STHE Overpressure Protection Colin Deddis, Senior Process Engineer, EPT 22 March 2010
STHE Overpressure Protection Changes in guidance & practice since 2000 Response times of relief devices Dynamic analysis of STHE overpressure and relief Defining the problem with implementation Incident examples Design & operational issues JIP Proposal
Changes in Guidance – API521/BS EN ISO 23251 Two-thirds rule replaced with: “Loss of containment of the low-pressure side to atmosphere is unlikely to result from a tube rupture where the pressure in the low-pressure side (including upstream and downstream systems) during the tube rupture does not exceed the corrected hydrotest pressure” “Pressure relief for tube rupture is not required where the low-pressure exchanger side (including upstream and downstream systems) does not exceed the criteria noted above.” Dynamic analysis: “This type of analysis is recommended, in addition to the steady-state approach, where there is a wide difference in design pressure between the two exchanger sides [e.g. 7 000 kPa (approx. 1 000 psi) or more], especially where the low-pressure side is liquid-full and the high-pressure side contains a gas or a fluid that flashes across the rupture. Modelling has shown that, under these circumstances, transient conditions can produce overpressure above the test pressure, even when protected by a pressure-relief device [64], [65], [66]. In these cases, additional protection measures should be considered.”
Changes in Guidance – API521/BS EN ISO 23251 Tube rupture design basis: “The user may perform a detailed analysis and/or appropriately design the heat exchanger to determine the design basis other than a full-bore tube rupture. However, each exchanger type should be evaluated for a small tube leak. The detailed analysis should consider a) tube vibration, b) tube material, c) tube wall thickness, d) tube erosion, e) brittle fracture potential, f) fatigue or creep, g) corrosion or degradation of tubes and tubesheets, h) tube inspection programme, i) tube to baffle chafing.”
Current Practice API521/BS EN ISO 23251 allows use of relief valves or bursting disks but states: “The opening time for the device used…..should also be compatible with the requirements of the system.” Opening times of relief valves considered to be too slow, hence bursting disks commonly used. Advances in heat exchanger design practice e.g. vibration analysis, materials etc. have decreased likelihood of tube rupture
Response Times of Relief Devices Bruce Ewan, University of Sheffield
Summary of test conditions and test numbers – phase 1 Relief device Relief diameter (in) 4mm orifice 8mm orifice 15mm orifice Relief pressure (bar) Open tube 39 38 37 Graphite disc 4 51 50 49 10 6 55 54 53 Stainless steel disc 41 42 40 15 (reversed dome) 8 48 47 46 2” Spring loaded pop action RV - 59 58 57 2” Bellows RV 62 61 60 2” Pilot operated RV 66 65 64
High pressure test 4” graphite disc. Rupture time = 1.9 ms
Low pressure test 2” spring loaded RV. 110% open capacity in 6 ms
High pressure test 2” spring loaded RV. 110% open capacity in 4 ms
Low pressure test 2” pilot operated RV. 110% open capacity in 4 ms
High pressure test 2” pilot operated RV. 110% open capacity in 2.5 ms
Summary of test conditions – phase 2 Relief Device Size Driver Pressure (barg) 4mm Orifice 8mm 15mm Safety Valve (SRV) 2 H 3 100 Test no. 2 Test no. 1 Test no. 3 4 L 6 Test no. 23 Test no. 24 Test no. 25 Relief Valve 4 in Test no. 21 Test no. 20 Test no. 19 Stainless Steel Disc 3 in Test no. 6 Test no. 5 Test no. 4 Test no. 7 Test no. 8 Test no. 9 Graphite Disc 20 Test no. 18 Test no. 16 Test no. 15 Test no. 13 Test no. 12 Test no. 11
4L6 safety relief valve 4” relief valve
Low pressure test 4L6 safety. 110% open capacity in 10 ms
High pressure test 4L6 safety. 110% open capacity in 4 ms
SRV, RV and Graphite Disc at High Pressure
Dynamic Analysis of Tube Rupture Ian Wyatt, Atkins
Dynamic Modelling of Tube Rupture Ian Wyatt - Atkins JIP on Bursting Disks for Shell & Tube Exchangers – 1st Stakeholders Meeting
API-521/BS EN ISO 23251 – 5.19 API-521.BS EN ISO 23251 does not dictate what has to be done: If a steady-state method is used, the relief-device size should be based on the gas and/or liquid flow passing through the rupture. A one-dimensional dynamic model can be used … This type of analysis is recommended, in addition to the steady-state approach, where there is a wide difference in design pressure [e.g. 7 000 kPa … There is a warning at the bottom: Modelling has shown that, under these circumstances, transient conditions can produce overpressure above the test pressure, even when protected by a pressure-relief device ...
Different Exchanger Configurations Similar Tube Rupture consequences apply to all of these configurations: Single pass gas, single pass liquid Multiple pass gas and/or multiple pass liquid HP Gas on tube side or shell side Cooling Duty or Heating Duty Horizontal or Vertical or Angled
Stages to Tube Rupture For all configurations there are four phases to the consequences of a Tube Rupture – identified in the tube rupture tests performed as part of the previous JIP: Phase I – Percussive Shock Phase II – Fast Transient Phase III – Liquid Discharge Phase IV – Gas Discharge
Phase I – Percussive Shock Rapid rupture creates percussive shock wave Extremely short lived <0.1ms Shell does not ‘feel’ the pressure spikes Not Model
Phase II – Fast Transient Gas entering shell is faster than time to overcome liquid momentum Fast transient pressure wave results travelling at sonic velocity Pressure wave usually breaks bursting disc Shell and pipework overpressures possible Simulated using software with necessary fast transient capability Shell baffle path ‘straightened’ – 1D Model
Phase II – Fast Transient Gas entering shell is faster than time to overcome liquid momentum Fast transient pressure wave results travelling at sonic velocity Pressure wave usually breaks bursting disc Shell and pipework overpressures possible Simulated using software with necessary fast transient capability Shell baffle path ‘straightened’ – 1D Model
Phase III – Liquid Discharge Gas bubble grows towards exits Liquid displaced through available exits Volume flow balance between bubble and displaced liquid Possible to over pressurise Shell and connected pipework Gas-Liquid interfaces affect pipe supports Shell baffle path ‘straightened’ – 1D Model
Phase IV – Gas Discharge Gas from rupture passes out of system Pseudo steady state depending on gas supply Usually not modelled
Results Relief device does not always protect against over pressure Even some below 2/3rds rule exceed limits – two of them lower pipework design pressures
STHE Overpressure Protection – the “problem” Increased use of bursting disks to protect STHEs over past 10 to 15 years Estimated frequency of guillotine tube rupture 0.0009 per unit per year (~1 per 1,100 years)[1] Frequency of bursting disk failures protecting STHEs 7 incidents in 13 years (~50 exchangers) 0.011 per unit per year (~1 per 90 years)[2] Future growth in numbers of high pressure STHEs requiring overpressure protection Has the balance of risk shifted? IP Guidelines for the Design and Sae Operation of Shell & Tube Heat Exchangers to Withstand the Impact of Tube Failure, Aug 2000 Estimate based on incidents known to BP
STHE Overpressure Protection – the “problem” Two major hazards associated with bursting disk failures: Impairment of relief system – liquid inflow & overfill Incident escalation - reverse rupture leads to uncontrolled hydrocarbon release from relief system
Incident #1 – liquid overfill Relief Header Flare Knockout Drum Flare PSHH Bursting disk rupture in forward direction PSHH in void space of bursting disk assembly fails to isolate exchanger Sustained cooling medium flow into relief system Liquid overfill & potential overpressure of knockout drum
Incident #2 – excessive backpressure 80 psig Burst 80 psig Burst 50 psig 100 psig 225 psig 225 psig Note: The top disc impacted bottom disc causing it to also rupture
Incident #2 ctd.
Any other incidents……? ???
Design & Operational Issues HSE Safety alert 01/2008 Steve Murray, HSE
Bursting disc failure: flare system impairment Stephen Murray HSE Inspector, Offshore Division
HSE Safety Alert 01/2008 Alerts: to advise industry of incidents http://www.hse.gov.uk/offshore/alerts/sa_01_08.htm Alerts: to advise industry of incidents enable lessons to be learned industry takes appropriate action to avoid similar incidents
HSE Safety Alert 01/2008 SWR HP Flare Drum Heat Exch. SWS gas
HSE Safety Alert 01/2008 SWR HP Flare Drum Heat Exch. LP flare drum PAH LAH ESD HP Flare Drum Heat Exch. LP flare drum gas ESDV Closed drain SWS ESDV Over-board
does not trip seawater pumps HSE Safety Alert 01/2008 What happened? press = 4 barg (no alarm) liquid @+40m disc failure does not trip seawater pumps water enters drum tell-tail blocked? no level >LAH SWR PAH no alarm overfills LAH ESD HP Flare Drum Heat Exch. LP flare drum gas ESDV not tight shut-off fills Closed drain fills SWS ESDV Over-board closed
Summary HSE Safety Alert 01/2008 uncontrolled flow of seawater into flare system several hours to identify source flaring event may have lead to serious gas release
Lessons HSE Safety Alert 01/2008 Be aware of potential for impairment of flare/relief system from uncontrolled cooling medium flow from ruptured bursting disc Ensure disc rupture will initiate measures to ensure isolation of cooling medium so that flare/relief system is not compromised
Legal requirements HSE Safety Alert 01/2008 Provision and use of Work Equipment Regs 1998 Management of Health & Safety at Work Regs 1999 Offshore Installations (Prevention of Fire & Explosion and ER) Regs 1995
Bursting disc failure: flare system impairment Stephen Murray HSE Inspector, OSD
Design & Operational Issues Bursting disks utilised for overpressure protection of STHEs Once opened, they maintain an open flow path from the process/utility system to the relief system. A sufficient margin (~30%) must be maintained between operating and set pressure to avoid rupture. In STHE applications, they are often located on cooling medium systems which can be susceptible to pressure surges. Failure in the reverse direction due to superimposed backpressures from the relief system.
Design & Operational Issues Bursting disks utilised for overpressure protection of STHEs Once opened, they maintain an open flow path from the process/utility system to the relief system. A sufficient margin (~30%) must be maintained between operating and set pressure to avoid rupture. In STHE applications, they are often located on cooling medium systems which can be susceptible to pressure surges. Failure in the reverse direction due to superimposed backpressures from the relief system.
Design & Operational Issues Selection of relief route Multiphase – high velocity liquid slugs HP or LP flare system (high pressure gas under relief conditions but large liquid volumes under a failure case) Should relief from STHEs be segregated from other relief routes? Is HAZOP effective at identifying potential failure modes and consequences? Additional protective measures required for failure cases.
Gaps in current guidance Broader design requirements associated with bursting disks and interface with relief systems not addressed At what pressure ratio are relief valves acceptable? Large differential pressure may actually favour relief valve – extent of overpressure may yield sufficiently rapid response Lower differential pressures – shell & nozzles may survive overpressure. What extent and duration of overpressure is acceptable?
Aims of JIP Eliminate or mitigate hazards associated with overpressure protection of STHEs Develop revised set of design guidelines for overpressure protection of STHEs principally to address: Heat exchanger design. Relief device selection.
Heat Exchanger Design (1) Determine criteria to assess if guillotine fracture is possible based on the mechanical properties of the materials of construction used in heat exchanger tubes. Determine any minimum tube thickness specification required to prevent guillotine fracture. Define the vibration analysis requirements that need to be applied to ensure that the likelihood of guillotine fracture is minimised. Define any sensitivity analysis of process variations which should be carried out to ensure that the design is robust.
Heat Exchanger Design (2) Determine if differential pressure limits can be established below which transient effects can be ignored. Determine the maximum allowable transient overpressures in the shell under tube rupture conditions to cater for peak pressures. This will require experimental and analytical work. Determine the impact of transient loads on the piping systems if bursting disks are not applied for overpressure and develop appropriate design guidelines to ensure that the piping design is robust but not overly conservative.
Relief Device Selection Develop a rule-set for relief device selection to accommodate the tube rupture case Scale-up to typical relief device sizes encountered in real applications. Testing of response times of a variety of relief valves to a range of overpressures . Establish mechanical integrity criteria for relief valves for use in tube rupture service. Establish the range of process conditions for which conventional relief valves could be utilised to protect against tube rupture and those for which bursting disks are required. This needs to consider aspects such as differential design pressure between low and high pressure side of exchanger etc.
Shopping list – issues captured in Stakeholders’ meeting Deliverable – software Criteria for selecting RDs Set points – selection criteria Overpressure/overpressure times Testing and inspection Type/size/etc of RD and response capability, affecting selection Design issues including Instrumentation Seawater system Learning from experience – what went wrong: capture findings/lessons learned Construction/operation/maintenance etc of whole system HAZOP – pertinent guide words (like RABS guidelines) HE stress distribution – revisit/extend Sheff Uni work