Ground Based Fuel Tank Inerting

1
Inerting of a Scale 747SP Center Wing Fuel Tank
During a Typical Commercial Flight Profile
William CavageAAR-440 Fire Safety
ResearchFederal Aviation Administration
October 30-31, 2002International Aircraft
Systems FireProtection Working GroupAtlantic
City, NJ
2
Outline

• Background
• Scale Tank Model
• Instrumentation
• Testing Scope
• Inerting Efficiency
• Test Data
• Summary

3
Background

• FAA is Seeking to Improve Upon Existing Fuel Tank
  Safety in Fleet in the Wake of TWA800 Air
  Disaster
• Inerting of Fuel Tanks Could Provide Significant
  Fuel Tank Protection.
• Focus of the testing is to validate the ability
  of the FAA simplified fuel tank inerting system
  to inert the CWT of a 747SP
• Use modeling results to validate the modeling
  tools developed during GBI studies
• Study inert gas distribution during the
  commercial mission

4
Description of Model

• Quarter-scale model of Boeing 747SP CWT was built
  from three-quarter inch plywood by scaling
  drawings from Shepherd report
• 24 length scale (1.4 Volume)
• Spars and spanwise beams simulated with
  quarter-inch plywood installed in slats with
  scaled penetration holes
• Vent system simulated with PVC tubing plumbed to
  an aluminum vent channel
• Removable lid to allow for model maintenance and
  modification
• Model in 6x6x7 altitude chamber
• Model inerted with manual NEA mixer

5
Photo of Model
6
747 SP Bay Diagram with Volume Data
Percent Difference 4.47
Reported Volume 1775
7
Instrumentation

• Onboard oxygen analysis system (OBOAS) acquired
  bay oxygen concentration data
• One sample port in each bay
• Sample returned to tank through manifold
• Thermocouple in chamber gave temperature
• Altitude measured by absolute pressure transducer
• NEA Flow metered/measured with mass flow
  controller and oxygen concentration determined
  with flow through galvanic cell type oxygen
  analyzer

8
Onboard Oxygen Analysis System Block Diagram
9
Scale Tank Testing Block Diagram
Altitude Chamber
NEA Generator
Flow Controller
Nitrogen
NEA Mixer
Oxygen Analyzer
Compressed Air
Scale Tank Model
DAS
OBOAS
Computer
Sample Return
T
Pressure Transducer
10
Scope of Testing to Date

• All testing involved single deposit in bay 6
• All testing used generic flight profile with
  different cruise times and different holds at 5K
  feet
• All testing uses same predicted system
  performance in terms of NEA flow and purity
• All tests had same venting configuration
• Right side (left wing tip) vent system blocked
• Plans to block aft port on open vent side for
  some tests

11
OBIGG System Model
12
Inerting Efficiency
Based on perfect mixing solution
with
Develop inerting efficiency equation
Redimensionalize perfect mixing solution
13
Volume Inerting Results

• Inerted tank twice with similar, constant
  inerting conditions at two different altitudes
• Compared nondimensional results
• Results were similar, but not identical
• Average Tank Inerting Close
• Bay inerting trends good, but values can differ
  greatly
• Some data trends unusual
• Inerting efficiency values very different

14
747SP Scale Fuel Tank Inerting Data
15
747SP Scale Fuel Tank Inerting Data
16
Inerting Efficiency Results

• Constant inerting data inerting efficiency trends
  similar but values very different
• Tend toward same value?
• When examining a complete simulation, k values
  vary greatly
• Trends are not consistent
• Makes prediction of inerting with perfect mixing
  solution difficult
• Some calculations respectable with a constant k
• Good k value found by iteration
• Can be very accurate in a narrow data band
• Never examined descent data

17
747SP Scale Fuel Tank Inerting Data
18
747SP Scale Fuel Tank Inerting Data
19
747SP Scale Fuel Tank Inerting Data
20
Descent Inert Gas Distribution Results

• All descent tests had similar results regardless
  of hold times which was better then intuitively
  expected
• Worst bays (bay 1 bay 3) were typically 13-14
  oxygen by volume
• Other bays typically between 10 and 12 oxygen
• Test showed that sample system has small effect
  resulting distribution

21
747SP Scale Fuel Tank Inerting Data
22
747SP Scale Fuel Tank Inerting Data
23
5K Hold Results

• Performed several descent tests
• Started at 39k feet, low tank oxygen
  concentration
• Examined the effect of holding at 5K on the
  resulting tank oxygen concentration as well as
  the worst bay oxygen concentration
• Results indicate short hold helps reduce tank
  oxygen concentration, but exhibits diminishing
  returns on longer holds
• The resulting oxygen concentration of the tank
  was generally around 12 , which is the predicted
  NEA purity at the 5k feet
• Holds tend to improve distribution

24
747SP Scale Fuel Tank Inerting Data
25
747SP Scale Fuel Tank Inerting Data
26
747SP Scale Fuel Tank Inerting Data
27
747SP Scale Fuel Tank Inerting Data
28
Run System After Touchdown Results

• Performed several descent tests
• Started at 39k feet, low tank oxygen
  concentration
• Obtained resulting tank oxygen concentration/distr
  ibution
• Ran system for 15 and 30 minutes at sea level
  in low flow mode (5 NEA)
• Results indicate running system in low flow mode
  decreases the average tank oxygen concentration
  well, but does little for bay 1
• Bay 3 is usually worst bay at touchdown, but
  generally decreases oxygen concentration quickly

29
747SP Scale Fuel Tank Inerting Data
30
747SP Scale Fuel Tank Inerting Data
31
747SP Scale Fuel Tank Inerting Data
32
Summary

• Inerting the tank at altitude was consistent with
  previous GBI research
• Inerting efficiency calculations could be used to
  predict average tank inerting in limited cases
  only with manipulated k values
• Preliminary descent modeling illustrates good
  distribution of inert gas with resulting average
  tank oxygen concentrations of approximately 12
• Short holds at 5K feet will improve the inert gas
  distribution in the tank and help lower the tank
  average oxygen concentration
• Running the system on the ground for 15 to 30
  minutes in low flow mode will improve tank oxygen
  concentration but has little effect on the bay 1
  oxygen concentration