature calculations. Acknowledgement The authors would like to thank Elisabeth Kjensjord for helpful input and comments.

17

References

/1/ /2/ /3/

/4/ /5/

“Size Depressurization and Relief Devices for Pressurized Segments Exposed to Fire”, Salater, Overa and Kjensjord, CEP (AIChe), (September 2002) “Guidelines for the design and protection of pressurized systems to withstand severe fires”, The Institute of Petroleum, (March 2003) “Guideline for protection of Pressurized Systems Exposed to Fire”, Rev 1, Norsk Hydro, Statoil, Scandpower Risk Management, (May 13 2002). Available at no cost from www.scandpower.com/?CatID=1071 “Developing, implementing and verifying an Engineering Tool”, Wilson, Overa, Stange and Majeed, 69th Annual GPA Convention, (March 11-12, 1991) “Determination of Temperatures and Flare Rates During Depressurization and Fire”, Overa, Stange and Salater, paper presented at the 72nd Annual Gas Processors Assosiation Convention, San Antonio, TX, (1993)

Depressuration Report

Recommend Documents