Appendix 4
APPENDIX 4
CONSIDERATIONS FOR FUEL TANK THERMAL MODELS
1. The type of thermal model that is needed to show compliance will depend upon the method of compliance selected for the different requirements. For fuel tanks shown to have equivalent flammability to that of an unheated aluminum wing fuel tank, a fuel tank thermal model may not be needed if the qualitative flammability assessment method defined in paragraph 11 of this AC is used. For other tanks where specific flammability levels must be shown, thermal modeling will be needed if fuel temperature affects fuel tank flammability. This appendix provides guidance for validating the thermal modeling of a fuel tank. It is based on using in the Monte Carlo Model defined by the User’s Manual (defined in paragraph 4e(11) of this AC). As stated in paragraph 4.1 of the User’s Manual, if flight test data or a detailed analysis of the fuel tank’s thermal behavior shows that this method cannot yield an accurate representation of the actual fuel tank’s fuel temperature profile, then modification of the thermal model code in the Monte Carlo model is necessary.
2. Validating fuel tank thermal coefficients, “Tau,” for use in the Monte Carlo Model.
a. When a fuel tank is heated or cooled by a change in air temperature, the response of the fuel temperature is to increase or decrease, respectively, following an exponential decay law. On the ground, air temperature is considered to be ambient temperature at the airplane location, and in flight it will be the Total Air Temperature (TAT) experienced by the airplane. This exponential trend is driven by the temperature difference between the fuel and TAT, and the response of the mass of the fuel and tank.
It can be represented by the system exponential decay time constant Tau, and an equilibrium temperature that the fuel temperature will eventually reach. The equilibrium temperature for a totally unheated tank will be very close to the air temperature, and can be expressed as a temperature difference from ambient temperature on the ground and TAT in flight. For this method, the temperature difference will be called Delta T.
b. By taking data from a flight test, the values for Tau and Delta T can be approximated as a function of fuel load and air temperature. This method requires recording fuel tank temperatures and TAT at regular intervals during critical operational ground and flight test conditions so that the fuel tank thermal characteristics can be established.
3. Compliance Testing. Testing should include ground and flight conditions with variable fuel quantities, and any heat transfer from airplane generated sources to the fuel tank at the critical conditions. The Monte Carlo Model includes the capability to use 6 different Tau values, and sufficient testing to confirm the validity of each of the values used in the model should be done. In addition, the thermal characteristics of the critical A4-1
Appendix 4 portion of the fuel tank relative to formation of flammable vapors should be used in the Monte Carlo Model and validated by the testing. Baffling incorporated into most fuel tanks results in segmented volumes that may significantly affect heat transfer and, therefore, the flammability within the tank volume. If barriers or walls result in separate volumes within the tank and prevent mixing of the fuel and/or vapors in the tank, then each of these volumes should be evaluated independently to determine the worst case exposure for that tank. The validation may include a qualitative assessment of the heat sources (warm fuel transfer, hydraulic heat exchanges, fuel return lines, ECS packs, etc.)
and thermal characteristics of different portions of the fuel tank to substantiate critical location for the test evaluation of the Taus.
a. Temperature Measurement. The location of test instrumentation should consider tank configuration and operational factors to determine which locations in the fuel tank require evaluation. The fuel temperature should be measured at critical locations in the tank for each of the critical fuel loading conditions. If an FRM that limits ullage oxygen concentration is included in the design, test instrumentation should be located in critical locations of the ullage to support analysis of the ullage density changes and it’s effect on ullage oxygen concentration. The locations of fuel and ullage test temperature sensors should take into consideration any structural barriers or baffles that may divide the tank into regions that may have different thermal characteristics.
b. Test Conditions. The applicant should conduct sufficient ground and/or flight testing that simulates the actual operation of the airplane type so that a validated fuel tank thermal model can be developed that will accurately model the fuel tank. Fuel tank temperatures should be measured so that the change in bulk average fuel temperature in each affected tank on the ground and in-flight versus time can be determined. The minimum number of test conditions needed to validate the thermal modeling of the fuel tank will depend upon the number of variables that may affect the fuel tank temperature.
For example, a minimum of two test flights would be required, one representative of a short mission for the airplane and one representative of a long mission for a typical CWT design so that variability in mission length, gate time and fuel quantities could be addressed. The short mission should include at least 30 minutes of ground operation prior to flight, and the long mission at least 90 minutes of ground operation prior to flight. For the entire mission (i.e., from the start-up of airplane systems to completion of the flight), temperatures should be recorded in the test tank(s) at locations to represent the bulk fuel temperature in each separate section of the test tank(s) unless less a qualitative assessment shows that fewer locations can be justified. For this discussion, TAT is used to represent ambient temperature on the ground and TAT in flight. The fuel temperature and TAT should be recorded at least once per minute. The test program should be designed to address the fuel tank thermal response as a function of day type including testing on warm days to understand the fuel tank thermal response when the tank is most likely to be flammable.
4. Example of a Flight Test. The sample flight test shown below is an illustration of how to demonstrate that the bulk average fuel temperature modeling used in the Monte Carlo Model, based upon use of Time Constant Tau, is acceptable for the particular tank A4-2
Appendix 4 under evaluation. The fight test shown is for a hypothetical aluminum wing tank on a long-range airplane. The data includes operational time before and after the flight as required by the method. The fuel temperature is recorded every minute and the bulk average fuel tank temperature is then determined from the test data. (Data at 10 minute intervals is shown for the example, to save space.) The Tau value is estimated for a full tank and decreased as the fuel is burned off in flight. Using the value of Tau, the tank bulk average fuel temperatures can be estimated using the exponential decay equation, (the term “Fuel Temp” in the equations is an abbreviation for “bulk average fuel temperature”):
(Fuel Temp change in time t)/(Fuel Temp-TAT)= 1-e (-t/Tau) This can also be written as:
Fuel Temp change in time t = (Fuel Temp-TAT) x (1-e (-t/Tau)) The resultant estimated fuel temperature is then compared to the actual and the Tau values corrected until a satisfactory match is obtained. This can be done manually or using a computer solver approach. If there is no heating of the tank, the Tau values can be obtained directly. If there is significant heating of the tank, the equilibrium temperature can be found by increasing the value for TAT to (TAT + Delta T), where Delta T is the tank equilibrium temperature offset from TAT. In this case, both the values for Tau and Delta T must be found for the tank. The Tau used in the Monte Carlo Model should be shown to predict fuel tank temperatures with an error of less than 3 degrees for a period of time not to exceed 5 minutes. If the accuracy does not meet this level, the applicant may choose to use a Tau that is adjusted so that the predicted temperature would be higher and produce a conservative flammability level or propose an alternative method of modeling fuel tank temperatures.
A4-3
Appendix 4
TABLE 4-1 Example of Wing Fuel Tank Flight Test Results (Part 1 of 2)
Shown at 10 minute increments for this example to save space.
Should be recorded at 1 minute intervals Delta T
Estimated Tau Tank Temperature Error Time OAT Mach Numbe r TAT Altitude Fuel Load Fuel Temp Fuel Temp TAT Actual vs.
Predicted
Min
Deg.
F.
Deg.
F,
Ft.
%
Deg. F.
Deg. F.
Deg F.
Predicte d Temp based on Tau value
0 80 0.00 80.0 0 100 80.0 0.0 149.8 80.0 0.0 10 80 0.00 80.0 0 100 80.0 0.0 149.8 80.0 0.0 20 80 0.00 80.0 0 100 80.0 0.0 149.8 80.0 0.0 30 80 0.00 80.0 0 100 80.0 0.0 149.8 80.0 0.0 40 80 0.00 80.0 0 100 80.0 0.0 149.8 80.0 0.0 50 80 0.00 80.0 0 100 80.0 0.0 149.8 80.0 0.0 60 80 0.00 85.9 0 100 80.0 -5.8 149.8 80.0 0.0 70 42.5 0.43 58.2 8750 100 79.5 21.2 149.8 79.4 0.0 80 5 0.57 32.7 17500 100 77.4 44.6 149.8 77.1 -0.2 90 -32.5 0.71 8.5 26250 100 73.9 65.4 149.8 73.4 -0.5 100 -70 0.86 -18.7 35000 100 69.2 87.8 149.8 68.3 -0.8 110 -70 0.86 -18.7 35000 100 63.8 82.5 149.8 62.7 -1.2 120 -70 0.86 -18.7 35000 100 58.8 77.4 149.8 57.3 -1.4 130 -70 0.86 -18.7 35000 100 53.9 72.6 149.8 52.4 -1.6 140 -70 0.86 -18.7 35000 100 49.4 68.1 149.8 47.7 -1.7 150 -70 0.86 -18.7 35000 100 45.1 63.8 149.8 43.3 -1.8 160 -70 0.86 -18.7 35000 100 41.0 59.7 149.8 39.2 -1.8 170 -70 0.86 -18.7 35000 100 37.2 55.9 149.8 35.4 -1.8 180 -70 0.86 -18.7 35000 100 33.6 52.3 149.8 31.8 -1.8 190 -70 0.86 -18.7 35000 100 30.2 48.9 149.8 28.5 -1.7 200 -70 0.86 -18.7 35000 100 26.9 45.6 149.8 25.3 -1.6 210 -70 0.86 -18.7 35000 98 23.7 42.4 146.9 22.4 -1.3 220 -70 0.86 -18.7 35000 94 20.7 39.3 142.2 19.6 -1.1 230 -70 0.86 -18.7 35000 90 17.7 36.4 137.5 16.9 -0.8
A4-4
Appendix 4
TABLE 4-1 (Continued) Example of Wing Fuel Tank Flight Test Results (Part 2 of 2) Shown at 10 minute increments for this example to save space Should be recorded at 1 minute intervals Delta T Estimated Tau Predicte d Temp based on Tau value Tank Temperature Error Time OAT Mach Numbe r TAT Altitude Fuel Load Fuel Temp Fuel Temp- TAT Actual vs.
Predicted Min Deg.
F.
Deg F.
Ft. % Deg F. Deg F. Deg F.
240 -70 0.86 -18.7 35000 87 14.9 33.6 132.8 14.3 -0.6 250 -70 0.86 -18.7 35000 83 12.2 30.9 128.1 11.9 -0.4 260 -70 0.86 -18.7 35000 79 9.7 28.3 123.4 9.5 -0.1 270 -70 0.86 -18.7 35000 76 7.2 25.9 118.7 7.3 0.1 280 -70 0.86 -18.7 35000 72 4.9 23.6 114.0 5.2 0.3 290 -70 0.86 -18.7 35000 68 2.7 21.3 109.3 3.2 0.5 300 -70 0.86 -18.7 35000 65 0.6 19.3 104.6 1.3 0.7 310 -70 0.86 -18.7 35000 61 -1.4 17.3 99.9 -0.4 1.0 320 -70 0.86 -18.7 35000 57 -3.3 15.4 95.2 -2.1 1.2 330 -70 0.86 -18.7 35000 54 -5.0 13.7 90.5 -3.7 1.3 340 -70 0.86 -18.7 35000 50 -6.6 12.0 85.8 -5.1 1.5 350 -70 0.86 -18.7 35000 46 -8.2 10.5 81.1 -6.4 1.7 360 -70 0.86 -18.7 35000 43 -9.6 9.1 76.4 -7.7 1.9 370 -70 0.86 -18.7 35000 39 -10.9 7.8 71.7 -8.8 2.1 380 -70 0.86 -18.7 35000 35 -12.1 6.6 67.0 -9.8 2.2 390 -70 0.86 -18.7 35000 31 -13.2 5.5 62.3 -10.7 2.4 400 -70 0.86 -18.7 35000 28 -14.1 4.5 57.6 -11.6 2.6 410 -70 0.86 -18.7 35000 24 -15.0 3.7 52.9 -12.3 2.7 420 -70 0.86 -18.7 35000 20 -15.8 2.9 48.2 -12.9 2.9 430 -70 0.86 -9.8 33906 17 -16.2 -6.5 43.5 -13.3 3.0 440 -20 0.67 22.1 22969 17 -10.4 -32.5 38.8 -8.1 2.3 450 30 0.48 55.6 12031 16 2.4 -53.2 34.1 4.0 1.7 460 80 0.28 80.0 1094 15 20.8 -59.2 29.9 23.5 2.7 470 80 0.00 80.0 0 15 36.2 -43.8 29.9 40.0 3.8 480 80 0.00 80.0 0 15 47.6 -32.4 29.9 52.3 4.7 490 80 0.00 80.0 0 15 56.0 -24.0 29.9 61.4 5.3
A4-5
Appendix 4
Comparison of Test and Calculated Fuel Temperatures -40 .0 -20 .0
0.0
20.0
40.0
60.0
80.0
100.0 0 200 400 600 Time -Mi nutes Fuel Tempe rature -De g F Tank Temp Deg F Predicted Temp based on Tau value Tank Temperature Error
FIGURE 4-1
Comparison of Test and Calculated Fuel Temperatures
5. Compliance Reporting. In order to show the fuel tank temperature modeling is acceptable using this method, the compliance report should include plots of the flight test data, as well as the predicted fuel tank temperatures and the associated error between the predicted temperature and those measured during the flight test.
A4-6
Appendix 5 APPENDIX 5
MEASUREMENT OF OXYGEN CONCENTRATION IN AIRPLANE FUEL TANKS
1. Introduction.
a. Background on Oxygen.
(1) Oxygen is the element most associated with life on our planet. It has an atomic number of 16 and is classified as an oxidizer, which means it reacts readily with other elements and is key to many reactions taking place. Oxygen is essential to every living thing. Air is a mixture made of 99.9 percent oxygen and nitrogen existing primarily on the molecular level in a single bond with itself, thus we use the chemistry term O2 when describing oxygen and N2 when describing nitrogen.
(2) Measuring oxygen in a given environment is accomplished by exposing a gas sample to a sensor designed to give feedback for a changing amount of oxygen in the sample. The various sensing technologies rely on one of several different natural properties of oxygen to change the voltage in a circuit, or more directly, the number of counts on an A/D converter. Virtually all oxygen sensors do not sense the volume percentage of oxygen in a sample, but rather, the partial pressure of oxygen in the gas sample in question. The analyzer is then calibrated for volume percent oxygen given a fixed sample pressure, flow, and temperature.
b. Types of Oxygen Sensors.
(1) Most sensing technologies work on some type of chemical reaction. A galvanic cell is a sensor containing a small amount of a liquid solution that reacts with oxygen, retained within a permeable membrane. The gas sample is exposed to the sensor membrane and oxygen from the sample enters or leaves the sensor solution to create equilibrium. The solution in these sensors is separated by an anode and cathode. A fixed current source changes the voltage across the anode and cathode given a changing amount of oxygen in the sensor solution (see Reference 1 of paragraph 4 of this appendix). This voltage can be converted to a fixed output with a voltage divider and some linearizing circuitry. Some sensors simply have a piece of chemical compound
(i.e., zirconium oxide) reacting between two electrodes. When a gas sample is exposed to the quantity of compound, the oxygen in the sample reacts with the compound and changes the voltage across the electrodes (see Reference 2 of paragraph 4 of this appendix).
(2) Virtually all sensors have a finite life, due to the fact that the reactants and solutions degrade and dilute over time. It is possible to measure oxygen with a truly nonconsumable sensor. One commercially available oxygen analyzer uses the principals of paramagnetism of oxygen to measure the volume concentration of oxygen. Because A5-1
Appendix 5 oxygen becomes magnetic when exposed to a magnetic field and the other constituents of air (nitrogen, carbon dioxide, and water vapor) do not, it is possible to make a detector that is sensitive to the amount of paramagnetic oxygen in an otherwise diamagnetic mixture of gases. In the case of the Rosemount 755R paramagnetic oxygen analyzer, a dumbbell-shaped, nitrogen-filled, hollow glass sensing body, wrapped with a single platinum wire, is suspended on a platinum/nickel alloy ribbon. When the sensing body is surrounded by a gas sample, “magnetic buoyancy” occurs creating a torque on the dumbbell. The current required to counteract the torque and restore the dumbbell to its null position is directly proportional to the amount of paramagnetic gas (oxygen) in the sample (see Reference 3 of paragraph 4 of this appendix).
(3) Some sensors do not require a dedicated gas sample to be removed from the environment to be measured, and thus can be used in situ, or literally in place. Some of these sensors work on the fluorescence properties of light, which change with partial pressure of oxygen in a given environment. These sensors detect changing fluorescence properties (intensity, time decay, etc.) of the light reflected from a coherent source using a spectrometer (see Reference 4 of paragraph 4 of this appendix). This measurement method is best applied in a controlled environment as the process is very sensitive to changes in temperature and pressure and all in situ methods impose unique calibration problems. In general, sensors that are reacting with oxygen make poor in situ sensors due to their propensity to consume oxygen in the general area of the measurement causing a type of sensor drift during measurements of quiescent environments.
2. Choosing a Measurement Method. To choose a measurement method several considerations must be made. The quantity of data as well as the number of gas sample locations required will determine the method of acquiring oxygen concentration for a flight test. The primary methods of measuring the oxygen concentration in a fuel tank ullage applied to date are continuous sampling, in situ measurements, and discrete sample measurements. Each primary method of measurement has an analyzer/sensor technology that is best applied given the measurement requirements. Each sensor and analyzer combination will have specific gas sample requirements. In the case of in situ measurements, sensitive sensors and fiber optic cables may need special protection.
Discrete (individual) measurements made during any test always require the carefully choreographed and documented actions of test personnel during the test. This can lead to problems with post-test analysis of the data if information becomes lost or mishandled.
Virtually all instrumentation has an effect on the environment being measured.
Regardless of the method used to measure oxygen concentration, care should be taken to minimize the sampling system impact on the ullage oxygen concentration. Ullage samples drawn from the fuel tank for measurement should be returned if possible to the tank in locations that minimize the effect of the samples taken for measurement. The analysis presented for compliance should include the effect of the instrumentation on the results.
a. Continuous Sampling. The ideal measurement of oxygen concentration in a fuel tank environment is instantaneous and continuous. This, however, is not very practical as A5-2
Appendix 5 some finite time is required to sample ullage gas and excite a particular sensor. By keeping gas sample lines short and small (diameter), while maintaining a relatively high sample flow rate, a gas sample system with a relatively short response time can be developed. Traditional sensors generally have a response time between 500 ms and 5 seconds, depending upon the technology applied.
(1) Flow Through Sensor System.
(a) One such method of continuous sampling was developed and applied by the FAA using flow through galvanic cells and a pressure regulated sample train. The sample train uses a powerful pump to draw a gas sample from the ullage sample location in question, through a forward pressure regulator, which regulates the pressure on the input of the sample train to a vacuum pressure lower than the lowest expected pressure in the sample environment (2.73 psia or 141 Torr @ 40,000 feet). The gas sample is then pumped up to a pressure to that just above sea level (14.7 psia or 760 Torr), allowing the gas sample to flow through a flow meter, the galvanic sensor, and then a check valve.
Lastly, the gas sample passes through a pressure controller that regulates the back pressure to a fixed value that can be maintained throughout the flight cycle. This pressure is critical to maintaining the calibration of the analyzer with the galvanic cell sensor. The gas sample can then be returned to the ullage to allow for a minimal effect on the sample environment.
(b) Calibration of this measurement methodology is accomplished easily by having a selector valve in the sample train before the input pressure regulator. The sample train is completely pressure regulated, providing calibration gas to the system input, allowing time for the system to stabilize, and adjusting the analyzer gain allows for calibration. It is best to check for adherence of your oxygen measurement system to a wide variety of calibration gases at least once before and after a test program. This is accomplished by calibrating in the traditional manner, and then checking several other calibration gases with different oxygen concentrations within the measurement range of the testing.
(c) When applying this measurement method to an airplane flight test, care must be taken to preclude liquid fuel or vapors from entering the cabin from the fuel tank gas sample lines. Traditionally, all sample lines are double walled and the shroud (outer wall) space is vented overboard to the exterior of the cabin. Leak checks of both the line and shroud prior to testing are essential. The measurement system sample volume should be isolated from the fuel tank with flash arrestors in the event of a failure that creates an ignition within the measurement system itself. Most oxygen sensors operate on very low power; however, complete isolation of the sensor, and other sources of energy within the sample system, from the fuel tank is essential in the event of a failure that could create a reaction within any portion of the flammable sampling system.
(d) The sampling of liquid fuel should be precluded by use of float valves on the ends of each sample tube in the fuel tank. Also, in the event of a failure that allowed liquid fuel into the sample lines, fuel should be prevented from entering the A5-3
Appendix 5 measurement system with either an automatic float shut-off switch or a warning system that allows for the operator to shut-down the sample system. The measurement system itself should also be shrouded and ventilated to allow for protection of the cabin environment in the case of a sample train leak. During ground operation, when natural suction is not available from the cabin exterior, the measurement system housing should also be continuously purged.
(e) A complete description of the galvanic cell continuous sample system applied by the FAA for the purposes of studying in-flight fuel tank inerting is given in Reference 5 from section 4 of this appendix.
(2) Flow-by Sensor System.
(a) Some sensors react significantly with the sample, and therefore may not be appropriate for returning the sample to the tank. A method of applying this type of sensor to a fuel tank environment is to create a volume flow sample train similar to the one described in the previous section. A small amount of this sample is then removed from the loop, in the positive pressure portion of the flow train, allowed to flow through the reacting sensor in question, and then deposited overboard. Minimizing the flow to the sensor helps minimize the effect of the consumed sample and creates a safer environment should the reaction of the ullage sample with the sensor become hazardous.
It may be necessary to remove flammable fuel tank vapors from the sample gas to safely apply some sensors for measuring ullage gas.
(b) Calibration of this measurement methodology is accomplished easily by having a selector valve in front of the sensor housing, outside of the primary sample loop. Care must be taken to ensure that the calibration gases flow through the sensor at same pressure and flow rate as the flow by sample. The flow-by sensor should stabilize quickly, allowing for adjustment of the analyzer gain. Again, it is best to check for adherence of your oxygen measurement system to a wide variety of calibrations gases at least once before and after a test program.
(c) Ensuring safety of a flow-by sensor installation is done much in the same manner as described in the previous section. Special care may need to be taken to ensure the reaction between the sensor and the ullage gas sample remains completely benign to the measurement system and the aircraft in general. It is for this reason that high energy sensors have not been applied to in-flight measurement applications.
Whatever installation is chosen, careful analysis and application is required, within the confines of the regulations, to minimize risk in accordance with the individual program plan developed.
A5-4
Appendix 5
FIGURE 5-1
Block Diagrams of Flow Through and Flow Bypass Sample Systems
b. In Situ Measurements.
(1) Application of in situ sensors for measuring fuel tank ullage oxygen concentration has been applied and observed by the FAA on a limited basis. However, recent advances in application of these sensor technologies make them a viable option for future in-flight oxygen concentration measurement in a fuel tank ullage. Optical methods such as fluorescence quenching are attractive to apply to measurements in fuel tanks due to the intrinsically safe nature of the energy required for the sensor, and in situ methods in general alleviate the need to remove flammable ullage gas from the fuel tank for sampling.
(2) Each in situ sensor will have its own unique installation issues due to the sensor size and shape, as well as any environmental limitations such as temperature.
Routing of any fiber optic cables will need to be within the requirements for the cable.
(3) The main difficulty with applying in situ sensors is the need to calibrate the sensor in place. This is not practical on a daily basis in an airplane fuel tank. This is complicated by the fact that optical methods generally have a very nonlinear response and need to be calibrated for a changing ullage temperature. In previous applications of fluorescence quenching sensors, the calibration was done in a controlled environment using multiple gases at multiple temperatures. The system was then installed in the tank and each channel was then “normalized” to a baseline intensity using air at some middle calibration temperature. This proved to be problematic given the variable of the baseline intensity of the different channels between the lab and airplane.
(4) To alleviate this, a sleeve assembly could be fabricated for the sensor that allows the sensor to be immersed in calibration gas through the sleeve, deposited from the outside of the fuel tank, to allow for daily checking of the calibration for several A5-5
Appendix 5 gases. At the very least, the sensors would need to be validated at a single point before every flight. This has been accomplished in the past by having the inert gas generation system create a fixed oxygen concentration gas for an extended period of time. This will eventually stabilize the tank at a single oxygen concentration and temperature to allow for validation of each sensor in the tank. It is easy to see that this could quickly become expensive and time consuming if the calibration had to be checked multiple times and/or with multiple gases on a daily basis.
(5) Safety requirements for equipment located within a fuel tank such as in situ measurement systems are already established in the existing § 25.981 and related sections (See AC 25.981-1B (or latest revision). These safety requirements establish limitations on power required for instruments installed and places limitations on materials for use in fuel tanks. Part 25 addresses analysis of failure modes in general (§ 25.1309) and sets general installation requirements (subpart F). The power required for any sensor installed in a fuel tank needs to be calculated to be less than the minimum ignition energy prescribed by the FAA for fuel tank installations. Any cables that need to be installed should be routed in accordance with standard industry practices within the requirements of all airworthiness regulations. Each individual installation needs to be analyzed and evaluated for risk within the confines of the regulations and agreed upon by all project stakeholders.
c. Discrete Sampling . Although not desirable, it is possible to evaluate the capability of a particular inerting system to allow for protection of a given fuel tank without acquiring continuous oxygen concentration data throughout a complete flight cycle. This may be because of advances in ullage oxygen concentration modeling or because of the existence of a large quantity of previously acquired data for a similar inerting system and application. Regardless, collection of discrete samples at several critical times for a few different ullage locations can provide a complete picture of capability of an inerting system to maintain an inert ullage in a specified fuel tank during a flight cycle for some applications. Discrete methods may provide a cost advantage over continuous sampling measurements and can be considerably less complex if large quantities of gas samples are not required. The primary methods of acquiring discrete samples are through vacuum bottle sampling or through single point calibration and sampling.
(1) Vacuum Bottle Sampling.
(a) One method of acquiring discrete ullage gas samples at various times is by the use of vacuum sample bottles, which are previously cleaned, evacuated, and hermetically sealed. Vacuum bottles are generally used by plumbing several bottles in series. Each bottle has an individual hand operated valve to let in a gas sample, to the desired sample line, which is located within the ullage of the fuel tank in question. Each vacuum sample bottle in the series will provide a single, discrete gas sample in time at that particular ullage sample location. A vacuum sample bottle generally has a second vacuum bottle adjacent to it which is deployed (valve opened) before deploying the A5-6