Wavelet Homepage                               Table of Contents

8 Engine Mount Simulation Test Plan  

            This section of the report will give some background information for the simulation; test plans for the four suggested tests; and refer to the equilibrium analyses in Appendices B and C, which must be redone for verification purposes.  Future ASE 463Q groups should implement the test plans in this section so that they can acquire dependable test data and maybe relate it to the engine mount on the Beech 1900C airplane.  We are concerned with relating fracture tests from the engine mount provided to the one on the airplane because of their differing geometries; however, we have constructed a series of tests that may allow the fracture tests to be related.

  8.1 Background 

            Although simulating an on-flight failure is not a trivial task, one may be able to pinpoint the cause of the crash through a signal analysis coupled with testing.  Such an analysis involves assuming certain failure modes, and simulating them in an attempt to replicate flight conditions.  Once the simulation is complete, it is imperative to determine whether the transients found on the CVR and those acquired during the test are similar enough to make on-flight failure conclusions.  In addition, there is the possibility that the transients of most fractures (whether a tube or a bolt) resemble each other, and depend on location on the airplane.  Also, the wires may have limited frequency detection capability so that only some of the frequencies are acquired.  Such a complex scenario requires that the analysts carefully consider possible conclusions, being careful not to misinterpret the facts.  Resemblances in signals do not necessarily mean that the analysis is powerful enough to determine exactly what the cause of the crash was, for they may only constitute the conclusion that something fractured or that the data is inconclusive.  In Section 5, we determined that for diagnosing a break, the orientation of the wires significantly mattered because an impulse force that did not result in a fracture yielded different results for different wire orientations. 

In this particular analysis, we assumed two possible primary failure modes, which are an engine mount bolt fracture or a tube fracture.  The engine mount tube fracture analysis involves a series of tests to be conducted in order to determine whether the location of the recorder wire and the geometric properties of the tube affect the signal acquired.  Closely simulating the engine mount failure is a key aspect of determining whether the transients found on the CVR are indeed fractures of engine mount tubes or bolts, and it is likely that an engine mount tube or bolt fracture caused the crash because the engine mount is in the infant mortality failure category [18].    The following section outlines the Wavelet group’s progress during the Fall 2002 semester. 

Originally, our group wanted to complete testing on an engine mount; however, after determining the need to analyze the effects of geometry on the signal acquired, our objectives changed because the problem became more complicated.  There are significant differences between the truss on the Beech 1900C airplane (see Figure 72) and the truss available to students in the ASE 463Q class (see Figure 73).

Figure 72: Truss on Beech 1900C Airplane

 

Figure 73: Truss Available for Testing

 

 

The arrangements of the truss members as well as the tube diameters are considerably different.  We therefore determined that a detailed Test Plan is a more significant contribution towards further progress.  The purpose of this Test Plan is to guide future ASE 463Q groups to successfully simulate an engine mount truss failure, and to determine whether the Beech 1900C airplane crashed due to an engine truss tube failure or a bolt failure. 

  8.2 Purpose of the Tests 

        There are four primary tests that should be completed in order to decisively determine the cause of the failure on the Beech 1900C.  These tests are an impulse hammer test, a single-tube fracture test, an engine mount test, and a single-bolt fracture test.  The purpose of the impulse tests and the fracture tests is to obtain the characteristic signature of the impulse force or the break, respectively, for different failure configurations.  The data will be obtained with a reel-to-reel audio recorder in order to properly simulate the aircraft’s CVR.  The data from these tests will first be digitized, and then it will be analyzed by a time-series assessment in Matlab.  After the analysis, the data will be evaluated with the wavelets toolbox in Matlab, and then finally, it will be compared to the CVR data from the Beech 1900C.  An alternative to the reel-to-reel recorder would be to directly connect the wires to the HP Signal Analyzer in the Aeroelasticity Lab.  The advantage is that the data would be directly digitized, while the disadvantage is the slow data transfer rate of one second per frame and dissociation to flight conditions.  If similar transients are found on both the aircraft data and simulated test data, and the test setup was engineered so that the signature of a simulated failure is similar to the one on the airplane, then it can be concluded that an engine mount tube or bolt on the plane did indeed fail.  Some of the test objectives may be answered by previous tests or with further research into the Aeroelasticity Laboratory’s computer files.  There is an archive of data that need to be identified and then analyzed.  The following sections will discuss in detail each of the proposed test objectives, materials, and procedures.

  8.3 Impulse Hammer Test 

            The idea behind this test is to strike an impulse hammer against a tube similar to the engine mount tube on the Beech 1900C that is believed to have failed (see Figure 25).  The purpose of this test is to determine if the wire geometry around the tubes affect the signal acquired by the CVR.  Note that since the exact engine mount tube that fractured during flight is unknown, a best guess will have to be made because the engine mount tubes vary in size and length; therefore, the chosen tube for this test may not exactly match the tube that failed in flight.   The goal for this test is to identically strike the tube multiple times, while changing the wire arrangement, the number of wires, and the length of the wires. The primary step is to see if the setup of the wires affects the signal that is recorded. 

       8.3.1 Test Objectives  

1)      Determine if test setup drastically changes the signal acquired by the CVR

2)      Establish relationships between wire configurations, tube lengths, wire lengths, the number of wires, and the characteristic signature of the signal

       8.3.2 Test Materials  

The primary test specimen is a tube that matches one of the engine mount tube’s material and geometric properties.  A tube nearly matching one of the engine mount members has already been purchased (it is cold-rolled steel rather than heat-treated); it just needs to be sized properly.  Other materials that are readily available in either the machine shop or the Aeroelasticity Lab are also needed for impulse testing.  Table 3 describes the major components necessary to complete this test. 

 

Table 3: Required Impulse Test Materials

 

      

The input parameter in this test will be the impulse force produced when the hammer is stuck against the tube, and the output will be the signal resulting from the wire or wires attached to the tube in various configurations. 

       8.3.3 Test Procedure  

1)      Clamp a V-Block on the base structure and connect the tube on the V-block

2)      Place a wire in one of the following configurations and connect the wire to the data recorder (single wire helical, single wire longitudinal, braided pair helical, braided pair longitudinal, silent, and microphone—the description of these wires is given after all of the test plans are described)

3)      Turn on the data recorder

4)      Impact the end of the tube with the impulse hammer recording both the input and the output signals

5)      Repeat test for remaining wire configurations (see Step 2) holding the length of the tube constant

6)      Repeat test for each wire configuration with wires half as long

7)      Repeat above steps for a 10 foot tube

  8.4 Single-Bolt Pop Tests 

The purpose of this test is to characterize the signal of a bolt fatigue fracture.  The type of bolt on the 1900C engine mount can be determined from the design drawings, which Dr. Ronald Stearman handles.  After the bolt is identified, testing can proceed.  The drawings for the engine mount did not describe the bolt; however, an assembly drawing of the engine should identify the bolts’ material and geometric properties.   Hopefully, the bolt can be bought at a local hardware store and if not, then we recommend McMaster-Carr or some other supplier Dr. Ronald Stearman might suggest.  The most likely mode of failure for one of the engine mount bolts is fatigue.  Fatigue was suspected after viewing wreckage photographs of one of the engine mounts [18].  To fatigue test a bolt, a final more detailed test procedure is required; fatiguing a bolt is something our group did not consider until the end of the semester.  After consulting with Dr. Stelios Kyakides and Dr. Ken Leichti, we realized that to compare the characteristic break of a bolt from our test to the CVR data, we need to load the test bolt to approximately the same conditions the bolt experienced in flight [19] [20].  Since we did not know the flight loads and did not have time to try to estimate them, a future group will have to calculate them.  Another primary difficulty we had was calculating the size of the crack to saw into the bolt to ensure that the bolt fails (critical crack length).  This was because we had not yet taking a Materials class.  Once these two issues are resolved, the test objectives should be achievable.

       8.4.1 Test Objectives  

1)      Accurately simulate the loading conditions on the Beech 1900C airplane engine mount

2)      Determine the frequency range and magnitude of a single-bolt fracture

3)      Evaluate the time span in which the fracture occurs

4)      Ascertain the shape of the transient of a single-bolt fracture in both the frequency and time domain

5)      Determine which frequencies are damped out by a microphone, and if the setup of the test affects the signature obtained

6)      Complete this fracture test for the type of failure that is most likely to occur (fatigue is our predisposition)

       8.4.2 Test Materials

            A bolt fatigue failure was not considered until the end of the semester, and we did not determine a complete list of the necessary test materials.  However, Dr. Stearman suggested that a bolt be screwed into the end of a metal plug attached at the end of a tube.  As of now, the required testing materials are described in Table 4. 

 

Table 4: List of Necessary Materials for Bolt-Pop Test

 

 

       8.4.3 Test Procedure

1)      Make a 3 inch plug to fill one of the ends of the tube

2)      Attach the plug to the tube using 4 ¼ inch screws

3)      Screw the bolt to be broken into the end of the plug, leaving an inch or sticking out of the plug

4)      Saw a (size) crack in the bolt with a fine saw

5)      Connect the end of the 3 ft. tube with the bolt to the MTS machine and support the other end with a V-block attached to a steel post.

6)      Wrap the wire around the bolt, starting from after the break point.

7)      Repeat for different wire configurations (if necessary after analyzing the results from the impulse hammer testing)

  8.5 Single-Tube Pop Tests           

The purpose of the single tube pop tests is to obtain the characteristic signature of a tube breaking, and then determine if the length of the tube significantly changes the signal acquired by the CVR.  This test would primarily be used to determine whether or not a tube fracture looks similar or is independent of test configuration.  One proposed theory is that the energy released from the break travels in a packet through material such that, the CVR would record the same signal regardless of the test setup.  On the other hand, both Dr. Kyriakides and Dr. Leichti assured us that the signal would be different because of the different materials, the unequal propagation distance, the type and strength of the signal input into the wire, and the age and type of wires that are recording the signal.  The single-tube pop test will verify whether or not the signature of a signal is dependent on the setup, and if it is, then maybe a relationship can be determined (say like a second-order system).  For instance, perhaps only the amplitude changes from different test setups, which would allow us to fracture this engine truss and relate it to an on-flight fracture magnitude.  On the other hand, maybe the signature depends on the test setup enough that a break could be confused with a nominal flight event.  Nonetheless, with careful planning, the above issues can be resolved and the following objectives will be met.

       8.5.1 Test Objectives  

1)      Accurately simulate the loading conditions on the Beech 1900C airplane engine mount

2)      Determine the frequency range and magnitude of a single-tube fracture

3)      Evaluate the time span in which the fracture occurs

4)      Ascertain the shape of the transient of a single-tube fracture in both the frequency and time domain

5)      Determine the ability of the non-acoustic channels to couple, or detect, the fracture

6)      Determine the type of waves (longitudinal or transverse) that propagate through the material as a result of different types of failures and determine the relative strengths of the transients produced by each

7)      Determine which frequencies are damped out by a microphone and if the setup of the test effects the signature obtained

8)      Acquire the effects of braided wires on the fracture signature

9)      Complete this fracture test for the failure that is most likely to occur, either fatigue, tension, or compression failure (fatigue is our predisposition)

       8.5.2 Test Materials  

Table 5 lists the materials necessary to complete the tube pop tests.    

Table 5: Required Testing Equipment for Tube Pop Test

Description	Quantity
3 ft 4130 Steel Tubes	5
1" Dia, 0.049" Thick
1 ft 4130 Steel Tubes	5
1" Dia, 0.049" Thick	?
Recorder at Least 6 Channel	1
MTS Testing Machine	1

These tests will likely be completed in the Aerospace Materials Laboratory at The University of Texas at Austin.

8.5.3 Test Procedure  

1. Saw a crack near the base of a 3 ft tube about the critical crack length size
2. Connect the end of the 3 ft. tube closest to the crack to the MTS machine and support the other end with a V-block attached to a steel support
3. Wrap the helical braided wire pair around the tube, starting from after the break point
4. Duct tape the longitudinal braided wire pair to the tube, again starting from after the break point
5. Lay a silent wire over the engine of the MTS machine
6. Attach microphones to the necessary wires.  Caution: Be careful to leave slack on the wires so that the microphones can be moved around if necessary
7. Turn on the data recorder and begin recording data
8. Turn on the MTS machine and set it to 5 Hz

9.      After the tube breaks and ceases to vibrate, stop recording and turn off the machine

  8.6 Engine Mount Test 

The engine mount that is readily available to students in the Aerospace Engineering Department is not an exact replica of the truss on the Beech 1900C airplane.  It is a different truss and the member configurations are not similar; however, it gives the students the opportunity to evaluate the effects of geometry on a CVR signature.  The effect of geometry on an engine mount failure is very important because it will help determine whether students can accurately simulate an engine mount failure with a single-tube, and then relate it to the airplane.  Also, this test has the capability of concurrently testing different wire configurations. 

Professor Stelios Kyriakides, the Director of the Structures and Materials Department at the University of Texas at Austin, believes that the type of failure must somehow be identified [19].  Whether it is a fatigue, tension, or compression failure may be determined by examining the post-incident truss.  In addition, he believes that there will be a strong difference between the signals produced by different sized tubes.  The loading conditions for the truss have not been determined, and are important in order to properly simulate the engine mount member failing.  The loading conditions for the airplanes’ drag and lift will determine the types of loads to which the truss should be exposed.  The stress calculation in Appendix B assumes a static load is probably not an accurate loading representation for flight, but will work for low frequency inputs.

If the Beech 1900C airplane crashed due to an engine mount failure, then fracturing a member on an engine mount should produce transient signals similar to the transients found on the CVR data.  If the impulse and tube-pop test prove that wire configuration and test setup is minimal, then the engine mount test fracture is a good test, and may be linked to the actual airplane.  However, turbulence can produce signals that are hard or even impossible to simulate.  Therefore, the engine mount simulation will focus on the actual breaking of the engine mount member that fractures under the most stress.  Any similarities before and after the failure are subject to errors due to the air turbulence causing unpredictable collisions or vibrations of the broken members. 

       8.6.1 Test Objectives  

1)      Accurately simulate the loading conditions on the Beech 1900C airplane engine mount

2)      Obtain the signature of an engine mount test failure in either tension or compression

3)      Fracture the member most likely to fail on the airplane that crashed

4)      Determine the frequency range and magnitude of a single-tube fracture

5)      Evaluate the time span in which the fracture occurs

6)      Ascertain the shape of the transient of a single-tube fracture in both the frequency and time domain

7)      Determine the ability of the non-acoustic channels to couple the fracture

8)      What type of waves propagate through the material as a result of different types of failures, and the relative strengths of the transients produced by each

       8.6.2 Test Materials  

            There was more progress made in the engine mount test that any other section of the testing phase.  The most current list of materials is in Table 6. 

Table 6: Necessary Materials for Engine Mount Tube Fracture Test

 

 

 

       8.6.3 Test Procedure  

1)      Saw a critical crack length sized crack near the base of one of the joints

2)      Attach the engine mount on the test setup (Figure 74)

3)      Bolt a 12”X 12”X 0.5” inch plate to the front of the engine mount

4)      Statically load the plate attached to the engine mount to accurately simulate the weight (100-150 lbs)

5)      Turn on the data recorder and begin recording data

6)      Utilize a hydraulic jack to dynamically load the engine mount at a certain angle (depending on the expected percentage drag contribution)

7)      After the tube breaks, stop recording and turn off the machine

Figure 74:  Potential Test Setup

 

  8.7 Wire Configuration 

            The test configuration will vary from test to test; however, for all of the test configurations the wiring of the data recorder will be the same.  For any given test, data will need to be recorded on three to six channels, including an audio microphone to narrate the testing.  For most of the testing, it is necessary to use silent wires (no microphone) because silent wires simulate an open circuit with virtually infinite impedance.  There are silent wires on the Beech 1900C airplane and we want to simulate its flight conditions.  One of the remaining wires will be a silent track placed on the motor of the MTS machine, and the last wire will have a microphone that will record a narrated/acoustic version of the test.  The test configurations for the non-braided wires will be exactly the same as the braided wires tests; however, the braided wires will be separated so that only one of the two are on the specimen.

 

Wavelet Homepage                               Table of Contents