LNG SHIP INSULATION EXPERIMENTS USING LARGE LNG POOL FIRE BOUNDARY - - PowerPoint PPT Presentation

<Title of Presentation> By: <Author Name>, <Organization> <Date>

<Title of Presentation>

By: <Author Name>, <Organization> <Date>

17th INTERNATIONAL CONFERENCE & EXHIBITION ON LIQUEFIED NATURAL GAS (LNG 17)

LNG SHIP INSULATION EXPERIMENTS USING LARGE LNG POOL FIRE BOUNDARY CONDITIONS

By: Thomas Blanchat, Charles Morrow, and Michael Hightower Sandia National Laboratories April 17, 2013

17th INTERNATIONAL CONFERENCE & EXHIBITION ON LIQUEFIED NATURAL GAS (LNG 17)

Program objective: Improve understanding of large LNG spills and the impact on hazards to the public

• Large LNG Pool Fires:

Goal is to reduce uncertainty in thermal hazard predictions for large-scale LNG pool fires (~100 m).

• Cascading Damage:

Goal is to assess cryogenic and thermal damage to an LNG ship from a spill event.

• Synergy between testing and

simulation.

LNG pool fire test

CFD simulation of LNG pool fire, 512 CPUs, 2.5 M elements Experiment (land based)

Large LNG Pool Fire Testing

• Conduct scaled fire height to diameter tests in controlled

conditions

• Construct a large experimental area to conduct up to 100m in

diameter LNG pool fires on water

• Conduct large-scale LNG pool fire tests (30m, 70m, and 100m

diameter) to determine large LNG pool fire hazards, including: – Large pool fire behavior, physics, and characteristics – Surface emissive power (flame radiant energy) – Fuel vaporization rate (pool size) – Flame height and diameter (view factor)

• From the test data obtained, obtain data to improve existing fire

models to enable better estimates of LNG pool fire properties and behavior at scales of interest (up to 500 m diameter) with complicated conditions and geometries

Fire Dynamics at Large Scale

JP8 – 2 m (SNL) JP8 – 3 m (SNL) JP8 – 20 m (China Lake) LNG - 10 m SNL 2005 LNG - 35 m 1987 Montoir 1987 LNG ~200 m

????

Methane has unique chemical reaction pathways resulting in an order of magnitude less soot production.

Construction

LNG experiment site – West view – 9-6-2008

Large Scale Test Configuration

• polyurethane foam and concrete lined reservoir – 5.2 m liquid depth
• aluminum cover with insulating concrete
• aluminum discharge plugs mate with SS pipe flange/elbows
• reinforced concrete pipes (91 m long) connect to spill diffuser
• 6 ton winches open plugs

Urethane foam and carbon mesh

Large Scale Test Configuration

• 1250 m3 capacity LNG reservoir
• 120 m diameter water pool (2 m depth,

lined)

• 3 discharge pipes (15 in, 24 in, 36 in)
• spill diffuser at pool center
• 11 m3 capacity LN2 trailer (inertion)
• Instrument towers (12) 110 m, 160 m, and

210 m (from pool centerline)

• Data Acquisition Systems (5)
• Reservoir

–

Liquid level (1 pressure, 1 floats)

–

Temperature (8 TCs)

• Pool

–

Spill Area (overhead video (2)

–

Heat flux to surface (13 DFTs)

–

Water temperature (36 TCs)

• Plume

–

Height (12 cameras) (4 high speed, 2 infrared)

–

Spectrometers (4) (400-800 nm, 1300- 4800 nm)

–

Heat flux (radiometers: narrow-angle (28), wide-angle (12)

• Meteorology

–

3D ultrasonic wind speed/direction (4)

–

ambient pressure/temperature/RH (1)

N (340°) E W S

LNG Fire Dynamics at Large Scale

The pool diameter does not determine the flame width on open water.

LNG - 10 m

SNL 2005

LNG - 21 m SNL

2/2009

LNG - 83 m SNL

12/2009

For the 10 m test: Spot SEPave ~190 kW/m2 (optically–thin, radiation vector not saturated)

Surface Emissive Power vs. Pool Diameter Heavy Hydrocarbons Compared to LNG (with large scale test data)

1000

Accident Scale On Land Test Scale

Dpool (m)

Dpool SEP (m) (kW/m2) 10 190 21 277 83 286

*

LNG Test #1

@SS (390-510s) Dpool 21 ± 1 m Wflame 22 ± 3 m Hflame 35 ± 3 m Burn Rate 0.15 ± 0.01 kg/m2s Wind Speed 4.8 ± 0.8 m/s

LNG Test #2

@SS (250-300s) Dpool 83 ± 4 m Wflame 56 ± 12 m Hflame 146 ± 8 m Wind Speed 1.6 ± 0.2 m/s

LNG Cargo Tank Insulation Fire Damage Testing and Relief Valve Capacity Analysis

Test Scope

• Review cargo tanker details that are susceptible to flame impingement and thermal

insult

• Capture the critical physics (to 1st order) and simplify construction details to allow

representative testing between various types

• Obtain or determine thermal properties for prototypic materials

– Small sample measurements of density, specific heat, thermal conductivity or data from manufacturer – Obtain prototypic insulation materials

• Perform insulation decomposition tests

– Various cargo types (Moss and membrane) with various insulation systems – Comparable tests at medium scale – Data from tests will be used to develop/validate models for thermal simulations to inform of damage potential at prototypic scale

• Perform experiments to address the concern that cryogenic insulation

(installed to preserve the cold conditions of LNG cargo) could degrade in the event of a large-scale LNG spill and fire.

• The degradation of the insulation systems could cause initially undamaged

tanks to become damaged, resulting in additional spills and larger and/or longer fires.

Methodology

Cargo Tank Insulation Systems

Membrane-type tanker PUF/PRF composite panel Polystyrene panel Moss-type tanker Polyurethane panel

Test Apparatus

Lamp Assembly (1 MW potential)

Test Assembly Enclosure (stainless steel members, Pyrotherm wall panels)

Test Enclosure

Test Assembly X-section (1 m x 1 m x ~2 m deep)

• 4 lamp holder assembly (1.2 m x 1.2 m)
• insulation mask (heated area 864 mm x

864 mm)

• steel plate representing Moss weather

cover or membrane outer hull (~16 mm, 5/8”)

• gap representing hold space or ballast

space (~1300 mm)

• insulation system (membrane or Moss,

~300-500 mm thick)

• Al tank (38 mm wall thickness, LN2 filled)
• top exhaust, N2 purge

Radiant heat lamp array Inner hull steel plate Insulation system LN2 tank Air gap Outer hull steel plate

No96 Composite Polystyrene MKIII

GTT (Membrane) Panel Fabrication

No96 perlite-filled plywood boxes MK III polyurethane foam panels

Moss Panel Fabrication

• 500
• 400
• 300
• 200
• 100

100 200 300 400 500

• 500
• 400
• 300
• 200
• 100

100 200 300 400 500 height (mm) width (mm)

1 4 6 5 3 2

• 500
• 400
• 300
• 200
• 100

100 200 300 400 500 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 height (mm) depth (mm) PL3 PL2 PL1 SL3 SL2 SL1 PL4 PL6 PL5

PUF/PRF Composite Polystyrene

Test Assembly

Steel plate and mask Water- cooled radiometer Lamps

inside enclosure (removable partial top for inserting test panel)

Typical Test Data and Insulation Thermal Response

Pre-test Cooldown Thermal Exposure and Insulation Response

Summary of Test Results

The following key observations were made: 1. The steel plate acting as a weather cover or outer hull temperature approached 1100⁰C and the insulation layers closest to the steel plate have reached 800⁰C near the end of the test. 2. In general, there was negligible heat flux into the aluminum tank that represents the cargo tank during any of the tests. Only after the thermal front passed through the layer closest to the aluminum tank was there an increase in flux. 3. The heat flux value shown into the LN2 tank were obtained using heat flux gauges attached to the tank and by evaluating the change in the liquid nitrogen boil-off rate in the LN2 tank. 4. Actual heat flux data for each insulation system tested is proprietary.

Potential for Overpressure or Thermal Damage to Adjacent Cargo Tanks - Moss

• In a 40 minute exposure, all of the styrene will vaporize.
• Radiative heat flux from the ~1100⁰C weather cover is greatly reduced

due to smoke production.

• Participating media heat transfer and free convection heat transfer

analysis for a Moss LNG cargo tank, support a heat flux estimate of up to 10 kW/m2 onto the cargo tank.

• There is a minimal likelihood of a Moss LNG cargo tank being damaged

from a fire due to vapor over pressurization.

Potential for Overpressure or Thermal Damage to Adjacent Cargo Tanks - Membrane

• The outer hull steel and the inner hull steel will see temperatures close to

1000ºC and 600ºC, respectively, significantly reducing structural integrity.

• However, the heat flux into the LN2 tank was very low, and the thermal

wave for both membrane insulation systems had not reached the LN2 tank at the end of the nominal 40 minute thermal tests.

• Pressure may rise in the inter-hull space (without venting, 1.5 barg in 30

min), slow enough that relief valves can protect this space.

• Expansion of gases from thermally-decomposing materials will increase

pressure internal to the insulation systems; this pressure increase may exceed the relief and venting capacity of this volume.

• It is likely that membrane tanks will distort during a large fire, based on

softening of the inner hull and pressure buildup from vaporization of the insulation system producing pressure sufficient to deform the membrane.

• However, we believe that distortion of the inner and outer hulls will be the

larger cascading damage driver, not the thermal and pressure integrity performance of the cargo tank insulation system.

Summary

All insulation systems showed some degradation, and some showed better performance than others. The smallest heat flux at the aluminum tank containing liquid nitrogen (a surrogate for the LNG cargo tank) was ~0 kW/m2. The largest heat flux at the aluminum tank was ~5 kW/m2. The following key observations were made: 1. Heat flux of ~5 kW/m2 will not cause high temperature/direct damage of the cargo tank. 2. LNG fire-induced boil off does not exceed the venting capacity of the cargo tank relief valves. 3. LNG-induced fire heat flux creates high temperatures (~1000-1100⁰C) on the ship steel. 4. All insulation systems will see degradation and reduction in mechanical strength . 5. LNG-induced high insulation temperature will lead to material pyrolysis, degradation, and flue gas formation.

Questions?