Finite element analysis simulation of a fireproof test for an aircraft propulsion engine mount structure made of titanium.
Aviation regulations requires that engine mounts, and other flight structures located in designated fire zones must be constructed of fireproof material so that they are capable of withstanding the effects of fire. Historically, steel is defined as being inherently fireproof, however, titanium was not. Therefore, a fireproof test was conducted using 6AL-4V titanium structure for the attachment of the propulsion system on a mid-size business jet to satisfy FAA Federal Aviation Requirement 25.865. To determine if the titanium structure would be able to support normal operating loads during the fire event, finite element analysis was performed on the titanium structure simulating the fire test. The fire test simulates a fire on the aircraft from the propulsion system by using a burner with jet fuel exposing the component to a 2000 [Degrees]F (1093[Degrees]C) flame. The 2000 [Degrees]F (1093[Degrees]C) Flame is calibrated based on FAA Advisory Circular AC20-135. The 2000 [Degrees]F (1093[Degrees]C) flame is modeled as a series of convection coefficients across the entire surface of the component. The conductive and convective thermal properties are used for the finite element analysis (FEA) model to simulate the heat transfer effects of the flame. A thermal transient analysis was performed to determine the component temperatures and correlation to the fire test showed excellent agreement. The peak temperatures in the vicinity of the flame on the titanium structure was about 1500 [Degrees]F (816[Degrees]C) but much lower at locations that were shielded by the structure. The transient thermal analysis also showed that after about 10 minutes the temperatures appeared to be at steady state conditions.
CITATION: Leicht, D., "Finite Element Analysis Simulation of a Fireproof Test for an Aircraft Propulsion Engine Mount Structure Made of Titanium," SAE Int. J. Aerosp. 8(1):2015, doi:10.4271/2015-01-2621.
Aviation regulations ([1, 2, 3]) require that engine mounts, and other flight structures located in designated fire zones, must be constructed of fireproof material defined as being equivalent or better than steel. Fireproof is defined in the definitions and abbreviations of the Code of Federal Regulations Section 1.1  as the following : (1) With respect to materials and parts used to confine fire in a designated fire zone, means the capacity to withstand at least as well as steel in dimensions appropriate for the purpose for which they are used, produced when there is a severe fire of extended duration in that zone; and (2) With respect to other materials and parts, means the capacity to withstand the heat associated with fire at least as well as steel in dimensions appropriate for the purpose for which they are used. The latter definition is what was used for the definition of fireproof for the Titanium structure.
The engine mount support system for an aircraft with fuselage mounted engine(s) consists of a top and bottom front mount attached to a yoke and one aft mount as shown in Figure 1. The airframe to engine attachment system includes a forward yoke structure attached to the engine intermediate case, with elastomeric isolators. The forward yoke structure made of 6Al-4V Titanium was analyzed and tested to comply and certify its use against the fireproof requirements.
The material used to manufacture the yoke was made from Titanium 6AL-4V beta annealed. The material composition is defined in AMS 4928 Q with 6% Aluminum, and 4% Vanadium in the annealed condition (2 hours at 1375 [Degrees]F (746[Degrees]C) and air cool) . It was then heat treated to the beta anneal condition which is defined in SAE
AMS-H-81200B as having the part soaked at a temperature which is 50 [Degrees]F [+ or -] 25 [Degrees]F (10[Degrees]C [+ or -] 3.9 [Degrees]C) above the pre-determined beta transus of about 1850 [Degrees]F (1010 [Degrees]C) of the lot for a minimum time of 30 minutes . The part is then air cooled and followed by a mill anneal, which is aging at 1350 [Degrees]F (732 [Degrees]C) for 2 hours .
A development fireproof test was conducted using 6A1-4V Titanium structure for the attachment of the propulsion system on a mid-size business jet to understand if the FAA Federal Aviation Requirement 25.865  could be satisfied. The testing provided data on the thermal distribution during a fire event and the ability of the yoke in the mounting system to carry normal flight loads during exposure to a flame of 2000[Degrees]F [+ or -] 150[Degrees]F (1093[Degrees]C [+ or -] 66[Degrees]C) for a minimum of 15 minutes per Advisory Circular AC20-135 . The basic torch or burner used in the fire test produced a 2000[Degrees]F [+ or -] 150[Degrees]F temperature within 0.25 inch of the specimen. Figure 2 shows a total of sixteen (16) thermocouples were placed on both sides of the yoke and in the air at both the forward and aft sides of the yoke near the clevis as shown in the red circle.
Figure 3 shows the fire test frame capable of placing a load representing a landing limit condition on the titanium yoke structure, while also being exposed to the 2000[Degrees]F (1093[Degrees]C) flame. Linear velocity differential transducers (LVDT) were used at each actuator to measure the deflection of the yoke test specimen at that load input. Load cells were used at each actuator to measure and control the load input to the specimen.
Before the actual fire test can begin, the burner must be adjusted to provide a 2000[Degrees]F [+ or -] 150[Degrees]F (1093[Degrees]C [+ or -] 66[Degrees]C) flame. This was done by adjusting the burner and using an array of thermocouples spaced apart to determine the flame temperature in air. Figure 4 shows the flame calibration per AC20-135  using the array of thermocouples. The heat flux density was measured using a calorimeter and was found to be 9.5 BTU/ft2-sec (104,467 W/m2-[Degrees]C), which meets the advisory circular requirement AC20-135 .
Figure 5 shows the load inputs simulating a landing limit condition of 2g forward, 6.31 g down, 1.12g inboard and full thrust load, where g refers to gravity. With the burner still on after finishing the calibration; the fire test was initiated by directing the burner flame onto the yoke at the pre-defined location. A thermocouple 0.25 inches from the part was placed in the fire zone and the temperature was recorded throughout the test. The 15-minute timer was started when the calibrated burner was turned onto the yoke then the landing limit loads were applied. This landing limit was held for 5 minutes at the peak simulating a fire at limit conditions. The loads were then reduced to 66% of limit without thrust for the remaining 10 minutes simulating the engine having been shutdown. After completing minutes of testing, the burner continued to burn for another 5 minutes with 66% of the limit load without thrust. This allowed the yoke to be completely saturated and reach a near stead-steady-state condition.
Using the thermocouple locations shown in Figure 2, the temperatures were recorded during the entire duration of the test from 0 to 20 minutes. Figure 6 shows that the peak steady-state temperature at about 20 minutes of test time was about 1850[Degrees]F (1010[Degrees]C) at the forward flame thermocouple and about 1500[Degrees]F (816[Degrees]C) at the bottom forward clevis arm. From Figure 6, both the inside and outside of the clevis arm are at about 1500[Degrees]F (816[Degrees]C) therefore, the clevis arm is totally saturated.
FINITE ELEMENT MODEL
The thermal analysis was done using ANSYS[R] finite element analysis software . The thermal analysis methodology was based upon the fire test results of the yoke. The FEA model of the yoke was then correlated to the fire test results by adjusting the convection coefficient bulk temperature. The conduction coefficient of each component defines how much heat flow is within each component and the convection coefficient of each exposed surface defines how much heat flow was transferred from the surface of each component to the fluid medium (air). Radiation was ignored. Table 1 shows the thermal properties of 6A1-4V that were used .
The flame was then modeled as a fluid medium (i.e. air) which was heated to a temperature and then forced over the surface of the yoke. The air flow due to natural draft was calculated by omitting losses due to duct wall friction and the resulting estimated velocity was 588 in/s (15 m/s) from equation 1 . This is based on a room temperature air density at 75[Degrees]F (24[Degrees]C) of [[rho].sub.RT] =4e-5 lb/[in.sup.3] (1.12 kg/[m.sup.3]) and for a flame temperature of 2000[Degrees]F (1366[Degrees]C) of [[rho].sub.F]=9.4e-6 lb/[in.sup.3] (0.26 kg/[m.sup.3]) and an imaginary chimney height of h=118 in (3m) . The peak velocity of the flame is determined to be 588 in/s (15 m/s) at the flame tip, and assumed to be 118 in/s (3 m/s) on the opposite side of the flame impingement, and 40 in/s (1 m/s) far away from the flame tip.
V = [2 x g x ([rho]RT-[rho]F) x h/[rho]r]0.5
where g is gravity, p is density, and h is chimney height.
Figure 7 shows the finite element which represents the overall test structure. The thermal model contains all of the structural components, which significantly contribute to the overall heat transfer of the test configuration. A three dimensional (3-D) FEA model was developed using 10 node tetrahedron elements. The FEA mesh was controlled using local mesh refinement in the yoke region and course mesh in the test structure and load fixtures.
The boundary conditions consisted of applying convection boundary conditions to all exposed surfaces. Figure 8 shows the convection coefficients. The convection coefficient bulk temperature was based on the temperature results obtained during the fire test of the yoke. In exterior surfaces far away from the flame such as the test frame free convection was applied. Figure 8 and Figure 9 show the convection coefficients on all exterior surfaces and the convection coefficient bulk temperatures.
The convection coefficients are based upon the flow of the air over the surface of the yoke. Since the yoke is a complicated structure, the flow of burning gases over the surface of the yoke is highly turbulent air flow. The convection coefficients are approximated by closed-form equations such as flow parallel to a plate and flow perpendicular to the plate . For flow parallel to a plate the convection coefficients were determined based on the Reynolds number-Re and Nusselt-Nu number using equation 2, equation 3, equation 4, equation 5, equation 6. The thermal properties to represent the air surrounding the yoke are used in the finite element thermal model and are shown in Table 2.
[R.sub.e] = V x L/v
where V is the velocity, L is the length, and v the kinematic viscosity. The Prandel-Pr number is defined as
Pr = v/[alpha]
where a is thermal diffusivity. The Nusselt number for laminar flow over a plate is given by
Nu = 0.664 x [R.sub.e.sup.1/2] x [P.sub.r.sup.1/3]
From these relationships the average convection coefficient for flow over a plate is determined from equation 3
h =[N.sub.u] x k/L
For flow perpendicular to the plate it was assumed to use noncircular cylinders in cross flow
[N.sub.u] = C x [R.sub.e.sup.m].[P.sub.r.sup.1/3]
where C=0.228 and m=0.731 for a vertical plate.
Finite Element Model Results of the Fireproof Test
Using the boundary conditions and the material properties described herein, a steady-state and transient thermal analysis was conducted using ANSYS[R] finite element analysis software . Figure 10 shows the overall temperature in the yoke from a steady-state analysis (FEA) using the boundary conditions and material properties described above. The peak temperature is in the bottom forward clevis arm near the attachment hole and was about 1750[Degrees]F (954[Degrees]C). The aft clevis arm was about 1250[Degrees]F (677[Degrees]C). The bulk section of the yoke was at about 700[Degrees]F (371[Degrees]C) and the top was at about 500[Degrees]F (260[Degrees]C).
Figure 11 shows good correlation between the peak temperatures from the FEA model and each of the corresponding thermocouple locations from the fire test as shown in Figure 2.
A transient analysis was done using the same material properties and boundary conditions as the steady-state analysis. The transient analysis was solved using two load steps. The first load step was to establish the initial condition, which was assumed to be at 70[Degrees]F (Room Temperature-RT) for the entire yoke and then load step 2 was set to 20 minutes. Figure 12 shows that the transient analysis approaches the steady-state results at about 15-20 minutes with the shape of the curves and peak values are in good agreement with Figure 6.
MATERIAL STRENGTH AT ELEVATED TEMPERATURES
To satisfy the fireproof structural material requirements not only requires the material to be "fireproof but also exhibit acceptable strength at elevated temperature. The advisory circular AC20-135 section 4a states: "Fireproof: The capability of a material or component to withstand, as well as or better than steel, a 2000[Degrees]F(1093[Degrees]C) flame [+ or -]150[Degrees]F (66[Degrees]C) for 15 minutes minimum, while still fulfilling its design purpose". The design purpose in this example is to be able to react load at elevated temperatures. Figure 13 shows that the strength at elevated temperature is much higher than the previous published data and also suggests that extrapolating from existing MIL-HDBK-697A data could be very inaccurate .
The fireproof test showed that under a limit load and in the presence of a 2000[Degrees]F (1093[Degrees]C) flame, the 6A1-4V Titanium yoke structure was indeed fireproof because it satisfactorily supported the load. The finite element analysis model using basic assumptions for flow parallel and perpendicular to a flat plate to determine the convection coefficients showed good correlation to the peak temperature as well as the overall temperature profile in a transient analysis. The transient thermal analysis showed that the yoke reached steady-state conditions somewhere between 15 to 20 minutes. The strength data at elevated temperature shows that the material retains about 20% of its strength compared to room temperature and can be used for design purposes to determine if a structural component could withstand the loads and have the desired strength to meet the fireproof requirements. As aerospace propulsion and airframe companies strive to build lighter weight planes the use of Titanium as a structural component will increase. The requirement to be fireproof will continue to be an important certification issue. Historically, only steel has been certified by definition as a "fireproof" material without any additional analyses or tests. However, this work shows that Titanium may be used as a lightweight alternative and can also be certified as a "fireproof" material.
[1.] United States Department of Transportation - Federal Aviation Administration, "Federal Aviation Regulations (FAR)", 14 CFR 23 and 25, Office of the Secretary, Distribution Service Branch, M-484.1, Washington, DC 20590, Sections: 23.865, 25.865.
[2.] Federal Aviation Administration, Advisory Circular AC No: 20-135, February 6.
[3.] Code of Federal Regulations (CFR), Title 14-Aeronautics and Space, Part 1-Definitions and Abbreviations, 2010.
[4.] SAE AMS4928Q, "Titanium Alloy Bars, Wire, Forgings, Rings, and Drawn Shapes 6AL-4V Annealed," SAE International Aerospace Material Specification, April 2001.
[5.] SAE AMS-H-81200B, "Heat Treatment of Titanium and Titanium Alloys," SAE International Aerospace Material Specification, July 13. 2003.
[6.] ANSYS[R] Mechanical Release 11, Finite Element Software, ANSYS. Inc, 2009.
[7.] MIL-HDBK-5H, "Metallic Materials and Elements for Aerospace Vehicle Structures", December 1, 1998.
[8.] The Engineering Toolbox, "Air Flow and Velocities due to Natural Draft.", n.p. n.d. Web. http://www.engineeringtoolbox.com/natural-draught-ventilation-d_122.html.
[9.] Incropera, F.P., De Witt, D.R., "Fundamentals of Heat and Mass Transfer", 4th Edition, 1996.
[10.] MIL-HDBK-697A, "Military Handbook: Titanium and Titanium Alloys", June 1, 1974.
To contact the author please send an email to firstname.lastname@example.org.
I would like to thank LORD Corporation for allowing me to publish this work and Matt McGill from LORD Corporation in acquiring the tensile strength data at elevated temperature and David Southwick from LORD Corporation for determining the flame velocity.
v - kinematic viscosity
h - Average convection coefficient
k - Material conductivity
C - Constant
L - Characteristic length
m - Constant
[N.sub.u] - Nusselt number for laminar flow over a plate
[P.sub.r] - Prandel number
[R.sub.e] - Reynolds number
V - Flow Velocity
[alpha] - Diffusitivity
g - Gravity Constant
[[phi].sub.rt] - Density of air at RT
[[phi].sub.F] - Density of air at Flame Temperature
h - Height of chimney duct
Douglas Leicht Lord Corporation
Table 1. Thermal Properties of Titanium 6Al-4V . Specific Conductivity-k Temperature-T Heat-C(BTU/Ib-F) (BTU/min-ft-F) (deg F) 0.13 5.83E-03 70 0.13 6.11E-03 200 0.14 6.94E-03 400 0.15 8.33E-03 600 0.16 9.58E-03 800 0.17 1.08E-02 1000 0.19 1.21E-02 1200 0.21 1.32E-02 1400 0.23 1.43E-02 1600 Specific Heat-Cp Conductivity-k Temperature-T (kJ/kg-K) (W/m K) (deg C) 0.54431 7.29E+00 21 0.54431 7.64E+00 93 0.58618 8.68E+00 204 0.62805 1.04E+01 316 0.66992 1.20E+01 427 0.71179 1.35E+01 538 0.79553 1.51E+01 649 0.87927 1.65E+01 760 0.96301 1.79E+01 871 Table 2. Finite Element Model Thermal Properties of Air at Temperature. Specific Kinematic Temp DENSITY-[ ] Heat-Cp Viscosity-[ ] (deg F) (lbm/[in.sup.3]) (BTU/lbm-F) (lbm/in min) 1500 8.389E-06 2.938E-01 2.232E+01 1800 6.991E-06 3.071E-01 3.060E+01 1300 9.679E-06 2.840E-01 1.721E+01 Specific Kinematic Temp DENSITY-[ ] Heat-Cp Viscosity-[ ] (deg C) (kg/[m.sup.3]) (kJ/kg-K) ([m.sup.2]/s) 816 0.2322 1.2300 2.400E-04 982 0.1935 1.2860 3.290E-04 704 0.2679 1.1890 1.851E-04
|Printer friendly Cite/link Email Feedback|
|Publication:||SAE International Journal of Aerospace|
|Date:||Sep 1, 2015|
|Previous Article:||Panel Assembly Line (PAL) for high production rates.|
|Next Article:||Design and evaluation of plug-and-play enabled IVHM architecture.|