The Ultra Light Aircraft Testing

APPENDIX E C*) : 2-D NASTRAN RESULTS APPENDIX F (*) : S-D NASTRAN RESULTS * APPENDIX E AND F ARE COMPUTER OUTPUT DATA WHICH ARE ATTACHED SEPARATELY.

LIST OF SYMBOLS Dimension Definition Symbol ksi Modulus of Elasticity E ksi Compression modulus of Elasticity Ec ksi Modulus of Rigidity G ksi Compression Yield Strength Fcy ksi Shear Yield Strength Fsy ksi Tensile Yield Strength Fry ksl Tensile Ultimate Strength Ftu Definition Acronyms Material MAT NASA Structural Program NASTRAN Definition Greek Symbols Poisson's Ratio Density (l_/Ft _) i. INTRODUCTION The purpose of this report is to create a three dimensional NASTRAN model of the Airmass Sunburst Ultralight comparable to one made for finite element analysis. A two dimensional sample problem will be calculated by hand and by NASTRAN to make sure that NASTRAN finds the similar results. A three dimensional model, similar to the one analyzed by the finite element program, will be run on NASTRAN. A comparison will be done between the NASTRAN results and the finite element program results. This study will deal mainly with the aerodynamic loads on the wing and surrounding support structure at an angle attack of i0 degrees.

- I

2. 2-DIMENSIONAL MODEL The purpose of this chapter is to create a two dimensional truss model similar to the Sunburst Ultralight front spar and the three flying wires. The static loads to be used are calculated from the aerodynamic loads at an angle of attack of 10 deg. The resultant element forces will be calculated manually and by use of NASTRAN. From these results, a comparative study will he made between the NASTRAN results the results achieved by manual calculation.

2.1 MODEL DESCRIPTION The purpose of this section is to describe the major assumptions used to create the 2-dimensional model of the Sunburst Ultralight. It is assumed for this analysis that the root beam and the two wire nodes are fixed. The resulting model will be essentially a fixed cantilever beam attached to three truss elements in tension. The following Nodes will be fixed: Node i, Front Spar and Root Beam connection Node 9, Cable end Node 12, Cable end Figures 2.1.1 and 2.1.2 show the dimensioned truss and the nodal data for the 2-dimensional model. The following subsections contain the information required for the NASTRAN program to be completed. The Sub-sections contain the following: Node and Constraint identification Element Description Material Description Wing Loading Calculations With this information, the resulting NASTRAN program can be run on the University of Kansas VAX system.

o o co ._J LLJ (D cO cO n-" F-- iii Z Z LO = c_ o LU LO < OE o © LL _F O_ D ED

I p

F- D u3 U3 rr" I-- w Z i O,J rr" U.._ l--- F-'l J F7 z _d "7.

U rt" © U_ 13_ F-'I u.l rn I-- Z Z3 2.1.1 NODE POINT AND CONSTRAINT DEFINITIONS The purpose of this section is to identify the grid points used and the constraint at each point. The constraints used by NASTRAN are as follows: 1 = Linearly constrained in the X-direction 2 = Linearly constrained in the Y-direction 3 = Linearly constrained in the Z-direction 4 = Constrained about the X-axis; Ox = 0 deg.

5 = Constrained about the Y-axis; Oy _ 0 deg.

6 = Constrained about the Z-axis; Oz = 0 deg.

The following table contains the GRID cards used in the NASTRAN program for the 2-D model. Table 2.1.1 also includes the single point constraints for each point and the GRIDSET card for the default constraints.

Table 2.1.i: GRID and GRIDSET Cards used in NASTRAN NASTRAN X Y Z CONSTR- CARD (AFT) (OUTB'D) (UP) AINT (IN) (IN) (IN) GRIDSET 1,4,5,6 GRID #i 0.0 2.6 66.3 123456 " #2 0.0 75.0 66.3 " _3 0.0 155.0 66.3 " #9 0.0 i0.9 28.3 123456 " #12 0.0 13.2 25.6 123456 2.1.2 ELEMENT IDENTIFICATION The purpose of this section is to identify the elements used in the 3-dimensional NASTRAN model. The following table shows the elements used and their descriptions.

Table 2.1.2: Element Descriptions DESCRIPTION ELEMENT NUMBERS (EID) Wing Spars; 1.75" Diameter 1,2,3 Tubes, t = 0.049" (Fig. 2.1) Flying Wires; 4,5,6 (4) Diameter = 3/32" (Fig. 2.1) (5,6) Diameter = 1/8" 2.1.3 ELEMENT MATERIAL IDENTIFICATION The purpose of this section is identify the materials used for each element of the ultralight model. The tube information is referenced from the ultralight model handbook. The cable information is experimental data taken from the analysis performed by students under the supervision of Dr. Howard W.

Smith. The following are the material identifications for each element in the 3-dimensional model and pertinent material information: Material ID = I; EID : 1,2,3 6061-T6 Tube, Spec : WW-T-700/6 Ftu = 42. ksi Fcy = 34. ksi Fsy = 27. ksi E = 9.9+3 ksi Ec = 10.1+3 ksi = 0.33 = 0.098 lb/in^3 (Ref. 3, Table 3.6.1.0(b)) Material ID : 2; EID : 4,5 Alloy steel cables, Experimental Data Ftu = 864. psi E = 29.0 +3 Ksi = 0.33 W = 0.283 ib/in^3 The materials used are assumed to be linear, temperature independent, isotropic materials. Therefore, MAT1 cards will be used in the NASTRAN program.

2.1.4 WING LOADING AND FORCE CALCULATIONS The purpose of this section is to determine the forces on the wing nodes which must be equivalent to the wing loading. The wing loading was taken-from test data in Reference I, Table 3.3.2. The table and the calculations used to obtain the forces on the nodes can be found in Appendix A. The following are the results of these calculations: Node i, F1 = 56.1 ibs Node 2, F2 = 55.6 ibs Node 3, F3 = 30.2 Ibs These forces are considered static and thus Force cards will be used in the NASTRAN program. The forces are considered to act in the vertical, (z) direction.

2.2 MANUAL CALCULATION OF RESULTING FORCES AND MOMENTS The purpose of this section is to calculate the resulting forces at each node for the 2-d model with the static loads.

Manual calculations for the 2-dimensional truss model can be found in Appendix B. The following are the resulting element forces and stresses: ELEMENT AXIAL FORCES AND STRESSES (APPENDIX B); AXIAL AXIAL ELEMENT FORCE STRESS (ibs) (psi) -186. 710. (COMP.)

i- (TUBE ) -89.2 430. (COMP.)

2 - (TUBE ) +94.2 13600.(TENSION) 4- (CABLE ) +0.70 i01. (") 5- (CABLE ) +i03. 8370. (") 6 - (CABLE ) 2.3 NASTRAN CALCULATION OF RESULTING FORCES AND MOMENTS The purpose of this section is to use the NASTRAN program to calculate the forces at each node for the 2-D model with the static loads. Appendix C contains the NASRTAN program for two dimensional model to be analyzed. The program was run and the resulting output from NASTRAN can be found in Appendix C, attached separately. The following are the nodal displacements and the element forces calculated by NASTRAN: NODAL DISPLACEMENTS (APPENDIX E); GRID POINT X Y Z (in) (in) (in) 1,9,12 0.0 0.0 0.0 2 0.0 -.00504 +.0314 3 0.0 -. 00809 +. 173 ELEMENT AXIAL FORCES AND STRESSES (APPENDIX E); !w SAFETY MARGIN ELEMENT FORCE AXIAL (ibs) STRESS -689. (COMP.) 4.8 1-(TUBE) -180.5 -377. (COMP.) 8.9 2-(TUBE) -98.7 8390. (TENSION) -0.90 4-(CABLE) 103.2 5350. (") -0.84 5-(CABLE) 65.8 4807. (") -0.82 6-(CABLE) 33.2 The displacements of the nodes 2,3 which are wing nodes are physically displacing in the correct direction. The wing, under the wing loading, will move in the up and inboard direction as if it were rotating about the root beam. It can be seen that for the experimentally calculated failure stress of the wire (Ftu = 842. psi) that all the safety margins are negative, as calculated by NASTRAN. This means that the wires are loaded beyond the experimental failure stress.

2.4 CONCLUSIONS AND RECOMMENDATIONS The purpose of this section is to comment on the results of the previous section and give some recommendations on the results.

2.4.1 Conclusions The purpose of this section is to provide a summary of the previous chapter. The following are the element forces calculated manually: ELEMENT AXIAL FORCES AND STRESSES (APPENDIX B); ELEMENT AXIAL AXIAL FORCE STRESS (ibs) (psi) 1-(TUBE) -186. 710. (COMB.)

2-(TUBE) -89.2 430. (COMB.)

4-(CABLE) +94.2 13600.(TENSION) 5-(CABLE) +0.70 i01. (") 6-(CABLE) +103. 8370. (") the forces calculated by use of the NASTRAN The following are program: ELEMENT AXIAL FORCES AND STRESSES (APPENDIX E); ELEMENT FORCE AXIAL SAFETY MARGIN (Ibs) STRESS I-(TUBE) -180.5 -689. (COMP.) 4.8 2-(TUBE) -98.7 -377. (COMP.) 8.9 4-(CABLE) 103.2 8390. (TENSION) -0.90 5-(CABLE) 65.8 5350. (") -0.84 6-(CABLE) 33.2 4807. (") -0.82 It can be seen that the results of the NASTRAN program and the manual calculations are compatible except for the values calculated for Element 5 and 6. The difference that does exist is due to NASTRAN taking into account the displacements of the wing root (Grid Points 2,3,4). It can be seen that the sum of the forces of elements 5 and 6 almost equals the sum of the same elements calculated by NASTRAN. The manually calculated values for element 5 and 8 must be off by a fraction of each. It is concluded that the NASTRAN program will produce correct results/ for the 3-dimensional model to be analyzed in Chapter 3.

iO

2.4.2 Recommendations is to give recommendations on The purpose of this section is recommended that the nodal the results of the chapter. It hand calculations to compare displacements be included in the with the NASTRAN output.

S. S-DIMENSIONAL ULTRALIGHT MODEL The purpose of this chapter is to create a S-dimensional ultralight model of the wing and surrounding structure to be used by the NASTRAN program. The forces, moments, and displacements of each node and the element stresses will be calculated by the NASTRAN program. These results are to be compared with those obtained by the finite element method calculated in Reference 1.

3.1 MODEL DESCRIPTION The purpose of this section is to describe the major assumptions used to create the model. It is assumed for this analysis that the root beam is fixed. Therefore, the following nodes will be fixed: Node I; Front Spar and Root beam connection Node 8; Rear Spar and Root beam connection Node 10; Forward truss attachment point Node II; Aft truss attachment point Nodes 1 and 2, however, are hinge attachments in which the front and rear spar are free to rotate about the Z-axis. This will be dealt with in the single point constraint for nodes 1 and 3. Figure 3.1 to S.3 show the top views of the model with the Nodes and Elements identified. The figures show the wing internal cables (Fig. 3.1), wing flying wires (Fig. 3.2), and the truss members (Fig.3.3). Figure 3.4 shows an isometric of the complete model for visual purposes.

information required The following subsections contain the The Sub-sections for the NASTRAN program to be completed.

contain the following: Node and Constraint identification Element Description Material Description Wing Loading Calculations With the information calculated and identified in these sub- sections the NASTRAN program can be written.

IZ

© Lr_ Z II = _.1 W Z LU k- Z C_ Z © Z L_ r_ Go UJ tU uJ iii ED -- (D {h 1.1_ _o _ 0 -- cD 0 0 co r-- 0 O LO ILl Z LO LLI _ Z LO W Z Iii Z _J klJ

\2

I-5 O 2; __1 Z O (/3 Z WU?

U_l _ o,1 u.J _ 2_W W t.U N_A LO W UJ IJJ LU Am 0 0 0 Z Z Z Z °° uJ E_ a3 k- I..t_ u._ o LD © II Z I11 .d © (D gO or" .__1 cJD LI'] UJ CO n- uJ uJ uJ LU /D o ED co _ ED uJ 0 0 _ 0 _ _ 0 /D tl_ Z Z iii z w W Z 0 kl_l _..1_

\\\\ \_

t.kl _ C3 O _- ...J <_ Zw m Z U_I w© LLI uJ _ _Z r'h E3 _ (D O 0 LI.I _ Z a_ Z z r_ Z ' j U.I m °_U._ C3 (D /,5 Z ft.)

t.kl rr" C3

-q

t.L t.l_ © z z C3 r_ r_ cO Z © Z

w

u_ uJ Z _jO z °.

(DLU Z_ t.U F'- 'F- rnO. ) .o

LU

C3

Z

w or" u_ rr_ % ( ,.z t,--..--

rr

S,.

L.L 3.1.1 NODE POINT AND CONSTRAINT DEFINITIONS The purpose of this section is to identify the grid points used and the constraint at each point. The constraints used by NASTRAN are as follows: 1 = Linearly constrained in the X-direction 2 = Linearly constrained in the Y-direction 3 = Linearly constrained in the Z-direction 4 = Constrained about the X-axis; Ox = 0 deg.

5 = Constrained about the Y-axis; Oy = 0 deg.

6 = Constrained about the Z-axis; Oz = 0 deg.

The following table contains the GRID cards used in the NASTRAN program for the 3-D model. The table also includes the single point constraints for each point and the GRIDSET card for the default constraints.

Table 3.1.1: GRID and GRIDSET Cards used in NASTRAN NASTRAN X Y Z CONSTR- CARD (AFT) (OUTB'D) (UP) AINT (IN) (IN) (IN) GRIDSET 4,5,6 12345 2.6 66.3 85 49 GRID #I 73.2 75.0 " #2 85 49 80.9 49 155.0 #3 85 87.0 " #4 85 49 218.0 84.0 0 218.0 " #5 115 77.9 0 155.0 " #6 115 75.0 70. 2 " #7 115 0 63 0 12345 2.6 " #8 115 0 28 3 39 10.9 " #9 60 89 0 123456 09 0.0 " #I0 82 84 0 123456 0.0 " #ii 139 6 25 6 2456 8 13.2 " #12 II0 74 1 5 22.0 " #13 205

I?

3.1.2 ELEMENT IDENTIFICATION The purpose of this section is to identify the elements used in the 3-dimensional NASTRAN model. The following table shows the elements used and their descriptions.

Table 3.1.2: Element Descriptions DESCRIPTION ELEMENT NUMBERS (EID) Wing Spars; 1.75" Diameter 1,2,3,5,6,7 Tubes, t = 0.049" (Fig. 3.1) Wing Ribs; 1.00" Diameter 4,10,13 Tubes, t = 0.035 (Fig. 3.1) Wing Internal Cables; 1/8" 8,9,11,12 Diameter (Fig. 3.1) Forward Root Tube Attachment Truss; 1.00" Diameter, (Fig. 3.3) t : 0.075" Tail Attachment Truss Tubes; 15,16 1.125" Diameter, t = 0.065" (Fig. 3.3) Flying Wires; 17,18,19,20,21 (17) Diameter : 3/32" (Fig. 3.2) (18-21) Diameter = i/8" Aft Root Tube Attachment 22,23 Truss; 1.00" Diameter (Fig.3.3) t = 0.049" 3.1.3 ELEMENT MATERIAL IDENTIFICATION The purpose of this section is identify the materials used for each element of the ultralight model. The tube information is referenced from the ultralight model handbook. The cable information is experimental data taken from the analysis performed by students under the supervision of Dr. Howard W.

Smith. The following are the material identifications for each element in the 3-dimensional model and pertinent material information: Material ID = I; EID = 1-7,10,13,14,15,16,22,23 6061-T6 Tube, Spec = WW-T-700/6 Ftu = 42. ksi Fcy = 34. ksi Fsy = 27. ksi E = 9.9+3 ksi Ec = 10.1+3 ksi = 0.33 = 0.098 Ib/in^3 (Ref. 3, Table 3.6.1.0(b)) Material ID : 2; EID = 8,9,11,12,17-21 Alloy steel cables, Experimental Data (Ftu) Ftu = 864. psi E = 29.+6 psi (Ref. 3) (..)

= 0.33 W = 0.283 ib/in_3 (,.)

The materials used are assumed to be linear, temperature independent, isotropic materials. Therefore, MAT1 cards will be used in the NASTRAN program.

lq

3.1.4 WING LOADING AND FORCE CALCULATIONS The purpose of this section is to determine the forces on the wing nodes which must be equivalent to the wing loading. The wing loading was taken from test data in Reference i, Table 3.3.2. The table and the calculations used to obtain the forces on the nodes can be found in Appendix A. The following are the results of these calculations: Node i, F1 = 56.1 ibs Node 2, F2 = 55.6 ibs Node 3, F3 = 30.2 Ibs Node 6, F6 = 20.0 ibs Node 7, F7 = 36.5 ibs Node 8, F8 = 32.4 ibs The forces calculated appear to be low. Since these forces are from the information from Reference I, the results should still be consistent. These forces are considered static and thus Force cards will be used in the NASTRAN program. The forces are considered to act in the vertical, (z) direction.

3.2 PROGRAM DESCRIPTION The purpose of this section is to describe the NASTRAN program created for analyzing the Sunburst Ultralight. The program was written with all the information identified in Section 3.I. The NASTRAN program output can be found in Appendix D.

The program is split up into three sections. The first section is the Executive Control Deck. This deck contains the user identification and administrative information. The second deck is the Case Control Deck. In this deck the codes identifying what type of analysis is to be performed is included.

This lets NASTRAN identify what the program wants it to do. The final deck is the Bulk Data Deck. This deck contains all the model information identified in Section 3.1. The program is ready to be submitted at this point.

3.3 NASTRAN RESULTS The purpose of this section is to document the NASTRAN program results. Appendix F contains the NASRTAN program results for the three dimensional NASTRAN model, attached separately.

The reader is advised to look at Figure 3.1-2 to help locate visually the grid points and elements. The following are the nodal displacements and the element forces calculated by NASTRAN for the wing and flying wires: NODAL DISPLACEMENTS; GRID POINT X Y Z (in) (in) (in) 0.0 0 0 00 1,8,10,11 2 0 0192 -0 00747 0 0440 3 0 0654 -0 0281 0 232 6 0 0564 -0 0236 0 146 7 0 0196 -0 0114 0 0462 9 -0 00184 0 0 00 12 -0 00196 0 0 0 00404 13 0.0722 0 0 -0 137 ELEMENT AXIAL FORCES; ELEMENT AXIAL AXIAL SAFETY MARGIN FORCE STRESS (ibs) (psi) 77.

1 (F.S.) -114. -436. (COMP.)

2 (") -72.2 -276. (") 120.

ii0.

6 (R.S.) -81.8 -312. (") 50.

7 (") -175. -668. (") (TENS ION ) 540.

I0 (RIB) +8.21 +77.4 510.

13 (RIB) -7.06 -66.5 (COMP.)

(") 190.

14 (TUBE) -38.3 -176.

15 (") 0.0 0.0 N/A 16 (") 0.0 0.0 N/A 85.

22 (") -57.T -394. (COMP ) (") 39.

23 (") +124. -844.

CABLES; SLACK 8 (Internal SLACK 9 wing) II +14.5 1179. -.28 (TENSION) 12 SLACK +58.6 8486. (TENSION) -0.90 17 (Flying 18 wires) +51.8 5350. (") -0.80 +112. 9125. (") -0.91 2O +72.8 5921. (") -0.86 +81.2 6603. (") -0.87

The displacements of the nodes 2,3,6,7 which are wing nodes

are physically displacing in the correct direction. The wing, under the wing loading, will move in the up and inboard direction as if it were rotating about the root beam. It can be seen that for the experimentally calculated failure stress of the wires (Ftu = 842. psi) that all the wire safety margins are negative, as calculated by NASTRAN. This means that the Ultralight flying wires, if this model is any indication, will fail in the i0 degree angle attack flight condition, if not before.

It can be seen that the highest cable stress is on Cable Element 19. This cable is the critical cable which will fail first. The cable runs from Node 12 to Node 3 (On Front spar).

This can be seen on Figure 3.2. The reason for the high stress level for this wire is the angle at which the cable makes relative to the front spar in the X=0 plane. The force at node three must be countered by a very large cable load for the small angle.

3.4 COMPARISON BETWEEN FINITE ELEMENT METHOD RESULTS AND NASTEAN RESULTS The purpose if this section is to compare the results obtained by the NASTRAN model used in this analysis and those achieved by the use of the Finite Element Method (Ref.l). Due to the different nodes and loading method used, only the cable axial stresses will be compared. The following are the resulting axial stresses for the flying wires calculated by each method: FINITE ELEMENT NASTRAN (3.3) (Ref. i) AXIAL AXIAL ELEMENT AXIAL AXIAL ELEMENT FORCE STRESS FORCE STRESS (ibs) (psi) (ibs) (psi) 17 +58.6 8486. 34 +76.7 10396.

18 +51.8 5350. 35 +44.3 3610.

19 +112. 9125. 37 +222. 18110.

20 +72.8 5921. 38 +145. 11818.

21 +81.2 6603. 36 +65.4 5336.

It can be seen that the values calculated by the finite element method are not very close to those by NASTRAN. This is due to the difference in models and loading scenarios used. The values, however, are comparable in that they follow the same trend. The critical wire is still Element 19 (NASTRAN) or Element 37 (Finite Element).

3.5 CONCLUSIONS AND RECOMMENDATIONS The purpose of this section is to comment on the results of the previous sections and give some recommendations on either the procedures used or the values assumed.

3.5.1 Conclusions The purpose of this section is to provide a summary of the results calculated in this chapter.

It was found that the critical element in the structure is Element 19. This is the flying wire which runs from the pilot cage (Node 12) to the outboard location on the front spar (Node 3). The large force was primarily due to the very low angle that the cable makes relative to the front spar. The axial stress on the cable was much greater than the tested maximum stress of 842.

psi (Experimental data from students under Howard W. Smith).

From the comparison between the NASTRAN results the Finite Element Program results (Ref. I), it was shown that Element 19 was critical in both. The values were not the same between both program results, but the calculated values did have common trends.

3.5.2 Recommendations The purpose of this section is to present recommendations on the results obtained in this chapter. It is recommended that the 3-dimensional model be redone using more nodes so that a better idea of the actual stresses in all the elements can be found. A more enhanced model could use quadrilateral elements for the wing with the actual calculated wing loading. This would get much closer results than the concentrated static loads used in this analysis.

. CONCLUSIONS AND RECOMMENDATIONS The purpose of this chapter is to comment on the results of the major parameters in this report that were to be calculated.

Recommendations will also be written about the values obtained and the methodologies used.

4.1 Conclusions The purpose of this section is comment on the results of this report. It was concluded in Chapter 2 that the results of the NASTRAN program and the manual calculations were comparable.

The difference that did exist is due to NASTRAN taking into account the displacements of the wing root (Grid Points 2,3,4).

It was concluded that the NASTRAN program will produce correct results.

The following are the resulting forces and displacements calculated in Chapter 3 for the 3-dimensional Ultralight Model: NODAL DISPLACEMENTS; GRID POINT X Y Z (in) (in) (in) 0 0 0.0 0 0 1,8,10,11 0 0192 -0.00747 0 0440 0 232 0 0654 -0.0281 0 146 0 0564 -0.0236 -0.0114 0 0462 7 0 0196 -0 00184 0.0 0 0 0.0 0 00404 -0 00196 0 0722 0.0 -0 137 ELEMENT AXIAL FORCES; AXIAL ELEMENT AXIAL STRESS SAFETY MARGIN FORCE (psi) (ibs) _ -436. 77.

-114. (COMP. ) 1 (F.S.)

-276. 120.

-72.2 (") 2(") -312. II0.

-81.8 (") 6 (R.S.)

50.

-668.

-175. (") 7(") 540.

+8.21 +77.4 (TENS ION ) I0 (RIB) 510.

-7.06 -66.5 (COMF. ) 13 (RIB) 190.

-38.3 -176. (,,) 14 (TUBE) 0.0 0.0 N/A 15 (") 0.0 0.0 N/A 16 (") 85.

-394.

-57.7 (COMP ) 22 (") 39.

-844.

+124. (,.)

23 (") CABLES; SLACK 8 (Internal SLACK 9 wing) -.28 +14.5 1179. (TENSION) Ii SLACK 17 (Flying +58.6 8486. (TENSION) -0.90 18 wires) +51.8 5350. (") -0.80 19 +112. 9125. (") -0.91 20 +72.8 5821. (") -0.88 21 +81.2 6603. (") -0.87 It was found that the critical element in the Structure is Element 19. This is the flying wire which runs from the pilot cage (Node 12) to the outboard location on the front spar (Node 3). The large force was primarily due to the very low angle that the cable makes relative to the front spar. The axial stress on the cable was much larger than the tested maximum stress of 842.

psi (Experimental data from students under Howard W. Smith).

From the comparison between the NASTRAN results the Finite Element Program results (Ref. I), it was found that Element 19 was critical in both. The cable stress values were not the same between the two program results, but the calculated values had common trends.

As a result of the analysis performed in this report it is concluded that the Ultralight Airmass Sunburst is unsafe. The outboard flying wire (Element 19) will fail due to the critically low angle it makes with the front spar.

4.2 Recommendations The purpose of this section is to present recommendations on the results of this report. It is recommended that the nodal displacements be included in the hand calculations to obtain the same results. It is recommended that the 3-dimensional model be reworked using quadrilateral elements for the wing with the actual calculated wing loadings used. This would get much closer results than the concentrated static loads used in this analysis.

5. REFERENCES Muller de Almeida, Sergio F., Aerodynamic and Structural 1.)

Analysis of an Ultralight Aircraft, May 6,1986.

Smith, Howard W., Aerospace Structures: Matrix Analysis, The 2.)

University of Kansas, December 1983.

Smith, Howard W. Phd., Aerospace Materials and Processes, 3.)

The University of Kansas, January 1978.

MCS/NASTRAN, Version 6.6.

MCS/NASTRAN, Handbook of Linear Analysis.

N93-2977|

IVo CONSTRUCTION, WIND TUNNEL TESTING AND DATA ANALYSIS FOR A 1/5 SCALE ULTRA-LIGHT WING MODEL Michael D. James Graduate Student Howard W. Smith Professor Department of Aerospace Engineering University of Kansas December 1988 P_tially supported by NASA Langley Research Center Grant #NAG 1-345 SUMMARY This report documents the construction, wind tunnel testing and the data analysis of a i/5 scale ultra-light wing section. The original ultra-light this wing model is scaled after is Dr. Howard W. Smith's structural test ultra- light located at the Lawrence airport.

Wind tunnel testing provided accurate and meaningful lift, drag and pitching moment data. This data was processed and graphically presented as: C vs. of L C vs. Qf D C vs.

M C vs. C L D The wing fabric flexure was found to be significant and its possible effects on aerodynamic data was discussed. The fabric flexure is directly related to wing angle of attack and airspeed. Different wing section shapes created by fabric flexure are presented with explanations of the types of pressures acting on the wing surface.

This report provides conclusive aerodynamic data about ultra-light wing. This topic is well worthwhile for continuing studies.

TABLE OF CONTENTS Item Io Introduction 1 2.

Wing Construction 3.

Wind Tunnel Testing Method i0 4. Test Trials 5. 12 Data Analysis 6.

Wing Fabric Flexure Recommendations and Conclusions 24 7.

A Data Analysis Appendix A: ii I. INTRODUCTION This special project was performed to study the basic aerodynamic characteristics of an ultra-light wing. Few known wind tunnel tests have been performed of ultra-light wings since they are designed to be very inexpensive. Thus, aerodynamic data such as the variation angle of attack with lift coefficient, drag coefficient, or pitching moment coefficient is relatively unknown. Another specialty about ultra-light wings is that aerodynamic data becomes a function of wing fabric flexure, which itself is function of airspeed and angle of attack.

To perform these wind tunnel tests, a one-fifth scale wing model of Howard Smith's experimental test ultra-light was constructed. Particular attention was paid to keeping the wing model true-to-scale so that hopefully scale aerodynamic characteristics could be studied.

This wing was sized to fit in the small subsonic wind tunnel in the basement of Learned Hall. The two column support rod was used for the test mount, the aerodynamic forces were read by a balance table and displayed on a scale. This data was processed and displayed as standard Cl, Cd and Cm vs. alpha data.

m 2. WING CONSTRUCTION The wing construction consisted of five phases: i) scaling the wing 2) plotting the airfoil coordinates 3) sizing the wing 4) selecting materials 5) construction Phase I. Scaling the wing The wing was primarily scaled down by measuring the chord and thickness of Dr. Smith's test ultra-light wing at the Lawrence airport and applying various scales to determine sizing. Scales of i:I0, 1:5 and 1:4 were considered. The scale of 1:5 was selected since it would size a model with a maximum thickness of 1.3 inches and chord of 10.2 inches; ideal size for the small subsonic wind tunnel.

Phase 2. Plotting the airfoil coordinates: In order to perform this step, I visited the Lawrence airport where Dr. Smith's ultra-light is currently hoisted and being prepared for structural testing. To plot the airfoil coordinates, two methods were used: i) plotting points measured on the wing surface 2) plotting points measured inside the wing By plotting both sets of coordinates, erroneous data points could be eliminated and the airfoil surface could be developed. An airfoil section is shown in Figure 2.1. Note the flat bottom of the airfoil and the constant slope in the upper camber between half chord and the trailing edge.

Figure 2.1 also shows the location and attitude of the mounting block in the wing. The mounting block is situated so that an angle of attack range of +20 to -I0 degrees can be achieved.

Phase 3. Sizing the wing: The wing was sized to create approximately 25 pounds of lift at maximum angle of attack at an airspeed of 75 feet per second. A maximum lift coefficient of 1.6 was assumed.

It was figured that a wing area of 2.3 square feet feet was needed. The wing span was incremented by a scale rib -spacing until the size was either 2.3 square feet or until the span was too large for the tunnel. A wing with four rib spacings was calculated to have an area of 2.0 square feet and a span of 2.35 feet. Perfect! the area requirement is close and it fits in the tunnel (with an inch on each wing tip to spare).

Phase 4. Selecting materials: Since "scale" materials were too hard to find and were usually to expensive or hard to work with, substitute materials were used. A list of the materials and their uses is: Material Slze Purpose i) Birch dowels 318" front spar (leading edge) 1/2" rear spar 3132" 2) Birch plywood wing ribs 5-ply 3) Oak block i" thck mounting attachment 4) Music wire 1/32" trailing edge stiffeners 1/16 lower surface fabric supports, wing chord trailing edge supports 5) Nylon fabric wing fabric used for wood-metal bonds 6) Two ton epoxy --- used for wood-wood and 7) Wood glue wood-fabric bonds Phase 5. Construction: Construction started by preparing the the wing ribs.

First the plywood sheet was cut, mounted together and bonded lightly. Wing rib templates were laid out and holes for the front spar and rear spar were drilled. Next the wing ribs were cut out by a ban saw which insured that each rib would be the same size and shape. They were separated, sanded and bonded together in pairs. A 2.5 inch section of music wire was epoxied into a groove cut in to the trailing edge to simulate the trailing edge shape of the airfoil.

The wing ribs were glued onto the front and rear spars maintaining a 115 scale distance between each wing rib and a 1 inch spacing betwee_ the two center ribs for mounting block.

Once the main wing structure was bonded together, the trailing edge music wire was added. The music wire in the model performs the function of the cables in the ultra- light. The music wire was soldered and glued to the trailing edge of the plywood wing ribs and the music wire extensions. Solder and epoxy lumps were files out to keep the trailing edge to a minimum thickness. 1/16" music wire supports were added in a criss-cross fashion between the • leading edge and the main spar of the wing lower surface.

These act as cables do in the ultra-light to provide fabric support. At this point, before the covering, the mounting

block was glued into place. Figure 2.2 shows two

photographs of the uncovered wing frame.

The wing was finally covered with the nylon fabric.

Wood glue was used since it binds between the fabric filaments. The fabric covering was stretched tight in the gluing process simulating that of the ultra light. An abundance of glue was used to provide a good rib-fabric bond since the fabric must carry the entire wing loading.

Overall, the model is an excellent 1/5 scale representative of the full size ultra-light wing.

1.0 2.0 3.0 4.0 5.o 6.0 Zo 8.0 9.0 I0._ I_I i I I I I I I I I --I .25C .50C .'75C I .9 I I I -I FIGURE 2.1 WING CROSS SECTION k: I, FIGURE 2.2 UNCOVERED WING FRAME 3. WIND TUNNEL TESTING METHOD Once construction of the 1/5 scale ultra-light wing was finished, the wing was mounted in the small subsonic wind tunnel in the basement of Learned hall. Figure 3.1 shows a 3/4 view of the wing in the test section. Figure 3.2 shows a front view of the wing in the test section from inside the wind tunnel.

Raw data from the tunnel testing appears in Appendix A. The following data is included in the upper portion of these data sheets: * Wind tunnel static pressure: P S * Wind tunnel total pressure: P T * Ambient temperature * Atmospheric pressure Once the tunnel is up to testing velocity lift, drag, and pitching moment were read off of a percent of range scale and recorded for a range of attack angles. The wing angle of attack is varied during the test run.

The basic purpose of the testing was to determine the aerodynamic data of the wing and compare it with regular airfoil data. During the testing it became apparent that the airfoil section shape, and thus aerodynamic data, depends highly on the fabric flexure. The fabric flexure is in turn determined by the airspeed and angle of attack of the wing. These compounding factors cannot be completely assessed individually but they are considered in explaining the aerodynamic data. Wing sections will be shown at varying angles of attack.

Eight individual tunnel test runs were performed for the ultra-light wing model. Tunnel speeds range in between 47 and 121 feet per second. Extreme caution was used in making certain that the wing would not receive loadings large enough to cause structural failure. This model is not designed to sustain lift or drag loadings over thirty pounds because of its light construction. This limit maximum limit loading on the model wing is, by the way, equivalent to fifteen pounds per square foot--the loading normally sustained by light all metal aircraft!

The aerodynamic forces carried through the wing are sensed by a force table beneath the test section of the wind tunnel. Strain gauges in the force table translate lift drag and pitching moment forces into electrical voltages through a Wheatstone bridge circuit. The data is finally displayed on a control panel which has selector knobs for lift, drag, pitching moment and scale factor and a percent of range scale for voltage reading. The scale factor knob has magnitude selections o£ 50, i00, 200, 500, i000 and 2000. The scale factor is read in percent o£ range which varies between -.5 and +.5. The scale factor and voltage are read for lift, drag and pitching moment for each angle of attack tested per trial.

Test runs #i and #2 are considered inconclusive evidence. It was discovered through these tests that varying the scale factor caused significant error because only one scale factor can be zeroed to at a time. For the remaining tests the percent of range scale was zeroed to a certain scale factor, which was used for the entire test.

ORIGINAL" PAGE BLACK AND WHITE PHOTOGRAPH i w FIGURE 3.! WING MOUNTED IN THE TEST SECTION: 3/4 VIEW FIGURE 3.2 WING MOUNTED IN THE WIND TUNNEL m FRONT VIEW I]'ab_DOlOHcl 311HM any >10¥'1_] 4. TEST TRIALS Eight different testing runs were recorded. Trial numbers 1 and 2 are inconclusive but served to demonstrate a more accurate method of testing; picking one scale factor and using it for the entire test run. The remaining tests all provide meaningful data. These tests were run at different wind tunnel velocities, which were selected as to maintain a useful range of data.

Test #3: The scale factor of this particular test was set at 2000. The tunnel velocity was incremented until the maximum drag reading (at 20 degrees angle of attack) read the maximum of .5 on the scale. The wing angle of attack was varied from +20 degrees to -12 degrees by increments of 2 degrees. Lift and drag data was recorded for this trial.

Noted are that buffeting occurred at -12 degrees and beyond +8 degrees. This was seen to be the case for the remaining trials.

Test #4: This test was run to obtain a complete record of lift, drag and pitching moment data. With the scale factor • set at 2000, the tunnel velocity was stabilized so that the maximum pitching moment reading was -.5. This tunnel velocity is the maximum limit for complete lift drag and pitching moment data. This also means that the wing is oversized: the aerodynamic forces that the wing capable of are larger than those that can be supported by the balance table. This test was performed for an angle of attack range of +20 degrees to -12 degrees.

Test #5: This test is the first "high speed' trial of the wing model. "High speed' for this model is considered to be greater than I00 feet per second, which is the approximate tunnel velocity of this trial. The angle of attack range selected is +12 to -12 degrees. Again, structural constraints limited the maximum wing angle of attack. Lift and drag data only were recorded.

Test #6: This test is the second "high speed' trial. This test is very similar to test #5 except a larger wind tunnel velocity was used; approximately 122 feet per second. This is the maximum recommended tunnel velocity to be used for this wing. Because of the high speed, the variation of angle of attack was maintained between +8 and -8 degrees.

The main purpose of this test is to compare the lift and drag data of high speed trials to lower speed trials.

Test #7: This test is a duplication of test #4. The same approximate tunnel speeds were used and the same angle of attack range was used. The purpose of this test is to determine the the test replicability of this testing procedure by attempting to duplicate the results.

Test 18: This test is the "low speed' trial. The scale factor used for this test was I000. Again, the pltchlng moment reading was the limiting factor: the tunnel velocity was set such that the maximum pitching moment registered -.5 on the percent of scale range. Angle of attack for this trial was varied between 20 and -12 degrees.

ii

5. DATA ANALYSIS Of the eight wind tunnel test runs performed, six trials had meaningful data. These data for these six wind tunnel tests was processed and they are displayed in this chapter in the following figures: mm Section Lift Characteristics for the 1/5 Figure 5.1: Scale Ultra-Light Wlng Model Figure 5.2: Section Drag Characteristics for the 1/5 Scale Ultra-Light Wing Model Figure 5.3: Section Pitching Moment Characteristics for the 1/5 Scale Ultra-Light Wing Model Figure 5.4: Drag Polar Characteristics for the 1/5 Scale Ultra-Light Wing Model The raw wind tunnel data is listed in Appendix A. The equations which relate percent of range and scale factor readings into actual lift, drag and pitching moment forces were obtained from an AE 245 laboratory exercise. These equations and along with lift, drag and pitching moment equations were written into a basic program to speed up the data analysis program. The final output of this program gives the tunnel speed, Reynold's number and the wing lift coefficient, drag coefficient and pitching moment coefficient. The output listing for runs 3-8 are in Appendix A.

The lift coefficient-angle of attack curve is seen in Figure 5.1. Data from trials number 6 and 7 were plotted.

Although these two trials were performed at 122 and 67 feet k.

m per second respectively, the data compares very well. The lift coefficients at higher angles of attack for the high speed case lies below those for the low speed case. This most likely indicates that wing section deformation at higher speeds lowers the wing's lift producing efficiency.

An unusual characteristic of this lift curve is that there appears to be two different and distinct lift curve slopes.

Between -4 and +2 degrees angle of attack the lift curve slope is roughly 7.6 per radian. Between +6 and 16 degrees angle of attack the lift curve slope drastically drops to 1.8 per radian. This indicates that this wing section does not generate much incremental lift coefficient at high angles of attack. Also evident is that lift coefficient is very sensitive to angle of attack change at small angles of attack. Another interesting characteristic of this wing section is the high lift at zero angle of attack. The angle of zero lift is approximately -5 degrees. Obviously this wing section generates a relatively large margin of positive lift at small negative angles of attack.

The drag coefficient-angle of attack curve is seen in Figure 5.2. Data for this plot was taken from test run #3.

Minimum drag for this wing section occurs between -4 and -2 degrees angle of attack. It should be clarified that this drag is for the entire model and support mount! No tare runs were performed due to time restrictions. Since most o£ the data runs were taken at low speeds and since the model is relatively large this wont create a significant error.

The drag bucket in this curve also seems fairly symmetrical between -12 and +8 degrees angle of attack. One interesting characteristic of this curve is the intense amplification of drag at large angles of attack. The drag reading at 20 degrees is a factor of 24 times larger than the drag reading at -2 degrees. This "amplification factor" in ordinary wings is usually not as large. This is perhaps caused by the wing fabric pocketing at high angles of attack and further destroying the air flow. Another possible theory is derived from the fact that the wing frontal area to tunnel test section area ratio is small at large angles of attack.

The airflow is constrained to this area, and normal flow probably cannot be achieved, and the air pressure is probably increased, thus the drag is increased. A third possibility of excess drag at high angles of attack could be due to the model flutter at these angles. The model was seen to flutter at -12 degrees and above +8 degrees angle of attack. Drag is known to increase with flutter.

The pitching moment-angle of attack curve is seen in Figure 5.3. Data for this plot was taken from test rum #7.

It should be reminded that this pitching moment data is about the main model support mount which is located at .18c i of the wing. Pitching moment data is usually referenced at .25c or the aerodynamic center. A simple transformation can be performed to shift the pitching moment coefficient to this point but time constraints limited this process. Never the less, the slope and shape of the pitching moment curve is accurate and can be commented on. The slope of a pitching moment-angle of attack curve should ideally be a straight line. The pitching moment curve plotted indicates three different upwardly sloping "troughs". The angle of attack breaks between the three troughs are 0 degrees and 14 degrees. It is uncertain what causes these distinct breaks, but again it is assumed to be the fabric flexure.

Apparently fabric flexure change at 0 and 14 degrees angle of attack is very critical to pitching moment characteristics of the wing.

The lift coefficient-drag coefficient curve is seen in Figure 5.4. Data for the two curves were taken from test runs #6 and #8, the high speed and low speed trials, respectively. The slope of this curve indicates the maximum lift to drag ratio of the model. For the low speed case (run #8) the maximum lift to drag ratio is 12. The maximum lift to drag ratio for the high speed case (run #6) is 7.

This indicates that the lift to drag ratio is reduced at h]gher speeds. This is probably because the fabric flexure at higher speeds is more warped and less conducive to lift.

l::i=:.LT. !7!:7-i-:ii!::ii:: .iii;_.i{!i;il. T::f:-L--.Y:i-ii:iiiH:--==:{_iii{:::ii£7_'i'-:.:ii:i;:-!ii::L:-Z_-'--ii. : !: : !!i::2L71 Z ..... -L..:+.i ....... , ........... ". ============================================================================= __=L----T_------:---=_=-:-:I_::= _:_Z_ .: = = : : : :: ' : Z=_.4 __T " = = ----'------ _ - :, " ":, ;,

-: 97__-.--:---U--_ _-_ ,,- : 1- i .... .;__ ,: i , ' ,,,, , ---,- "-_

7--- " __i. :___._l==. _ .... !__-_x_.'__._--__-- ........ L ;-. ........... _ .... .'-"-"-_I --- i I' I _ '

_-7;i---:_--_,--N(_L---E _-OE-I; AT:TA:C K :_--:_---_!:_-:-_ :DE -G-6_E-E.---$--

°- :- CALC FIGURE 5.1 SECTION LIFT CHAR-

AIE 592

ACTERISTICS FOR THE 1/5 SCALE ;CM_C_ ULTRA-LIGHT WING MODEL APPO APPO UNIVERSITY OF KANSAS _*_ 14 _-_!_>i _ii_...-i..-..._!i_!_!iii_!iii_iii_i_!!:..4.i_ii!iiii!ii_iiii_!!i_i_.::..i!i_ i i:iiii_iii_ _i_!i!.:!-i!!il DATE CALC FIGURE 5.2 SECTION DRAG CHAR- /(j. _.% _v,s_o

AE 592

ACTERISTICS FOR THE 1/5 SCALE CHECK ULTRA-LIGHT WING MODEL APPO APPO UNIVERSITY OF KANSAS p*_E 15 ......... " ° " " . .......... i .....

, - :...::: :-- : .... _..._ -_ _.<_-_RUN _'7 -:-: :;=-:==:: .. . ..... __- - ! . ..... • ; "-'-: ,:. ::L:.-:.-:. ":';..::: ":;.-" ---4 • i • _ : .. . ... :._...!i:: ':':_ :'_. i :: !"7-: --i:---:":! i.... i .

[:i---._.9 - 1-:,-. " - : : : ....!-:---[:.-r-.iil]__:i-i:_ i:._::_ii !:-: ::-::::_ - i_i_-_-_:fl_--- '

P_ :_::_-:_ -- - .... -:::-:---- -:-: --:-I ....::-:. .... :i=.:f:-_i:>. i .:.i

.--:_ :.-= :::.":- . --_-- L: :-: -- "2-- " .T_':: ...... : "'_-- "---=:_ • "- ::::::::::::::::::::::: -..:!!:___::-.: .... __:_.L_--_:I.._. : -{::-:-::--: ....

._::i-:+-]--i::4::_--: :: _:-::-.:::ii-:-__.i....i :-:---:.::" ::-:---i---==.._i.i ..... :7::--.!--::: I:::: := ::=--'-::=::.:-: :, -.--.;-.- -'--:. '::::'.:.i:J:;..:-'::-- i - .- ::= "_=- - -- E--.:::_:::"--_" -::- i " : t" . ". :"" . _ : 2 ": .::::2-:? .... ::{.2"2:: ".'.:_: :'" : :7::. :'.:'-:.L':7:'-: • _ " ' "_ " :4--, --U .... ' " " " "" "_:': " _ _ -: :_: ._._:i_:.i'i-_i-!._;_::::i:_::,"--:X:f: ..:-.-.?!_::-.:::: :i::.:_ :::i::_-::":i:_:: ::--'..-::---ff:::-._:_!i-:-:_.__}_-:--L--#--:!::i:_.

_-q.:.:::::.:.;.. ==::_:.. -.:.._.::-_--:.:...-.-=- --j ....... :.;... ..... _-:-:::..... :=...---..--- -- ..:.... - --- ----:F__--_.___== ::F:-:: --.-=.=:::=7-- --:=_=---( .-=-_-'_. 2=:--: .._.- ,,--T-.-T----.--T - .- -:. - CALC /_ , _ _ REVISED OATE FIGURE 5.3 SECTION PITCHING

AE 5':32

CHECK MOMENT CHARACTERISTICS FOR THE 1/5 SCALE _LTRA-LIGHT WING MODEL APPD A_PO

UNIVERSITY OF KANSAS --_E 16

w ,,v,s,o OAT_I FIGURE 5.4 DRAG POLAR CHAR- c,Lc ,,4'I. _,,,c_

AE 992

ACTERISTICS FOR THE 1/5 SCALE CHECK ULTRA-LIGHT WING MODEL A@PD APPD UNIVERSITY OF KANSAS PAGE

I

6. WING FABRIC FLEXURE The topic of wing fabric flexure was mentioned often in the previous chapter. The section shape of an ultra-light wing is highly variant to airspeed and angle of attack.

Airspeed tends to vary the magnitude of the fabric flexure.

Angle of attack varies the location and direction (inwards or outwards) of fabric flexure. The fabric flexure for five different angle of attack settings were sketched in Figures 6.1 to 6.5. The many different (and odd !) airfoil shapes should be noticed for the range of attack angle settings.

These figures show generalized airfoil shapes. The wing model was constructed with wire cross braces on the lower surface between the leading edge and main spar for fabric support (as stated in the construction chapter) which obviously are reflected in the lower surface fabric flexure shape. These helped to limit the fabric deflection in that particular area, but the exact shape they create is not determined in the figures.

-I0 degrees angle of attack: This setting is shown in Figure 6.1. The upper surface leading edge and trailing edge are indented signifying a pressure force exerted downward on the wing. The entire lower surface is bubbled outwards, again displaying a downwards pressure force.

w There is a very interesting bubble in t_he fabric on the upper surface of the wing at about .25c. This perhaps is the only upwards pressure force on the wing, and serves to form a very unusual airfoil surface.

-6 degrees angle of attack: This setting is shown in Figure 6.2. The upper surface leading edge and trailing edge are indented, and so is the lower surface trailing edge. These indented surfaces are all handling inward pressure forces. The surfaces bubbling outward (experiencing outward pressure forces) lie on the middle upper surface and the lower leading surface of the wing.

0 degrees angle of attack: This setting is shown in Figure 6.3. The upper surface leading edge and entire lower surface of the wing are experiencing inward pressure forces. The remaining upper surface is bubbled outward and is experiencing lift.

6 degrees angle of attack: This setting is shown in Figure 6.4. It is virtually identical to the setting of zero degrees in Figure X.4. The only difference is that the upper surface fabric bubbling is more marked.

20 degrees angle of attack: This setting is shown in Figure 6.5. This is quite similar to the previous two settings (0 and 6 degrees), however the upper surface leading edge and lower surface fabric deflection is more marked, and the upper surface bubble is shifted more aft.

/

x _2 O_ _0 ,-I O i

I

/

I

\

\

\

\

x U _0 ,,<Z

/

=:,<

/

Z_ O_

/

r..)_

/

w,D 0 I

I

Z _[.-,

I

i.D

\

\

2O

/

/

D x

I

/

U

/

< .4

/

< 0 E-.

/

E-.

E- <

/

D

/

/

< 0 u_

I

I

me.

/ I

B_ r_ ! , o _D 2_

I I

E-.

I I

I I

I I

I

D

I

I

\ I

\ I

\ I

x I.-4 cZ _O _fJ Z_ _0 O_ CfJ Z _D

\

\

\

\

/

U I,--I ,,.,.t.

r..j ,,_ [.-, ,.1 t..O rt.

,.I- r,.O o .4- l,,,,-4

\

\

\

\

w 7. RECOMMENDATIONS AND CONCLUBIONS This project is an initial attempt to provide aerodynamic data for an ultra-light wing. Conclusive and fairly accurate lift, drag and pitching moment data were recorded and analyzed for the wing model. Some of the important findings are: I) The lift coefficient-angle of attack curve indicated the presence of two entirely different lift curve slopes at different angles of attack.

2) The change in drag between small and large angles of attack is quite marked.

3) There occur two distinct break points on the pitching moment coefficient-angle if attack curve, indicating a particular sensitivity at these two angles of attack.

4) Lift to drag ratios for this model are 12 at low speeds and 7 at high speeds.

5) Aerodynamic data for an ultra-light wing is a function of the fabric flexure, which in turn is directly related to angle of attack and airspeed.

There are range of other tests that could be performed with this wing model. Hopefully a structural failure test will not be one of them. Ideas for future experiments with this wing may include: I) Building a rigid model of the ultra-light wing to provide base data so that a more accurate study of the effects of fabric flexure can be studied.

2) Re-doing the drag data and taking drag tare data.

3) Calculating the pitching moment about a more useful reference point such as 0.25c.

4) Performing this testing in a different wind tunnel that can register the maximum forces endured by the wing.

Overall this was a very enjoyable project and it is encouraged that other students use this wing in individual or group testing--such as an AE 245 laboratory exercise.

Appendix A: Data Analysis

c,_r_ : 12 /I'_ /_s

AS

59Z

{,,_, t,'u 6"

- 73. $°6

P,.,_ -- Zq.zg

"_ #0 " z£,3B 7"E.S T" , O<.

DRAG

_.I='. r?. G.An t/u -C

_00 -33.3 I000 5"OO

qqqzS _IZ ° _ '-I.0

50o Z_.5 500

-I0 °

3_, 5 ZOo

qq'1 (,8 Zq. 5 500

2-00 2-0.6 "" Z_.o 500

qq'_B8 DO0

I0o 0o006 16.5 I o00

500 + Iq,6 500

oo028

-Z ° 5o0 SoO 2.000

3q,_ 1"7, o

0 °

ooo _8 Io00

Zs. "5 500 Zoo o

Zo ,8 O00G _, +2" I0O0 3o. 5 Zooo 5OO -5o ,_ O0 OBB 50o _o._ Z ooo

+1"I ° 3a. ,5

10o0 ZZ.Z

+G" I000 38.0 I000

O0 I0_ Zooo

i

Oo 128 +8 ° IOOa I000 Z'7,Z _0.0 + I0 ° O0 iN8 _z.o /-4_. o Iooo I000 IoOO +.IZ ° 3_. o O0 168 /4_.0 I0oo I0o0 +lh*

ooj g_ [0oo

+16 °

I000 _ooo Z5.5

ooZo_ z4_. o

2-ooo

,18 ° lOOO 50. Z

OoZgS

56.0 Zoao I0O0 S'I. O oo Z_g .I-2.o ° 59Z 5P_C IaC_ m_ 0 &-._z ;- , {/5 3C_tL-S, t2bT_A- - L-/T_ 7-6_ P = 73,8°F Zq.Z4 Pa = 7-w/_& _ 3q -- 29.50 Pr - 7-ue_ J* HO ,, 25./2_

o<. LiFT

5. F. fZC-.4 0 t/,JG" ooz_ 8 ,2_0° 1000 33,0 +18 ° I ooo 3/.5 +16 ° 31.8 /ooo 3D. 6 looo *lZ" Iooo 2_.5 +/O ° I 0oo 2-.._. S I ooo Z_ .o -+_,° I 0oo Z5.5 _Z. B 4.zi° Io00 4-.2" 5o0 36.0 0 ° 5oo Z3.3 5"oo 3g .5 Zoo 3z_. 0 /oo _ 8o - Z8.5 Zoo -54.0 -/0 ° -IZ ° _ /.,,7.5 AS DA_ :

S_Z

_z /t_ 188

Z S_o '7(, .0 P._,.4 -- _-_ . Z'7 HO " 25.5_ "T'-E._ T .

o<.

2.00 o 4 20 ° .320 _ooo ,320 .. IG ° 2-Do 0 .320 I_ ° ZOoo

.303

_. 12 ° ZOO0

.2q8

+ I0 ° ZOOO _ooo 2-_3 Zooo

,z_8

+If" Zooo ,__15 _2 ° gooo ,/83 0 °

Zooo ./_5

_ _o 2ooo _Ifo ?_.ooo .O[o ZDoo -,O._o

Zooo

-.o85

- I0 _ ZOOO .._IO- _17_ ° Zooo -IZ5

A_

,42 /_//_,/_ .5"9?

_'_- P_' _" {_' _'_ + 71-I 7-_,_p = T6.Z. o D4,,',_ -- -?'9..3_ = T_,_ == 39 = E"7. ts3 - TV/_G.

_,- 40 ,, Es.;c_ _<.

L..I F T

_. F. _4 0 t,/VG" 4 20 ° _.OOo 21.5' 4-I_° Zo.'_ ÷i6 ° 20.3 7_o.o _- I_ ° lq .0 +. I0 ° I_,0 4--8" +9" Iq.8 /3.o _ O ° ro . o _.Z.

_Ll° Z.8- 0.?_ -I.2 - |O ° _.o -12" -5._ A-_ GA T-_ :

/z//7/89

AS

5_Z_ _P_GfAL _f.O.7"_z--..T" : O0 5o _ l,,u C 7-6,_P = 76, Z

_0 ,, ZS,q_

Z._57" S6r- _v 0<.

DRAG

P I TF_H I/lJ_; _ O_6.V ._

7_..0o0 1"tl.0 Z O0O IZ ° /15.0

3_. 5

q3,0 lo"

29,3

38.8 3_. 8 +_° 2q.3 32,5 ÷ lfl ° 18._ 2_'1.0 + 2 °

Ii,o

0 ° Zi.3

"1.3

/3.5

-ld° q.O

6.Z

-I.5 IZ.o

I&,o

-6" -5.5

Zo.o

- I0"

-&3

Z&.5

-lZ"

-_o.5

AS _PEC :A_ p£ _ ,.7"_C.7 ; S_Z (_jtM_ W- (u'o)E • C)Y" M J'x_ .T/_65 ff6C0_ 9 P#,_ -- D_. 30" ;IT" ze¢o A,uc,_ o_ (h'rterCK.

/Vv,_ "

L_iTT

O4,.

DRAG ._._.

_o0o _o. o Z_oo 38.5 _ _2.Z &° _2-3 Z&. o £° 20.8 0 o ZB.5 16.5 __ l&'7 -Z.f" 6.0 /3. fl __ -33 18.o __° Z5. 8 - _^ _r _DDO _ P_ = Tu&_ _ 3_ = Za.38

t_lTT

o<.

_. F. /_ C_J_ +.IS ° 18,_ _. lf_ ° ,-IH ° I_.o , 12 ° 1'7, o ._ I0 ° IG, Z 15.5 __ (_°

+ l-i _ _.D

11.5 0 ° B.5 _Z ° i-_.-1 1,5 -OA -I.5 -10" -_."I -1'-2" -&.l C4T_; / z./ J-;/_9 AS 59Z k 6 3c,c-_ u_7_ - z.-1r_ 7"E_ P = (3 Y /4 j _ 4"a_C : Pa = 7-u&_ _ 5q - Z_,.58 7--__5T Pr -- 7-w_ _ NO - Zg.Ol _2_:o _c/_._ : 2,S_o6o z_ /ooo

L. IFT

o<.

DI_AG P 1TEa laJC_ ka 0 _&:J 7" R._a 0 _m __. _-.

I o'._o _3 .5 /00o -5"o.o / ooo go.5 ,_ 20 ° 2'7.5 Jq.o - _.5 +18 ° 2e.o tq.O - Lt#.O + Ig' + Iq ° - 35.5 18.5 + IZ ° 18.o - 3q. o la.5 4- IO" IqO _Z,a -33.0 t0.o +8 ° Ig.6 -3Z.o 8, o -300 15._ 5:7 &.O -Z&.7 -z_. 8 3._ IZ.'Z- ,Z ° i_.5 Z.5 ,0 ° 5.'7 2.5 Z,?- -I_.7_ Z.'7 -%7 -O.Z _(o ° &?_ -5.5 -Z._ -I.o _ I0 ° 7._ -_.0 _-8.0 ,IZ ° I1.0 -6.5 J • _r 3. _/_(_,,_ /,,t_¢- _-..¢c,, (c.A_ _ pat _"1 y /Z, I-- I-- I-- --- _ laid i_ lIMl_ E -'G,-,_ -?pee3 " p,, e,I elf P ) [ _.z_'_ ) J - F_r--_ _'. _,._'c_.l_ _F /_/v,,_ 6 e/'v I0 -ra "F

.3. _ O & ,, ,_-."z_ 4. 5. _ & ,, .r( )

f_.

e v ..

_ l

P--D---

_s

) RHO=() 2_-I L=O 30 0=0 40 TEMP=O 50 F'RESS=O 6(:) PM=O 7(") MU=(') 8(:) V=O 90 CDTARE = 0 - 100 INPUT "TRIAL RUN NUMBER =";NUM llO INPUT "STATIC F'RESSUPE =";PS 120 INPUT "TOTAL F'RESSURE =";PT 130 INPUT "PRESSUPE IN INCHES HG =";P 14(:) INPUT "TEMPERATURE IN DEGREES FARENHEIT =";T 15(:I INPUT "WING CHORD IN FEET =";C 16_:I INPUT "WING SPAN IN FEET =";B -170 S = B*C 175 PRINT "WING AREA =";S 18(:) F'RESS = P-70.722 185 PRINT "PRESSURE =";PRESS 188 PRINT "PRESSURE ="PRESS 190 TEMP = T+459.6 195 PRINT "TEMPERATURE =";TEMP --2(:)0 RHO = PRESS/(1716*TEMP) 21(:) V = ((PS-PT)*3.2174/RHO)^.5 ZO Q = .5_RHO*(V'"2) .."25 PRINT "DYNAMIC PRESSURE =";Q 23(:) MU = ((5.48-10 ...... 10)*T)+(3.408"10^-7) =3F. PRINT "'.2ISROSITY = ;MU 240 RN = RHO*V*C:/MU --'25(:) LPRINT "WIND TUNNEL RUN NUMBER ";NUM 260 LPRINT "TUNNEL VELOCITY IN FT/S =";V 27r) LPRINT "REYNOLDS NUMBER =";RN It CD I CM 280 LPRINT " ALPHA 1 CL I II 290 LPR INT ..........

3r)(:} INPUT "SF:ALE FACTOR =" ; SF 305 INPUT "ANGLE OF ATTACK ="; ALPHA 310 INPUT "LIFT READING =";LREAD 3?0 L = LREAD*SF/27 "'" CL = L/(Q*S) _(_) 34(_') INPUT "DRAG READING =";DREAD 35(:) D = DPEAD*SF/IC)C) 36(_) CD = (D/(Q*S))-CDTARE _ 37('.) INPUT "PITCHING MOMENT READING =";PMREAD 38(.) PM = PMREAD*SF/348 390 r:M = F'M/(Q*S*C:) 4(')0 LPRINT TAB(4) ALPHA TAB(15) CL TAB(30) CD TAB(47) CM 410 GOTO 305 _ A -IO WIND TUNNEL RUN NUMBER 3 TUNNEL VELOCITY IN FT/S = 86.4358 REYNOLDS NUMBER = 432672.'9 CD CM ALPHA I CL I 20 1.410434 .5950268 0 18 1.410434 .5093429 16 1.410434 .4046182 14 1.335505 .3510658 0 12 1.313466 .3034637 0 10 1.234129 .24'39112 0 8 1.115124 .1820782 6 1.04901 .1606572 0 4 .9476352 .1130551 c) 2 .8065918 7.14c)321E-02 0 .5509507 4.165187E-02 -2 ._.ouai_ 2.380107E-02 -4 4. 407606E-02 2. 380107E-02 0 -6 -. 2203803 4. 760214E-02 -8 -._ _o_ 8 33c)375E-02 -10 -.4848367 .1249556 -12 -.55c)9507 .1844583 c) WIND TUNNEL RUN NUMBER 4 TUNNEL VELOCITY IN FT/S = 61.3908 --REYNOLDS NUMBER = 307416.8 CD CM ALPHA _ CL I -.3985074 20 1.877324 .7732828 -.3690178 18 1.772543 .6365438 -.3427163 16 1.772543 .5540288 -.3108358 14 1.746348 .4361504 12 1.65903 .3772112 -.2972865 -.2869253 I0 1.571713 .3253446 8 1.4843'96 .2781932 -.278'9551 -.263O149 6 1.423274 .2239691 -.2391044 4 1.292298 .1'356783 2 1.135126 .1249512 -.2072238 0 .873173'9 8.958764E-02 -.163388 -2 .5413678 .0777998 -.1378836 -.1155671 -4 .2444887 .0777998 -8.767162E-02 -6 1.746348E-02 9.430279E-02 -4.622686E-02 -8 -.1047809 .136739 -7.970147E-03 -10 -.34'92696 .1886056 6.774625E-02 -12 -.5064409 .2829084

_IND TUNNEL RUN NUMBER 5

TUNNEL VELOCITY IN FT/S = 104.4355

REYNOLDSNUMBER= 522'364.7

ALPHA : CL : CD _ CM 12 1.357762 .3340095 0 10 1.297417 .281c)568 0 8 1. 17c)693 .2386946 0 6 1. 080175 .1979617 0 4 .9806059 .1531556 0 2 .8750022 .1205693 0 0 _ .6426741 .08'36123 0 --2 .4073286 7.576313E-02 0 -.4 .187c)695 7.331915E-02 0 -6 -4.525874E-02 9.775888E-02 0 -8 -.1659487 .1303452 0 ._ 2 10 -. 2504317 .162'3315 0 -12 -. 3107767 .2158842 0 IND TUNNEL RUN NUMBER 6 0UNNEL VELOCITY IN FT/S = 121.847'3 REYNOLDS NUMBER = 610158.1 ALPHA : CL : CD : CM 8 1.108258 .2322023 0 6 1.03068 .192704 0 4 .9375865 .1555995 0 2 .8Cr30285 .1244796 0 0 .6272741 9.754889E-02 0 -2 .3701582 6.882284E-02 0 -- 2.329E_ -4 .132991 8.318586E-02 0 -6 -7.314504E-02 .1077227 0 -8 -.2504664 .1544025 0

WIND TUNNEL RUN NUMBER 7

TUNNEL VELOCITY IN FT/S = 66.6¢]797

REYNOLDSNUMBER= 333876

ALPHA _ CL ', CD CM

20 1.482431 .6364076 -.3382829

18 ° 1.3'33485 .5403461 -.3179859

16 1.3'33485 .440282 -.29(YB233 -.2638607 14 1.334188 .3582294 12 1.260066 .3061961 - 25371o0 10 1.200769 .2561641 -.2455934 8 1.148884 .2261448 -.2367'38 6 1.074763 .1761128 -.2252'364 4 .9858166 .1360872 -.2049994 2 .8523978 9.406024E-02 -.17929 0 .6300331 6.804358E-02 -.1420788 -.1204287 -2 .3483713 5.203333E-02 ---'_ 1111023 --, , . -- _ -4 .1111823 3.202051E-02 -.1008083 -6 -.0667094 7.404743E-02 -7.645194E-02 -8 -.1111823 .1080692 -4.465334E-02 -10 -.3483713 .1561 -1.353132E-02 -12 -.4521415 .2181397 4.262365E-02 -- WIND TUNNEL RUN NUMBER 8 TUNNEL VELOCITY IN FT/S = 47.32559 REYNOLDS NUMBER = 237222.1 ALPHA _ CL CD CM 20 1.504973 .6640236 -.3350507 18 1.3'34853 .545094 -.3048962 16 1.394853 .4757184 -.2948446 14 1.358147 .3429137 -.237886 12 1.32144 .2874132 -.2278345 I0 1.248027 .2457879 -.2211335 8 -1.233344 .I'_8216 -.2144325 6 1.167272 .1585728 -.2010304 4 1.(}27787 .1129831 -.1789171 2 .8956426 7.730424E-02 -.1460821 0 .624¢}133 .049554 -.1172677 -2 .418456 .049554 -.0971647 -4 .1835333 4.360752E-02 -8.175237E-02 -6 -1.468267E-02 5.351832E-02 -.6499984 -8 -.1908747 8.325072E-02 -3.685558E-02 -10 -.3670667 .1466798 -6.701014E-03 -12 -.4771867 .2180376 5.360811E-02

- -13

V. STATIC TEST OF AN ULTRALIGHT AIRPLANE Reprinted from J. Aircraft, Vol. 25, No. 1, January 1988 Howard W. Smith Professor Department of Aerospace Engineering University of Kansas Partially supported by NASA Langley Research Center Grant #NAG 1-345 VOL. 25, NO. 1, JANUARY 1988 J. AIRCRAFT 37

Static Test of an Ultralight Airplane

Howard W. Smith* University of Kansas, Lawrence, Kansas This paper describes the work neeesmuT to perform the static test of an aitralight airplane. A steel reaction gantry, loading whiffletree, hydraulic actuation system, and Instrumentation systems were designed. Load and streaa analyses were performed on the airplane and on the newly designed gantry and whiffletrees. Load cell culibraUon and pressure indicator ¢ullbratlom procedur_ are described. A description of the strain and deflection mea.mremant system is Included. The engine, propeller, fuel, and pilot were removed and replaced with ma._am to fulfill center-of-gravity requlramenta prior to tesUag. Data obtained to date are compared to the analytical predictions.

Nomendature Analysis CL = wing lift coefficient Design Criteria d = displacement, mm In the early days, an airplane had to be able to carry the F_, = ultimate compression stress, ksi limit load without permanent deformation and the ultimate h = altitude, ft load for 3 s passing the static test sequence was a time of joy M_ = wing bending moment, N-m and celebration for the structures engineers. Nowadays, air- n = limit load factor craft are governed by much more rigorous specifications.

RN = nose wheel reaction, lb The static strength requirement has been retained, but is now RL = left main wheel reaction, lb only one element of a much larger array of specifications RR = right main wheel reaction, lb under a comprehensive umbrella known as the structural S ffi wing area, ft z integrity program. Among the factors included are: corrosion, V ffi airplane speed, ft/s durability, damage tolerance, and flutter. Aircraft that are to W0 = empty weight, lb be certified prior to use must meet or exceed specifications.

War = basic flight design weight, lb These requirements are specified in either Federal Aviation Regulations or Military Specifications and the "meet or exceed" phrase is satisfied by analysis or by test or both.

A set of design guidelines for an utralight has been published by the Powered Ultralight Manufacturers Associa- Introduction tion (PUMA). 4 However, there are no specifications govern- S the service life of the fleet of ultralight vehicles ing the structural integrity of an ultralight airplane. For this increases, the number of fatal accidents is expected to analysis, the ultralight was treated as though it were a normal increase as well. Several cases have been documented by the category general aviation airplane governed by FAR-23. All National Transportation Safety Board_ in which the integrity related Mil-Specs and Mil-Standards were invoked as well.

of the structure was questioned. When fimilarities between It should be noted that student interest in this research cases occur, it is logical to formulate a plan to investigate the project was very high. One student elected to write a report on basic behavior of a typical vehicle.

a structural integrity program for ultralights, s probably the The opportunity to formulate a plan presented itself in early only one of its kind in existence.

1985. Research on the aerodynamics and flight characteristics of an Airiness Sunburst "C" was drawing to a close and a master's thesis by Blacklock 2 was published. Consequently, a full-scale ultralight airplane was available for further reseaxch.

A proposal was written and presented to the NASA.Langley Research Center. The primary goal of this proposal was to perform a structural test to destruction of an ultralight airplane.

The structural floor and the ultralight airplane specimen are shown in Fig. 1. To perform a static test, a steel gantry and its sway bracing was designed. 3 Similarly, the upper and lower •'hiffletrees were designed and integrated with the loading de- vice. Finally, the strain and deflection systems were designed.

This paper describes the details of the work accomplished.

Presented as Paper 86-2600 at the A/AA General Aviation Technology Meeting, Anaheim, CA, Sept. 29-Oct. 1, 1986; received Oct. 28, 1986; revision received June 12, 1987. Copyright _ American Institute of Aeronautics and Astronautics, Inc., 1986. All rights reserved.

*Professor, Aerospace Engineering. Associate Fellow AIAA.

Fig. 1 Sunburst "C" ultraiight.

H. W. SMITH VOL. 25, NO. 1 Table I Lift distribution Table 2 Weight breakdown of test aircraft, Ib Structure Speed (maneuvering) 69 ft/s Altitude h 1000 ft Tube WG-I 5.31 Weight Wsl, 468 lb Wing skins 16.25 Cc (max) 1.48 Landing gear S 150.9 ft2 Wheel-nose 3.12 n (limit) 3.8 Main wheels and tires 10.90 Rear axle 7.01 Seat 8.7 I Powerplant Lift Distribution Engine and propeller 78.38 Muffler 5.70 Ordinarily, a structural test engineer begins with air load Propeller shaft 8.88 distributions as "known" values. Both spanwise and chord- Misc., each < 3 lb Remainder wise pressure distributions must be given beforehand to allow Wo Weight empty 277.48 determination of "patch" loads. For this ultralight, six Fuel 15.52 spanwise and two chordwise stations were selected to simulate Pilot ("Bellerophon") 175.00 the subsonic pressure distribution. In reality, the airfoil W_ Basic flight weight 468.00 behavior is unknown, since it is only sail cloth stretched over the front and rear spar tubes. During a maximum positive load factor condition, the airfoil is taut and has a particular set of \ ordinates. During any other flight condition, including \ inverted flight, the ordinates are variable.

\ Since an air load distribution was not available, one was 2.0 calculated using a quasivortex lattice method. This work was

\

done by a student who favored this method and the analysis was performed with ease. s'7 With this knowledge, patch loads could be determined. Those data were incorporated in the upper whiffletreedesign.The design maneuvering speed at a limit load factor of 3.8 was 69.0 ft/s. (See Table l.) The spanwise liftdistribution is shown in Fig. 2. The spanwise drag distribution was assumed to be negligible.

1.0 Dead Weights l.S The weight breakdown for our test condition is given in Table 2. The engine, propeller, shaft, and mounts were 0.5 removed and replaced with a mass whose magitude and center of mass were correctly located. The lower whiffletree mass was included to correct the Ig dead weight loads. Fuel was replaced with water of the correct weight. 0.O

t Our ultralight pilot, named Bellerophon, was constructed o_- 0.0 0.2 0.4 0.6 0.8 l.O

army coveralls, worn-out army boots, a cap, and a mask SPANWISE NON-DIMENSIONAL COORDINATE (Halloween) for cosmetic purposes. The cap was adorned with

i

a NASA logo_ Bellerophon's center of gravity was built up with concrete cylinders at the buttock and thigh locations. The Fig. 2 Wing spanwise lift coefficient.

remainder was constituted from plastic bags and Kaw River sand. Weighing and loading him into the aircraft required the assistance of four strong students. The upper whiffletree arrangement for the left-hand wing is t Overall airplane weight and center-of-gravity location was shown in Fig. 4.

checked and rechecked by actual weighings with three balance The lower whiffletree is a loading mechanism as well. A scales under the wheels. Results of the wcighings were: pair of steel straps connect at the engine mount holes and the RN= 11.49 lb, Rc = 127.0 lb, RR=133.2 Ib, for a total of U-straps bear directly on the fuselage case tubes. These 271.69 lb. (See Fig. 3.)

whiffletrees are commercial grade steel and are designated tiers 6 and 7. Tier 6 is adjacent to the fuselage and tier 7 (the Point Load Calculations lowest) connects to the 10,000 lb hydraulic actuator. A load cell is in series with the actuator. These linkages are bolted With many scientific developments, the creators of the directly to a floor fitting where they are reacted. The floor breakthrough cannot foresee the eventualapplications of their fitting, called the "alLigator," was specially designed for that work. Likewise, Joseph Fourier could not have known that his

!

purpose. It is located directly below the air load center-of- work with sinesand cosineswould be used tocalculate air load pre&sure vector P, shown in the lower whiffletree sketches pressures on an ultralight airplane nor could Fred Whipple (Figs. 5 and 6). All of the lower whiffletree members are made have known that his method would be used to approximate from standard AISC steel sections: rectangular tubing, tees, that air load.

and flat straps.

The upper whiffletreesare simple three-point beam pairs made from. ordinary 2 × 4 and 2 × 6 pieces of lumber. There Internal Loads Analysis are five :'tiersTof trees. The first isthe highestand the fifth the Iowes_.. The treesare connected with heavy-duty turnbuck- A stress analysis of the wing structure was performed using les. Tier l is connected to the steelgantry with a singlesteel the alt loads discussed above. Availability of the Polo finite-element method and its ease of use were the reasons for strap.Tier 5 isjust below the wing and isin directcontact with T its selection, s Results are given in DeAlmeida's report. 6 The the tubular spars. Plywood bearing platesare used to spread flying wire loads at the design limit load factor of n = 3.8 are: the load along the spars. Tiers I-3 are the spanwise trees,

t

while tiers 4 and 5 assure the chordwise center-of-pressure Forward inboard 44 Ib location. With no load in the actuator, the ultralight is Aft inboard 65 Ib suspended above the hangar floor in straight and levelflight.

)

l

JANUARY ]988 ULTRALIGHT AIRPLANE 39 ENGINE uP

i

I Fig. 3 Weight and eemer of F_H_.

I i

I

I

i Z(UP) I

_I

TIER 7 _p Fig. $ Lower whiffletree, left side view.

_.. A IRPLAJIE J ING X (FORWARD) ELASTIC AXIS lql&. 4 Upper whiffletree.

Forward outboard 222 Ib Aft outboard I45 Ib Wing bending moments Mx and spar displacements d are shown in Figs. 7 and 8.

Systems Design For this study, the test rig was divided into four independent systems. The design and assembly of each system is described below.

/

Hydmuflc System A 3000 psi hydraulic system was designed to apply the load.

An Allis-Chalmers 10,000 lb, 8 in. stroke actuator and a Prince hand pump were purchased from a surplus madxinery supplier. A pressure gage and short hydraulic lines were obtained from the same supplier. A schematic of the hydraulic Fig. 6 Lower whiffletree, rlar view.

system is shown in Fig. 9.

The Boeing Company supplied the hydraulic lines, a four-port Barksdale valve, and several hydraulic fittings. The eye shaft could then be gripped in test machine jaws. Excellent linearity was achieved. A calibration constant was determined 2 gal reservoir and hydraulic oil were purchased locally. These parts were assembled and the lin_ purged of air by two to be 82 lb per unit readout. 9 students. The system was tested during the two-by-four destruction test described below.

Defie_oa Measurement System Large deflections were measured with a sliding scalesystem.

Load Cell System In hazardous situations, a telescopeor transit was used. This A 5000 lb Baldwin-Lima-Hamilton load cell has been in the was the case when cable failureswere imm/.nent. When deflectionswere small (lessthan l in.), a dialindicator was Aero Department for a number of years. A pair of load cell "eyes" had to be purchased to match the spec/al internal used. Tip deflections of 3.70 in. limit were expected. The threads. The eyes have ! in. diameter self-allgning bearings. A sliding scale concept was proved during the wood bending pair of links connect to a smaller eye at each end. The smaller destructiontest, which was recorded on video tape.

p, 40 H. W. SMITH VOL. 25, NO. 1 ""I'"' l'"'l""i"" ._ 20 A a_

-

• _ Front Spar /I • Rear S

io2 4o

;x L Front SparJ 20 -60 _- -80_ Rear SparJ .lO01],,tl,,,,It,,,l,ltlltt lj ,,,I,,,m JJJ,,J,tt,J_,_ O.O 0.2 0.4 0.6 0.8 1.0 O.i 0.2 0.4 0.6 0.8 1.

SPANWISE NON-DIMENSIONAL COORDINATE SPANWISE NON-DIMENSIONAL COORDINATE Fig. 7 Wing limit bending moments.

Fig. g Wing limit deflections.

Strain Measurements System ALLIS-CRAIJERS All strain gages were single-element foil gages from Micro RESERV.

Measurements. A 10 channel switch and balance unit and a _L GAGE. HARSH--_ J- I578 \ strain readout unit were available from previous research. The ,, 3,ooe psi HOSE_XNSION strain gage terminal board was borrowed from the Aerospace Medical Research Laboratory. The resulting strain measure- OW i • NIC44 8REATNER-7 _ /i II_RESSURE _..1 e ,RBSU,E ment system design was proved during the tube tension IISUPPLV _ -- mC_ R ?N-HP-IO "_ . CO_RE$$I_ component tests described below. Data were taken with a PRINCE P_JRP -_ _F.1"UA TOR Vishay-Ellis switch and balance unit and strain indicator.

...... "-"T" c./ _ _ Ca. 9-6_1 Component Tests HANDLE JJ VALVE. 4-HAY -- " BARKSOALE Tube Compression LOW PRESSURE RETURN tAZ R3 AC3 Compression tests of the 6061-T6 tubes were run to verify Fig. 9 Hydraulic system.

the heat treat level. The ultimate stress in compression was: Fee (measured)=47.8 ksi and Fee (MIL-HDBK-5A)=42.0 2) Almost nothing is known about the behavior of an ksi.

nitralight structure under repeated loads. A durability and damage tolerance research program is highly recommended.

Wood Bending

F

Wood bending tests were performed on a pair of medium- Acknowledgments grade "S-P-F" lumber. The test simulated an upper whiffle- tree and was performed to spot check the modulus of rupture Many people freely volunteered to work on this project: Steve Waddell, Geoffrey Smith, Ron Schorr and Paul of "spruce-pine-flr," another unknown. Both the stress magnitude and the failure mode were missed. The modulus of Oeischlaeger. Thanks to the Caroline Wire and Rope rupture in bending, not to be confused with the civil Company who suppLied the cable and assembled the test specimens at no charge. This work was supported by NASA engineering design value, was estimated to be 9600 psi. The wood beam ensemble failed in horizontal shear and "prying" Langley Research Center under NASA Grant NAG 1-345 near the point of maximum moment. The magnitude was 85_ and the Aerospace Engineering Department of the University of Kansas.

of the predicted ultimate load• For this test, the load-deflec- tion curve was linear up to 50% of the failure load.

References Cable Tension "'Safety Study: Ultralight Vehicle Accidents," National Traa_ Cable testing was very interesting and informative. Four portafion Safety Board, Rept. NTSB/SS-85/01, Feb. 7, 1985.

assemblies of V, in. diameter, 7 x 19 aircraft cables were ZBIacklock, C. L. Jr., "Summary of the General Powerp[ant.

designed to represent the "flying wires"on the ultraiight. They Weight and Balance and Aerodynamic Characteristics of an UltraBght Aircraft," M.S. Thesis, University of Kansas, Lawrence, Aull. 1984.

were fitted with thimbles, grommets, tangs, and Nico-prets sSmith, H.W., "Design of Static Reaction Gantry for an Ultralight clamps. Failure load for the cable is estimated to be 1740 lb.

Airplane Destruction Test," AIAA Paper 85-4022, Oct. 14, 1983.

None of the cables carried more than 975 lb. All "failed" by "Airworthiness Standards for Powered Ultralight Vehicles,'" the cable sliding out of the Nico-press fitting. Cable testing is Powered Ultralight Manufacturers Association, Annandale, VA, Dec.

incomplete at this time. ALl cables will be fitted with double 9, 1983.

clamps and retested in an attempt to rupture the cable strands.

Vrurnipseed, Michael E., "Aircraft Structural Integrity Prolpram Special safety precautions have been taken to keep humans for Ultralights," University of Kansas, Lawrence, May 7, 1986.

out of a I00 in. cable whipping lethal radius drawn with each 'DeAlmeida, S.F.M., "Aerodynamic and Structural Analyses of an cable end as an arc center. Ultralight Aircraft," University of Kansas, Lawrence, May 6, 1986.

'Lan, C.T., "A Quasi Vortex Lattice Method in Thin Wing Theory," Journal of Aircraft, Vol. 11, 1974, p. 518.

'Lopez, L.A. et all., "Polo-Finite," University of Illinois, Urbarm, Recommendations 1985.

I) Unscathed portions of the ultralight, such as the wing 9Page, L., "Cable Testing for Ultralight Airplanes," University of tip,can be sawn off and used in futurewind-tunnel work. The Kansas, Lawrence, May 6, 1986.

two-dimensional lift and drag coefficients should be obtained t°(All) Engineering Drawings, University of Kansas, Lawrence, (125 from minimum to maximum CL. drawings total).

N93-29779

k VI. SELECTION AND STATIC CALIBRATION OF THE MARSH J1678 PRESSURE GAUGE Charles R. Oxendine Graduate Student Howard W. Smith Professor Department of Aerospace Engineering University of Kansas March 1986 Partially supported by NASA Langley Research Center Grant #NAG 1-345 TABLE OF CONTENT List of Tables and Figures .....................................

Summary ....................................................... I Diseusslon .................................................... I0 LI ST OF TABLE S TABLE NUMBER TITLE ASHCROFT DEAD WEIGHT LAB DATA II RICHARD GEBAUR HYDRAULIC LAB DATA CALCULATED DATA OF THE ASHCROFT TEST IIl IV CALCULATED DATA OF THE RICHARDAGEBAUR HYDRAULIC TEST MARS}{ GAUGE SELECTION GUIDE LI ST OF FIGURE S FIGURE NUMBER TITLE PHOTOGRAPH OF MARSH PRESSURE GRUGE PHOTOGRAPH OF THE AS}L'ROFT TESTER ASHCROFT CALIBRATION CURVE S RICHARD_GEBAUR CALIBRATION CURVE ASHCROFT GAUGE TESTER RICHARD GEBAUR GAUGE TESTER L.

SUMMARY During the experimental testing of the ultralight, it was determined that a pressure gauge would be required to monitor the simulated flight loads. After analyzing several factors, which are indicated in the dis- cussion section of this report, the Marsh J1678 pressure gauge appeared to be the prominent candidate for the task. However, prior to the final selection the Marsh pressure gauge was calibrated twice, using two dif- ferent techniques. As a result of the calibration, the Marsh gauge was selected as the appropriate measuring device during the structural test- ing of the ultraligh_.

J Although, there ar_ommerical pressure gauges available on the market that would have proven to be more efficient and accurate. However in ord_ er to obtain these characteristics in a gauge, one has to pay the price on the price tag, and this value is an exponential function of the degree of accuracy efficiency, precision, and many other features that may be designed into the gauge. After analyzing the extent of precision and ac- curacy that would be required, a more expensive gauge wouldn't have proven to be a financial benefit towards the outcome of the experiment.

Page I INTRODUCTION There are several manufactures that design and produce a large variety of measuring devices with specific capabilities that are predetermined for each %nstrument.

The-i_-are two primary objectives of this report. First, it will justify the logical deductions that lead to the selection of the Marsh J1678 pres- sure gauge as the measuring instrument to monitor the experimental loads that would be exerted on the structure of the ultralight at any given time.

Second, it will indicate the two different techniques that were used to calibrate the Marsh pressure gauge, and the margin of error thats associat_ ed with each reading as a result of each calibration.

Also, this report was written in partial fulfillment of course _qulre- ments in A.E. 592. This report is rated with a worth of 3/4 of a semester hour out of the two hours of_A.E_ 592.

page 2 MARSH PRESSURE GAUGE Figure I ASHCROFT TESTER Figure 2 ORIGINAL PAGE BI..ACK AND WHITE PHOTOGRAPH Calibration Process There were two calibration tests performed on the Marsh J1678 pressure gauge prior to its acceptance as an experimental measuring device. The first test was completed with an Ashcroft dead weight tester (model no. 1300, and serial no. 1788).

The following procedures were used during the test process and are illustrated in Figure (5) in Appendix (B).

1 The reservoir was filled with a light mineral oil.

2 - Value B was retracted, so that the compression cylinder could be filled with mineral oil from the resevoir.

3 - The Marsh pressure gauge was connected to the Ashcroft tester at point E.

4 - Value B was closed to prevent the mineral oil from escaping back into the reservoir.

5 - Value D was opened to expose the port of the pressure gauge to the mineral oil contained in the compression cylinder.

6 - Weights of desired increments were added to the platform of piston F.

7 - Value H was screwed until the piston floated freely approx- imately two inches above cylinder G.

8 - The platform was spun.

9 - A pressure reading was read from the pressure gauge.

After each incremental weight increase, the steps that followed the addition of weights were compiled. With the Ashcroft dead weight tester, the Marsh pressure gauge was calibrated up to 500 psi. Even though the tester had the ¢apabflity of calibrating a gauge above 500 psi, the accessories that were required to continue the calibration process were not available. The calibration data can be observed in Table I and Figure 3.

page 4 ASHCROFT DEAD WEIGHT TEST LAB RESULTS GAUGE READING CALIBRATED PRESSURE (PSl) (PSI) I.

.

o 485" .

TABLE I page 5 ...... y.-T+ Z-. - .-.- _+_ . + -: ] : +_T: • :} • __+.+ _--_ _ i . +__L.

. -.;.+ -LLT_ ..... _-- +'-:.!

.... "I :7Z: "i. --'_ _+-+J . -tL -_ .... , ----++ _+--_.+ "+_ :'Z .ZZT'- "+ Z L.++-:I+ +_-+ !.- + -- --o .Z +2{:':- .....

+--+ :ZHZ _ L___: _ :;f+ :q+fi "+-:+ 7 7 .:+_.; . + +-Z-. + ..... 5m

i ,iii +gl i - +++

_7"-+ : ++---+-+_+- L+'_+-_-++: • _ - : -+_'-+--_

___ __ +_- -- ...... --+._ +_

..+__-+ _+++_+_ :k-++-_ _+-. +:_.... -__ -+_

-_ 3m

-+++_+:-+ +: +!4! + u--+: +_+. _+:+ _ +++ + iil!}++

?

+- - -- _. .'_-+%_._:_+ -.+---+ _ _+:J.-_/ _+-_ . + __...,_._ +::+-3£:_ -:.-:_ .-6-7 ::Z _- --!._ L_-___ _ _+5E-. :.--++-G X- : :--+.:" +++-i: __ :_:_:.--_. L:_: +--_

+".C_+/ '171+i+: "__: " +":++" ;:-_+ • _ +-" _ '"

_-+_: s-:_-+_ ,u_-:_ a_-_- s ss+L -:-_m+:-- :+ms ._

_ +- • _ - ......... - .......... . +-- _---- . ..... +__ .

--.--_--7 -+-+-+---+_U +:,-.=:_+-_-!_ _ZZ ---+-4 __:_+__

__"+: : :_. + 3_": +_.T: : + . i:::+','+ + +-: _--++ +- : "Z+_- ?_++_T -. "--.--+: :" :.--'_ ...... -: -:+ .... :_:.+::_-_ =+:-:.::_+---.'- :-x +-:::: ...... -.... +__++:: +- _-i.U: Z- U: :--++:- " + ::-+":.-: ....... _ ..... _+_--- + . ........... +- _+ __ ............ --+ --- •-._-: i_ ._---_ ........ ....... _T:__:'_ :.EEL+ -_-,--u'_r-'_-. ......... •-; :- ._,.+-- -: .+-:.... .+zt"-: +"-+-_ + . _...+_ +"_%"r_: :. _ .... " - _'-- - _-'_:.:+ .._ • :+++ZZT': + + ....... ..'_+..-+.. .'3----_-:+- ........ _-- -_--L_.

.. --+ . ___ -':- .- +_: _ +-','+j+ .

ZT:----Z+ :+ _'_ ___--t'-z_- _____,_Z" Z_ OXENDINE 03/86 REVISED DATE CALC A.E. 592 ASHCROFT DEAD WEIGHT TESTER CHECK CALIBRATION CURVE FOR THE 'IGURE 3 MARSH PRESSURE GAUGE APPD APPO PAGE UNIVERSITY OF KANSAS 6 CALIBRATION PROCESS (CONTINUED) The second calibration was accomplished by using the facilities at Richards- Gebaur Air Force Base in Missouri.

Initially the test equipment was prepared for testing. The steps that were involved in preparing the test equipment are outlined in appendix B. Once the equipment was ready, the calibration process was completed by using the fol- lowing steps: A) Isolate the gauge from the test stand system by closing the associated shut off valve.

B) Using an independent source of pressure connected to a master gauge of known accuracy, connect this pressure source to the test port of the gauge to be calibrated.

C) Remove the ring and glass from the gauge and use a screwdriver and ad- just the position of the pointer by turning the self-locking worn adjustment screw D) Then check the calibration of the pressure gauge at several different pressures, when the adjustment :is satisfactory replace the glass and ring Howeve_ when the Marsh pressure gauge was tested, the gauge didn't need to be adjusted, and this fact can be observed fron the data that was obtained during the calibration process at Richard Gebaur. This data can be observed in T.ABLE II, and the calibration curve can be observed in figure 4.

page 7 RICHARD GEBAUR CALIBRATION RESULTS

A

CALIBRATED PRESSURE INDICATED GAUGE PRESSURE (PSl) (PSI) l° 4OO 2.

5OO 3. I000 I000 4.

5.

6.

2O0O 7.

8. 2300 9.

I0.

Ii.

TABLE II page 8 v nEVISEO DATE CALC A.E. 592 m/'EI,IS CALIBRRTION CURVE CHECK FOR FIGURE 4 APPD MARSH PRESSURE GUAGE APPO PAGE UNIVERSITY OF KANSAS 9 DISCUSSION When a pressure gauge or any other measuring device is being considered for a particular task, several factors have to be analyzed to ensure that the proper gauge has been selected for the job. Because, if the time is not taken to properly analyze these factors, complications as well as inaccuracies can result directly from an improper selection. From the available gauges, the Marsh J1678 pressure gauge was preferred over the other models and brands.

Our decision was based on several factors which included the gauges's op- erating environment, readability, accuracy, measuring range, recalibration ca- pabilities and versitility for future usages.

Readability During experimental testing the scale on the measuring in- strument should be highly visible and relitively easy to comprehend. On the Marsh pressure gauge the scale is marked with slashes in I00 psi increments.

The face on the dial gauge has a white enamel background with slashes and nu- merical values painted in black enamel. The needle is also painted black which enhances the reader's ability to accurately interpret the correct pressure.

Accuracy In experimental testing the degree of accuracy in the laboratory data is an extremely important consideration. Therefore, methods should be developed and practiced in the lab to enhance the accuracy of experimental data, as long as the results of the experiment are more important than the cost.

The Marsh company publishes a handbook on standard gauges. This book shows that the Marsh J1678 gauge has a margin of +/- 2% error for the middle half of the scale, and +/- 3Z for the remaining half. From Appendix A, it is evident that the margin of error is much less than either 2 or 3 percent, except at pressures below 175 psi.

Measuring Range The measuring range is a factor that can be easily over looked when selecting the proper gauge. However, through a theoretical analy- sis, it was determined that the ultralight structure could withstand approxi- page I0

mately up to four G's, which is equivalent to 600 psi, prior to catastrophic

failure. With this information, the range of loads that are of interest can

be determined and used in selecting the proper gauge. On the Marsh pressure

gauge, the effective range is from approximately 750 psi to 2250 psi, which is

the middle half of the gauge.

Recalibration When recalibrating a pressure gauge it is beneficial to have the ability to adjust the location of the pointer so that it can be re- adjusted to rest within the zero band when the pressure applied to the gauge is zero. The Marsh pressure gauge includes a zero band denoting that the pointer may fall anywhere within this band when the gauge is properly calibra- ted. In addition the gauge is designed in such a way that the needle can be adjusted within a limited range so that a seriously damaged instrument can not be falsely recalibrated.

J Vers{tility_ When a gauge is selected for vers/tility a decision has to be made as to whether the gauge will be used for a specific task or for a va- riety of tasks. If the selection was based on a specific task then, gauge vers/tility can be limited. However, if the gauge was selected based on ava, riety of tasks, then the gauge will have to be versatile in order to be used efficiently. When the Marsh pressure gauge was selected, the selection was based mainly on precision and accuracy. Even though vers/tility was not a de- ciding factor, the manufacturer designed the gauge with versatility in mind.

The universal design features of the Marsh pressure gauge can be observed in Table V.

page II CONCLUSION From the limited selection of gauges that were readily available the Marsh J1678 pressure gauge was selected as the proper gauge for the task. However, there are gauges on the market that would have proven to be more efficient in accomplishing the same task. Also, it is evident from Figure I that accurate scale reading will be difficult to obtain. Although the margin error (inac- curacy) is not suspected to exceed +/- I0 psi. Although even with this error and after analyzing the extent of accuracy that is required during experimen- tal testing, in conjunction with the capabilities of the Marsh pressure gauge, it was concluded that the Marsh gauge would be an acceptable measuring device.

In determining the accuracy and precision of the Marsh instrument, the Gaussian distribution method was used and the calculations are outlined in Appendix A.

The results of the Gaussiam distribution for the _/- 3s approach are as follows: FOR THE DEAD WEIGHT TESTER Po =" (0.9795 + 0.0392)P i -2.62 +_ 12.78 FOR THE HYDRAULIC TESTER Po " (1.002 +_ 0.004)P i - 4.22 + 10.77 Where: Po - Marsh Pressure Reading (out-put) Pi " Calibrated Pressure (in-put) page 12

APPENDIX A (CALIBRATION CALCULATIONS) page 13 CALIBRATION CALCULATIONS In the calibrating a pressure gauge =he relationship between the cali- brated input pressure and the output ( Gauge Reading ) pressure is ideally a straight line. However in reality nothing is perfect. Although the cali- bration curve is still considered to be a straight line. This line was de- termined through the least squares method. This method minimizes the sum of the squares of the vertical deviations of the data points from the fitted curve.

USING THE LEAST SQUARES METHOD PO= MPi + B Where: Po - Output Quantity Pi " Input Quantity M - Sole Of The Line B - Intercept of the Line On the Vertical Azis N_PiPo - (ZPi)(_Po) M ml N_;Pi 2 " (_Pi)2 (_Po) (_Pi) 2 - (ZPiPo) (_;Pi) B :m N _Pi 2 (_ Pi) 2 Where: N is the total number of data points.

page 14 STANDARD DEVIATION N Spo _2 m N_Pi 2 - (_%.Pi)2 Sb = N_p2i - ( _,Pi )2 The numerical values of the mean and standard deviation were calculated for both calibration processes. The data that was substituted into the above equa=ions were obtained from Table Ill and IV Where: 2 I Sp ° -_ (_MPi + B - Po ) page 15

FORTHEASHCROFT lEST

MEAN

( s )(5.25xi05 ) - (1405)(1365) M r.

(5) (5.39xi0 s) - (1405) 2 7.072XI05 - 0.9795 7.219XI05 (1365) (5.39XI05) - (5.25XI05) (1405) B s 7.219XI05 - 1.89XI06 I' - 2.62 7.219X105 STANDARD DEVIATION S -2 m • , 1.308X10 (5) (123.66) 7.219XI05 + -2 FOR 35, S m - - 3.92 Xl0 page 16 S b ....... =, 4.26

I (123.66) 2 (5.39XI05) i_2

7.219 XI05 / -+ 12.78 FOR 3sj S b FOR THE RICHARD GEBAUR HYDRAULIC TEST MEAN (Ii) (4.368XI07) - (1.989Xi04) (1.991XI04) M :m (11)(4.369 Xl07) - (1.991XI04) 2 8.474 XIO 7 , - 1.002 8.458 XI07 (1.989 XlO 4)(4.36 XIO 7) - (4.368 XlO 7)(1.991 XlO 4) B s 8.458 XI07 -3.573 XlO 8 •, -4.22 8.458 XIO 7 STANDARD DEVIATION - 1.38 XlO "3 S " I (II)(161"49) _ _'a m 8. 458 XI07 - 4.14 XI0 "3 For 3s I S m page 17

(161.49) (4.369XI06)

8.458 XI07

= 3.589 S - 10.77 For 3s 7 m page 18 CALCULATED DATD OF TfIE ASHCROFT TSET Po P o PiPo p2 _2 O 25.0 25.0 25.0 2O5 4.1 Xl04 4.2X104 4.0X104 2 90 8.85X104 9.30X10 4 8.41X104 3 90 I. 58X10 4 I. 64XI0 4 49O 1.52Xi05 2.38X105 2.41X105 2.35X10 5 5.25XI05 5.39XI0 5 5.12X105 TABLE III page 19 Q G _ _ O 0 0 Q _ Q _ X X _ X X X X I_ X X X X Q co i-.-i ,...I X X _ I_I I_I X X X X X X i-I Q X ,-I page 20

APPENDIX B (CALIBRATION PROCEDURES) paEe 21

MADE ONLY BY

ASHCROFT GAUGE DIvrsION

OF Fizure V EaSe 22 MANNING. MAXWELL 8- MOORE INC.

BRIDGEPORT CONNECTICU

T.O. 3ZA2-2-'_ 5-I Sc<,icn \'[; F='=_-aph$ 7-1 to 7-4

SECTION VII

CALIBRATION

7oi. O ENEI_AL.

(2) Then increase the setting Uy h_.tf for a margin of stabili_'.

7-2. A cAlibt, mtion check _ required every- 180 Clays, however, caJ.ibracion of the complete test stand ss a _it Ls e. H obser_-ation of _,e temperat-.:.-e control2er sho_ not considered practJc_l. Refer to paza@aph 3-5 for the Lhzt the cent.rolled temperature is _lu_ish or _-ander-

[ initial adjustmen_ to be made before opera:ion of the test

ir, g, proceed a.s follows: stun d.

7-3. FLUID TEMPERATURE CONTROLLER.

(I) Tuxn proportio,'_! band ad_-z_tmen: (12) with a (15, figure 4-2.} screwdriver to decrezse the proportional band in steps until measurement Ls jittery or just cycles a bit.

7-4. To adjust the fluid temperature controller, pro- ceed as follows: (2) Increase proportional banduntil control is stable.

(3) Then increase the setting byhalf for a margin of stability.

The fluid temperature controller requires Note clean, dry, oil free air at 18 to 20 psi. A piece of tmrd paper (flat free) placed between An attempt to sec'.:_ea ,¢lne o.:erat',n_adj'_t- the nozzle (I0, figure 7-I) and the flapper (9) ment which is just s.'nbleunder Uhe operazing wilt show the presence of moisture, oil, or conditions of the moment is net adv'.sed since dirt. Add dryers or filters to the air supply sllghtly chan_ed cperatin_ co=dr:ions w:ll line a.s required to obtain clean dry air before probably result in Lns_bt_.ity zr.d cyclir.g.

operating or calibrating the temperaraxe con- troller. Be sure the flapper is lined up with the nozzle and makes a square contact.

f. Normal adjustment of the temperat,.tre con_oller should not requLre excessive adjustment. If the pro* a. Turn on air and dralnfilter (15, flgure4-7) through cess being controlled is sub}ect to ex'*-"eme tempera- its drain v_ive. Adjust pressure regulator (6) to 20 lure changes or frequent shut-dowrts and start-ups the psi supply pressure as shown on supply g'Age (15, fig- temperature controller should be observed through ure 7-i). Set red index pointer (I) at 100*F by tttrn- the period of upset to rn;tke certain t.l'.Rt it remains in_ index setting knob (6). stable.

b. Operate the tests:and to pump ell past the sensing g. If continued adjus:.,r.ent does not bring the process eie.--r.ent of the temperature controller (refer to para- under control, refer to the trouble shooting table in graph 4-5 and step j of paragraph 3-5 for this oper_tt- Section VIaad check for erratic behavior in the hy- in_ procedure).

draulic system, water system, an_ _emperat..u-e con- troller. To determine ifthe controller or the process c. Observe the operation o( the temperature con- is at fault,operate the controller m_nually as follows : troller.

Note (I) Setrcd index po-nter (I)well alcove b[ackindicat- ing pointer (2) and above Cne d:.=:Te_ temperatu.-e of Temperature control processes respond slowly ',.he hydr:u::,, iiutd.

(as corr.p_.red with pressure). Be sure that the period o[ observation is ofsufficient length fcr (2) Adjust Lhe air suyply press_re regulator valve the controller to respond to changes in ell to _ary the press,J.re c_ th.: coal-,:Icr diaph.ra_:-2, and temperature. Also, the position of the sensir4_ thus manu:liy regulate the zctio:_ _: Lhe controller.

element in the hydraulic circuit will cause long delays in _djustmg due to load changes.

(31 When temperature stabilizes at desired value, record the pressure on the output ga_e (S).

d. If observation of the temperature controller shows that the controlled temperature cycles too much, pro- (4) Move the red index pointer (I) back toward the ceed as follows: oswed tcn'peratu.,-e ur.t_i U_c pr, ssure on the output age (5) Ju t drops. Restore Ll,ea:r supply" pressure (I) Turn proportional band adjustment (12, figure : , 20 psi. ' _ 1 " I _ _ [ the red ind, cx p,.,,_ter to be su_r t2:,.

7-I) with a screwdriver to increase (widen) the pro- . _u'_._u.rc ,.., th(" output k.-a::,: _t; l:_:u_.ht to the exact portional band in steps unt_l the controller is Just stable.

:due r,..r_. :._ in step (3) _._v{,. " / /

, #,r 1

Cl;an_-.cI '7 April 1(Ld39 7 i j • _l TO. 33A2-2-35-] i. Red [=dex PoLnter B. Feed_ck DLmphr_..,., Assembty 2. BL.%ck [nd[c:_tm_ Pointer 9. F_pper 3. Process Connection S[ock i0..\'oz z Le II. Pro_or_',on_l DEal 4. ,_,_es.surin_ Herod A.sser.,Dl 7 12. Proportional Bnnd Adjustment (mercury _c_ated) 13. [_e!/y Asse.'=.b[y 5. Ou/put Ga@e i_. Or[£{ce Cleaner Bu_ton 6. _ndex _ettm_ Knob 7. SynchronLz_n._ Nut 15. SuppLy Ga._e I_gure Ftuid Temperature Controller, Door O_en Note c.._...__. _e -_uze ._[[Q;v _,e_-_'J.re.'r.e.-,t :o iet-:e zt _'_e _,e_red _m.lue be£,=re proc-=ed_c_.

[_ the process can be controlled ..,.,_n",all?

(s_e_s _. _'_rou_h .3) but not 3.utorn_!¢_._7 (step 4) L. L[ cu=put _,'a_e pressure Ls signi£[czncly dLqerent _he _rcuble L_ _n u_e controller. [/ u%e process [ro.--.. 9 [_sl rand uhe rest stand {s :o De oper1:ed _ one cannc_ be controlled m_nu,_Hy, the L_0uble _S tem._er:_c,lre for z [on_ per{od, the red po{:'_er m-'*y {n the _-ater system or the hydrzul{c _ysteun.

be brou_5_ to ;1 ,'-- -_: ,_ .. ..,a.,nz..$ position "_'tLh ,ie black potn:er _v :urntn=_ sy_:chron_.z{n_ nu: (7).

Note r.Ic l,.t:l'?,'i '.,irC'4 :o ::._ ,'onu',,L'.. ::d page 24 T.O. 33A2-2-35-1 Sec'._on VII Paragraph_ 7-5 to 7-9 to be frequentl 7 v3.rted, synchronization wi.',.h c. P.en:ove :he r:('_ _nd gL_ss fro,"_ "..',e gage. Use each load change is not necessary; proceed as scr_wd.-:,..: and ac:i_st :_e pos_:_._n of '_,e po:n:er by in step h _.bove.

_'J._nm_; ".._,e _e,:[-_oc_,_ worm adlv'_:.-:-..e_t sc.-ew.

7-5. HYDRAULIC INDICATORS ZERO ADJUSTMENT.

_-. C.hec:< :?:e call,rat:on 9,_ "_he _aRe .t_. severctl d:f- The pressure g;_es suppl, ied ,utr.h. the test st:lnd h_','e _ren,. pr,.'s_u'.__s. '_Vhe,: a__;._stmer.t t_ sat-s,'_c=ory, td,_ust_ble po;nter_¢ :o permtt rec_].Lbra::ng the ,,,_g'e_.

to rec:tlibrat_, 0. g_.ge, proceed as .fo[lo,,vs: e. P._-.p,,,Z ,''_ .1:l :R:_cc;rD./e _';._e :,_l,}.c C"':,,'1,2: .';e :"ecCz_[- a. [solz:e t.he g"_ge from the test st'_nd system by _ ;" aA...,._% c],osmg L_.e _ssoctated shut of[ valve.

b. Use an Lndependent source of press:m _. ,hand 7-6. ELECTRICAL INDICATORS ZERO ADJUST?dENT.

pump) connec:sd :o a master &_ge o£ known acc'=:a,'y; "I_ae;'oltme_.er _d ammeter ace. suppl!ed ;;'i,.h _ ex:emal c_nnect :h's pressure Source to Lhe :est. port of ".he go.ge co be c_.itbr_,ted. zeroZer° adjus;ment.with r.o cu_ntL'_e,,o,.v, "_a_crewdriver to adjust, poin _er to ] 7-7. RESERVOIR AIR RELIEF VALVE ADJUSTMENT.

T'ne :tit relier" v_/ve, ,_or '_he hydraulic rese_:oiz. (93, figure 1-5) mUst be set.to re_.ievei£ pressure in the line exceeds 125 psi, By applying regulated air, it can be determined at what psi the relief valve opens. Tl_e pressure ar which the v-alve initial/y opens can be adiusted by inc=easing or dec=en.iing the spring :enston.

7-8. INSPECTION OF RESERVOIR LEVEL FLOAT SWITCH. The switch, S15 :'i._ure ].-6, _=ll cut o£t" the elecI:.-ic immersion heaters it" :he hyd:au!ic fluid level re/Is below 3;4 full. If _he sw_tch does not function properly when inspected replace it.There is no adjus{ment.

7-9. MANOMETER CALIBRATION. The accuracy of" the manometer is confirmed by initial preparation and :he be- fore use adjustment requirements con:aiRed in paracraph 3-5. I. Furthe: c_libration is not required.

page 25 Cl:,In(I': 33

APPENDIX C (MANUFACTURE'S INFORMATION ON MARSH GAUGES) page 26

Marsh Standard Gauges

Dial

JSl B40.1 Grade B accuracy

New cupped dials are made of steel, with

s ±2% of span inmiddle half

white enamel background and black printed

3f scale, '-3% of span for rest

matter. 2" and 2'/_" only.

3f scale.

Case patterns and construction

Plain Case. Slip Ring--drawn steel, 1½% 3½", 4 'I=".

Specifications

Plain Case. Twist-lock Ring--drawn steel, 2" and 2.'/=".

Plain CIearfront--drawn steel, 1½_.

Accuracy

Stainless CIearfront--drawn stainless steel, Grade B Pressure and Vacuum Gauge 1 >;" and 2".

_ s_uecifications as established by ANSI Standard Flush Case, Snap Ring--drawn steel, 2% 840.'t --1974 sta_es that the permissible error 2Y, 3½".

shall not exceed 2% of span at any point between 25% and 75% of span; in the rest of Liquid-filled Plain Case, Nonremovable Ring-- the scale, 3% is permissible.

phenolic, 2 '/=".

Drawn steel cases and rings are finished in black semi-gioss enamel.

.¢_zes and connections

Drawn steel cases in a flush pattern have a clear zinc finish.

1 ½% 2", 2½", 3½" and 4½" dial sizes. All connections are male N.P.T. 1 '/=" size has _" Drawn stainless steel cases have a brushed bottom or center back outlet. 2" and 2½" sizes stainless steel finish.

have ½" or ',/," bottom or center back outlets.

3'/-_" size has ¼" bottom or center back outlet.

,= size has "" outlet. Lens

4 ''_ ,, botte4"n All Standard Gauges are supplied with flat glass lens except for Clear'front cases, which have

Bourdon tube assembly

a molded acrylic press-fit front. 1½" Plain Case Gauges have a fiat plastic crystal.

For Vacuum and Pressures to 600 psi Tube, tip and socket are copper alloy.

• Phenolic case liquid-filled gauges--

For High Pressures, 1,000 to 5,000 psi

special construction features

Ni-Soan-C Bourdon tube; copper alloy tip and Neoprene plug seals fill port.

socket.

Strap-in, nonremovable _olypropy|ene retaining ring.

M ovement

Accuracy is ±3% of span in middle hail of scale.

300 series stainless steel internal construction is Standard movement for all 2", 2;_', 3_£', and available in bottom connection in selected 4½" gauges is the new Accutite" 2000. It is ranges.

made of glass-filled thermoplastic polyester.

and is available either with or without 2½" dial size only.

Recalibrator in some models (see Selection Cupped aluminum dial with black numerals on Guide).

white background.

1½" Standard Gauges feature a copper alloy Restrictor screw is supplied as standard.

movement.

Glycerin filling dampens pulsation and vibration, Suitable for use from --30" to 150"F. Other See _.-=ge 3 for fuller descriptions of both fills available on s#ecial order.

increments.

page 27

Tdarsh Standard Gauge Selection Guide

01AL SIZE 1 ½" t ; 2" ! i CASE MATERIAL Steel t_tain(ess Steel! Sleel I CASE PATTERN PYain ;lain I CIearfront ! CIearfront i Center C0q;IECTfON LOCATTON Bottom Battom Back 8ott_ml Back 8ott:m Back Back C0,*;NECTIO,*I SIZE 7," ¼'" W' 20," I i No No No No No No AECALIBRATOR No No i Yes _;o No _Jcne ,_ane _]ene AJone )Jane RESTR_CTOR '(_s _lane 13lane :,,:he" Nan° ] i V,._CUUM 30" Ng/--Ir'Jo kPa 'JII_5:jT_'C5 i! _j -_-:,_.,_ -,n Ik I " .J..05 t i i I • * !JlllZ J141Z

I !JZO_Zi

,_0" Hg × 30 psil--100 x 7.10 kPa

i tJl11_ J1411

_0" Hgx 50 0;i/--IC0x 400kPa i / i .

• • ° • • I 30"HgxlOO_si/--lOOx 700kPaj CCMPOUNO I _J111_ J1413 IJlal8

30" Hq x 150 psii--lOOx I000 kPa (

l x 200 _si/--lO0 x 1400 kPa

I 30" Hg

x _00 ;asi[--lO0 x ZTO0 kFa 30" Hq x 400 psi/--lO0 x/._00 kPa 30" H_ COPPER ° • • )J1840 _.. ALLOY !Jl_42 JZ94Z 80UROON JI047. IJ1647. Jl14_. JI44Z

•

JOO42 ,oz_z IJo4.z JO64_I Jo.2

TUBE !J1146 Jl_t,.t8

:j

1143 j144_ i I I

_J1152 JT452 iJ1a_Z t J7.057.1

PREaSUF.E J0252 • 150 psill.lO0kPa J0057. J1552 I

!J;_54 }j_a_4 IJ13.-IJ_.S4

J07.54 J0454 200 p_i/1.4OO kPa JOOE& J165_ .-i.- i 1 300 _silZ.l_O kPa IJT.0S_l !J1160 J14_O 500 psit3.50O kPa Q • J1464

.osil4.0O0 kPa I "

I • JZ064 J1671.'

1.000 HIGH 1.500 NI-SPAN.C 80UROON PRESSURE 2.000 psi/14.000 kPa TUBE 3._C0 psil?.1.000 kP3 5._00 psiiaS.P_00 kPa [') all high-lo_es,=u_s gauges hays restrictor= as standard equil=ment •.A B I_ V page 28

VII. DESIGN OF STATIC REACTION GANTRY

FOR AN ULTRALIGHT AIRPLANE

DESTRUCTION TEST

AIAA paper #85-4022

Howard W. Smith

Professor

Department of Aerospace Engineering

University of Kansas

October 14,1985

Partially supported by

NASA Langley Research Center

Grant #NAG 1-345

DESICN OF STATIC REACTION GANTRY FOR AN ULTKALICHT AIRPLANE DESTRUCTION TEST Howard W. Smith* Uaiverslty of Kansas Lawrence, Kansas Abstract Discussion The steel gantry superstructure needed to perform an airplane static test is described.

Approach Standard civil engineering design practices are used to react the loads generated by an airplane Since rime and funds were limiting factors, in flight. Reaction columns are mounted on a it was decided that a structural test to structural floor to carry the wing alrloads and destruction would be performed in the same r,=nner the downward acting fuselage loads are carried as an FAA statlc test would be performed for directly into the floor. The gantry can certification of a new general aviation accommodate a general aviation airplane or airplane. Testlng was abbreviated to include rotorcraft. An immedlate use for an ultrallght only one flight condition. The "point" to be alrpla_e is shown as an ex_mple conflguratlon of tested was chosen as point "A" on the V-n the four main steel frames.

diagram.

Introduction Airplane Descriptio n There have been several accidents involving The manufacturer called the airplane an "Airmass Sunburst Model 'C'." It is nine feet ultrallght aircraft. In some of these the integrity of the structure was questloned, [1]**.

high, sixteen feet long, and has a wingspan of As a result it was decided that a structural test thirty-slx feet. Additional details are shown in should be performed. Figures I and 2, and Table i.

=

/° • ° Airmass Sunburst Model 'C' Fig. I *Professor, Aerospace Engineering **Numerals in brackets are references.

Associate Fellow, AIAA Table I Hangar Description A specially designed hangar houses A1r_Iss Sunburst Ultrallghc Hodel 'C' university-owned airplanes. The eastern half alao has a structural cesc floor, which Is a scaled version of the structural floor at the Specifications: Beechcraft Plant in Wlchltaj Kansas. Figures 3 and 4 show the salient features of the floor. A Length 17.58 ft cruciform rest sec_lon is fourteen inches of Helghc 9.69 ft reinforced concrete, with "l-Beams" embedded in Wing Span 36.00 ft floor. These embedded beanas provide "up reaction" where needed, and also serve as a Wing Area 150.93 ft 2 foundation for the steel columns of the gantry.

Anpect Ratio 8.59 A major shortcoming of the hangar is the MGC 4.19 fc lack of an overhead crane. A clearance of Wln E Taper Ratio 0.92 EwenEy-one feet six inches is available for mobile crane operations.

_ncldence Angle 5.50 deg Tall Area 28.04 ft 2 Loads Tall Span 9.33 ft The empty weight of _he airplane is 273.9 Dihedral Angle -40.00 de E pounds, determined by three-polnt weighing.

Total weight ("Basic Flight Design Welght" ) Is_ Performance Specifications : Fuel 15.5 # Pilot 175.0 1.45

%

max Alrp. 273.9 TOTAL 464.4 OWE 277.48 lbs S r.all Speed 43.11 fc/sec 50-75 ft/sec Cruise Speed Cuyuna 430 c= 30 lip engine.

PxRT_a, I } ._; -8 _" [rlCMIES IIN_MES 2 _-J _ _0._2 52284 [ 3 wG-27,.,_,,37 R2 4774d 5L588 4 I v4G -SS H3 78.252 5(1_6

s l _;-s2

, _4 I 12248 a _ *q3 -37 R_ '" 14_-_.244 4_5_ R{XI TS -3._ WING AREA: S: #_'JO_3 FT z ( \L / | I0 TS ._:l WING 6Pk_4; b: 36.0 FT ] [ _ It 05.49 : ' FO_aO 0¢ w_ _G EOGE / _ SUN.JeST " _CO_- "--C" Ime_:_, ._,so.. _ w,s_s -- Fig. 2 Planview - "Sunburst".

FiB. 3 Cruciform Floor.

_' - O* 4- _ - 0 ° 4-' - 0" "2' - O" I _Z= EXP._NStON JO| N-r ALL

i • i/_z7 i

2: : :_" ;jCAPILLAEY WAT E l_, j, EACI-I W_,Y I_AI_ R 1EP,.

_ WATI= p.. • E./_4:_4 FACE.

TYPICAL SECTION

©

Fig. 4 Embedded Beams. - For structural test purposes, the design 11mlt Each column base plate was centered over a load factor was assumed to be. n - 4.0. A factor floor beam. Each of the three parallel floor

of safety of 1.5 was assumed, TM •

beams is on four foot centerlines, and the columns are located on the outer beams. Since Using these values, the estimate maximum _he portal height was chosen to be sixteen feet, ultimate load is: a portal is twice as high as 1_ ls wide. Each of _hese portals acts as a slender frame, and 1.5 (4.0) (464.4) = 2786.4 pounds requires sway bracing normal to the plane of the portal. An external brace is located on every Rounded, the design ultimate load is 2,800 ibs.

column ten feet from the floor and extending outward and downward at a forty-five degree Steel Gantry angle, Fig. 6. The sway brace itself consists of clevises aC each end, a turnbuckle and two five- Steel used for the superstructure was eighth inch diameter rods. Each column is tied designed for a general aviation airplane of the to its nearest neighbor wlth a short sway brace, "King Air" class. Using the 12,500 lb. limit as and the four columns near the wlng-body root are prescribed by FAR Part 23, the ultimate load diagonally tied with long sway braces, Fig. 7.

would be 1.5 x 4.0 x 12,500 = 75,000 pounds.

This load can be carried by four reaction All the steel is type A36 ag_ all bolts are columns, gound off this number, a column load of type A325 per the AISC Handbook, TM • X list of 20,000 pounds was used for the steel design. A the s_andard steel section chosen is given in beam connectiog each pair of columns was designed Table 2.

for a 40 kip load. A beam and two columnsj called a "portal', was provided for each wing, the aft fuselage, and the forward fuselage. The four portals are connected co each other with beams in the water plane, Fig. 5.

9- 3//3

\

\

\

\

Overall Steel Installation.

Fig. 5

qE

]E -7

'N r , m -$ O/E-I...

(s_. 3) "x

"-R,E JM<.V IEV_."

Fi E. 6 External Sway Bracing.

7..

I

I

sy_ I//-- g-dO//-Z

/

/

T c-I

Z. *_ _-_'.

t

PUAI4 VIEW ..

Fi E . 7 Internal Sway BracinK.

Table 2. Steel Secclons.

8 Columns wgx24 16' 4 Channels C12x20.7 5' 4 Beams Wl6x40 5' 2 Beams WI8x40 12' 4 Beams W8x24 8' AX1 bolts loaded _n termion and shear are three-quarter £nch diameter. Bolts at column base plate clamps are f£ve-eighch inch diameter. Bead-co-beam connections are _ade by "good civil engineerlng practices." A pair of angles is fillet welded to the beam web ac each end. The outstanding flange has a hole pattern chat matches the repeating pattern in each column flange. Beam "sear" angles are provided for easy construction and disassembly. All assemblies were cleaned and grey primed after welding. All assemblies were painted royal blue before installation.

Acknowledgements The _ork described in thls paper was performed under NASA Langley Research Center Grant mz_ber LRC/NAG 1-345/3-3-83. Six students from the University of Kansas did the drawing of the steel gantry, (in alphabeclcal order): Albers, Roger Bultman, Myron Clune, Mike deAlmeida, Sergio Martin, John Eobertaon, Greg There were ,any ocher people, including students, staff, faculty, and townspeople Chac gave freely of chair tl_e. People from Lawrence and Kansas City have given advice and aervices freely.

References [1] "Ulcralight Vehicle Accidents: Safety Study," NTSB/SS-85-01, Feb. 7, 1985, 405 p.

"A/rworthiness Standards for Powered

[21

Ulcralight Vehicles," Powered Ulcrallght Manufacturers Association, 7535 Little U/vet Turnpike, Suite 350, Annadale, Virginia 22003, Dec. 9, 1983.

Manual of Steel Construction, AISC, Inc.,

[31

400 North Michigan Ave., Chicago, IL 60611, 8oh Ed., 1980.

G