Ground Based Fuel Tank Inerting

1
Modeling Inert Gas Distribution in
Commercial Transport Aircraft Fuel Tanks
William M CavageProject Manager - Fuel Tank
Inerting FAA AAR-440, Fire Safety RD Branch
22nd AIAA Aerodynamic Measurement Technology and
Ground Testing Conference June 24th-26th, 2002
Adams Mark Hotel - St. Louis, MS
2
Outline

• Background
• Equipment Procedures
• B-747SP Ground Test Article
• 24 Scale Tank
• Data Analysis
• Modeling Methods
• Results
• Summary

3
Background

• Ongoing FAA Rulemaking Seeks to Improve on the
  Existing and Future Fuel Tank Safety
• Consistent Accident Trends are a Concern
• Focus of Concern is on Heated Center Wing Tanks
  (CWTs)
• Fuel Tank Inerting is a Well Established Method
  of Reducing/Eliminating Ullage Vapor Flammability
• Has Been Meet with Resistance by Industry Leaders
• FAA Would Like to Develop Cost Effective Methods
  of Modeling Inert Gas Distribution in Commercial
  Transport Fuel Tanks

4
Equipment

• Boeing 747SP Full-Scale Inerting Test Article
• Decommissioned from Airline Service and Purchased
  by the FAA for Ground Testing Only
• All Major Systems Fully Operational
• Has Independent Power for Test Equipment and
  Instrumentation
• Full Complement of Ground Service Equipment
• Aircraft Modified to Study Inerting
• Inert Gas Deposit System Installed on Aircraft
• Inerts CWT from Ground Source of Nitrogen
  Enriched Air (NEA)
• Instrumentation
• Gas Sample Tubing at 8 Locations for Oxygen/THC
  Analysis
• 32 Thermocouples in Tank (Ullage, Fuel, Walls,
  Floor, and Ceiling)

5
Boeing 747SP Aircraft
6
Boeing 747-100/SP Center Wing Tank
7
Boeing 747SP CWT Top Diagram
8
Equipment

• Scale Tank Test Article
• 24 Scale Model of Boeing 747 SP CWT was Built
  from 3/4 Inch Plywood By Scaling Drawings from
  Shepherd Report
• Spars and Spanwise Beams Simulated with ¼ Inch
  Plywood Installed in Slats with Scaled
  Penetration Holes
• Vent System Simulated with PVC Tubing Plumbed to
  an Aluminum Vent Channel Plumbed Similar to
  Aircraft
• Instrumentation
• Oxygen Sensor in Each Bay and in One Vent Channel
  Plumbed in Unique Sample Drafting Method
  Returned to Each Bay
• Thermocouple in Each Bay
• Variable NEA Manifold Allowed for NEA to be
  Deposited in Any and All Bays of the Tank at
  Different Flow Rates

9
Scale Plywood CWT Model
10
Procedures

• 24 Scale Tank Testing
• Series of Tests Done to Examine Different Deposit
  Schemes
• Deposited Different Amounts of NEA in Different
  Bays to Determine the most Efficient Method of
  Deposit in a Half Blocked Venting Configuration
• All Work Presented is for 95 NEA and 128 CFH
  Total Flow Rate
• Focus of Testing was to Find Best Method of
  Depositing NEA
• Boeing 747SP Full-Scale Inerting Testing
• Series of Tests Done to Examine the Efficiency of
  Inerting
• Single Deposit (Optimal from Scale Testing)and
  Venting Case
• Tested for Different Day and Operational
  Conditions
• All Work Presented is for 95 NEA and 140 CFM
  Total Flow Rate with ACMs Running (Vertical
  Mixing Stimulated)
• Focus of Testing was on Operational Effects and
  Predictability

11
Data Reduction Analysis

• Volumetric Tank Exchange is the Ratio of the
  Volume of Deposited Gas to the Volume of the Tank
• This Gives a Dimensionless Quantity of Inert Gas
  Given the Volume of the Tank
• Average Tank Oxygen Concentration
• This Gives a Representation of the Tank Oxygen
  Constituency Given Varying Oxygen Concentrations
  in Different Bays

12
Inert Gas Distribution Engineering Model

• Model Calculates Inert Gas Distribution in 6 Bay
  Tank, in terms of Oxygen Concentration Evolution,
  Given NEA Purity and Bay Deposit Flow Rates
• Based on Original Single Bay Inerting Model, by
  FAA CSTA for Fuel Systems, which Tracks Oxygen In
  and Out of Each Bay Assuming Perfect Mixing
  During the Time Step
• Assumes an Outward Flow Pattern and Splits Flow
  into a Bay to Adjacent Bays Using Out Flow Area
  Relationships
• Basic Formula for Volume of Oxygen in a Bay

13
Assumed Engineering Model Flow Pattern
Flow Out
Bay 1
Bay 2
Bay 4
Flow In
Bay 3
Flow Out
Bay 5
Bay 6
14
CFD Model

• A CFD Model was Developed with the Analysis
  Package FLUENT
• Used the Fluent CFD Solver Which Uses a Finite
  Volume Method Where the General Conservation
  (transport) Equation (Mass, Momentum, Energy,
  etc.) is Solved for Each Finite Volume
• Has Ability to Track Fluid Species (O2
  Concentration) at Given Locations
• Model was Solved Using a Laminar Flow Throughout
  (Oxygen Evolution is Based Entirely on Flow
  Diffusion)
• For Administrative Reasons, Model was Developed
  of the Scale Tank and not Full-Scale Test Article
• The Model Developed had Approximately 700K Cells
  and Ran on Several Platforms Over a Weekend.

15
Results - Full Scale Comparison

• Scale Tank Test Data Compares Well with
  Full-Scale Test Article Data
• Bay 4 Does Not Compare Well for Any Modeling
  Method
• Engineering Model Compares Fair
• Trend Data Very Good but Some Bays have Large
  Discrepancy in Some Bay Oxygen Concentration
  Values when Compared with Full-Scale Data
• CFD Model Comparisons Initially Poor
• Trend Data Marginal with Large Discrepancies in
  Some Bay Oxygen Concentration Values when
  Compared with Full-Scale Data
• Subsequent Data Has Much Better Agreement

16
Scale Tank Data Comparison
17
Engineering Model Data Comparison
18
CFD Model Data Comparison
19
Modeling Methods Compared
20
Results - Mock Trade Study

• Comparing Scale Tank Test Article Data with
  Engineering Model Data for Different Deposit
  Scenarios Has Mixed Results
• Bay to Bay Oxygen Concentration Comparisons Vary
  for Different Deposit Scenarios
• The Average Oxygen Concentration Trend Data for
  the Different Deposit Scenarios is Consistently
  Biased High for the Engineering Model Except for
  the Single Deposit Case
• This Results in a Discrepancy Between Which
  Deposit Method is Optimal (Most Efficient) for
  Each Modeling Method

21
Engineering Model Compared with Scale Tank
22
Engineering Model Compared with Scale Tank
23
Full-Scale Data Compared with Modeling Methods
24
Summary

• Scale Tank Testing Produced Good Results when
  Compared with the Good Mixing Full-Scale
  Testing
• Cost Effective Modeling Method
• Simple Engineering Modeling Methods Can Produce
  Fair Results in a Very Cost Effective Way
• Additional Work Needed to Improve Model
• Additional Research Required to Resolve
  Discrepancies between Engineering Model and Scale
  Tank for for Multiple Deposits
• CFD Data Labor/Resource Intensive and Eventually
  Resulted in Good Comparison to Full-Scale Data