Design of A Minimum Mass Wing-Box: Professor J. Cooper
Design of A Minimum Mass Wing-Box: Professor J. Cooper
AEROSTRUCTURES 2
Alex Farrant
Christopher McGarry
Josh Plant
Oliver Weston
15 December 2011
Professor J. Cooper
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
Contents
1.0 Introduction .............................................................................................................................. 3
2.0 Design Intent ............................................................................................................................. 4
3.0 Square Wing-box Design ........................................................................................................... 4
3.1 Distributed Loading Analysis ................................................................................................. 5
4.0 C-Section Design ....................................................................................................................... 7
4.1 Wing Deflection .................................................................................................................... 7
4.2 Wing-box Shear Stress and Twist .......................................................................................... 9
4.2.1 Mohr’s Circle Analysis ................................................................................................. 11
4.3 Wing-skin Buckling Analysis ................................................................................................ 11
5.0 Final Wing-box Design............................................................................................................. 15
References .......................................................................................................................................... 20
6.0 Appendices ............................................................................................................................. 20
2
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
1.0 Introduction
The wing-box of an aircraft is the component that makes up the wing’s internal structure, to fit within
the important aerodynamic shape which defines the aircraft’s performance. The wing-box must handle
all expected operating loads whilst retaining structural integrity, despite routine exposure to fluctuating
loads. The main difficulty in the design of a wing-box is the challenge of achieving a design which has
minimal mass, to ensure that the aircraft can maintain maximum operating efficiency.
Conventional wing-box designs consist of spars which run along the length of the wing, form the
connections to the fuselage, and effectively handle all experienced loads. Wing-boxes also require a
wing skin, which consists of thin metal sheets that enclose the wing-box, and provide some shear and
bending stiffness. For added strength, ribs are added which run from the trailing edge to the leading
edge, and resist twisting in the wing to help maintain the desired aerodynamic shape. This resistance
can also be achieved through the inclusion of stiffeners, which run parallel to the spars to prevent
wing-skin buckling.
The design of a wing-box must take into consideration the specified wing profile, which limits the
wing-box dimensions, and support the aerodynamic shape in normal operating conditions through
handling bending, torsion, and shear. Many modern wing-boxes are based on similar designs, with C-
or I-beam spars supported by ribs and stringers. Modern wing designs are also tapered, which
distributes the wing area more towards the wing root, which reduces deflection.
This report outlines the design of a minimum mass wing-box for a given aircraft. The design involved
analysis of deflection, torsion, and shear experienced by the wing, to develop the lightest wing-box
structure which will operate safely within specifications, when constructed from a specified aluminium
alloy.
The report demonstrates the step-by-step analysis that was carried out, but only in the general case.
A MATLAB programme was then constructed, which iterated through a number of possible wing-box
dimensions to find the most light-weight design that would handle the expected loads in normal
operating conditions. The results of this numerical analysis were then compared with finite element
analysis carried out in Pro-Engineer to verify the final design.
3
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
It was also followed that the design should include only two spars: one at 20% chord, and the other at
80% chord.
Using ProEngineer, the height of the wing thickness at 80% of the chord was found to be 250mm. The
lateral distance along the wing from 20-80% of the chord is 2.88m. Figure 2 shows the dimensions of
the square wing-box analysed in this section.
4
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
B (2.88m)
t (20mm)
D (250mm)
The thickness, t, of 20mm, was a value taken from typical wing-box designs. The second moment of
area of this box is found using the equation shown below:
L (m)
z (m)
With this term, the deflection, y, can be found at any point along the beam. To find the two integration
constants, the conditions y’ and y’’ =0 at z = L were used, as the wing-root is fixed. With this, the
5
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
Although non-uniform distributed loading arises from wing taper, triangular loading analysis was
carried out on the same wing-box shown in Figure 2, to demonstrate the analysis method and to
develop an understanding of wing-box analysis. For simplicity, it was assumed that wmax occurs at the
root, and decreases linearly to w = 0 at the tip. This is shown below, in Figure 4:
L (m)
z (m)
Again, the Euler-Bernoulli beam equation was used to derive an equation for the maximum deflection.
The z=L conditions of y’ and y’’ = 0 still apply. In this case:
-3
Referring to the dimensions in Section 2.0, the second moment of area, Ixx, was found to be 1.56x10
m4. With this, an aircraft mass of 74,000kg, and a Young’s Modulus of 70GPa, the maximum (tip)
deflections are:
S&L Flight (m) Worst Case – Gust (m)
Uniform Load case: 1.70 6.81
Triangular Load case: 0.91 3.63
From these results, it can be seen that the beam under triangular loading is deflected about half as
much as the uniformly loaded case. Despite this, tapered wings have increasing I xx from the tip to the
root, and so it is expected that triangular loading on a tapered wing would cause a higher deflection
than either of the cases shown above. Taper will be in the final design, however, as it is much more
weight-efficient than the simple uniform section that was analysed in this section.
The importance of considering worst-case scenarios is also highlighted – with the gust causing a
deflection four times that experienced in straight & level flight.
The maximum allowable stress of the material is 350MPa, so the wing-box design must not
experience any direct stress above this value. The stress on the beam was calculated using the
formula shown below, with the maximum stress occurring at the wing root, where the bending moment
is at its maximum value:
6
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
b2 B2
d1
d2
Hav
D1 y1 y2 D2
To calculate the second moment of area for this section, the Ixx values for each C-section were added,
along with the area contribution for the skin – the bd3 term for the skin was neglected as the skin is a
thin section, so t3 0. The full formula for Ixx is shown below, and the last term is the skin contribution
to Ixx.
The rate of change of Ixx over the wingspan was simplified to λ. This relationship was substituted into
7
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
the Euler-Bernoulli equation, along with M(z) for triangular loading, shown previously in Section 3.1:
This equation was then integrated once, to find the first integration constant:
The integration solution was found using an online integral solver(Vanovschi, 2006). This was then
integrated again, to find a term for the deflection containing two integration constants, c and d:
The constants, c and d, were found using the same method employed in Section 3.1 – y’ and y were
assumed to be zero at the wing root: where z is 16m. With this, the wing deflection is:
8
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
To reduce human error in these lengthy calculations, all arithmetic was conducted on MATLAB.
Figure 6 - Bending Moment along Length, L, of Tapered Cantilever from Root to Tip
(eFunda, 2011)
To calculate the shear stress, a boom approximation analysis was carried out on the wing-box at the
root. The boom approximation equates the C-section wing-box to an open section with four point Ixx
components (booms), where the shear acts through the shear centre, and a section with a constant
torque. This is done by equating each member to a zero-mass length with booms at each end:
b
Δq1 1 2
3 4
Sy Sb
Figure 7 - Wing-box Boom Approximation
From the boom approximation, the shear from boom 1 to boom 2 was found to be:
In the above equations, the letter B represents point areas, which were found by equating the Ixx of
each member to two booms, which themselves have no Ixx. For the final wing-box section shown in
Figure 5, B was found using:
9
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
In order to calculate q0, the unknown shear flow due to the moment (constant), moments were taken
about Boom 3, the point where Sy acts:
From this, the shear flow of the open section, q 0 was found to be:
Shear stress in each boom due to the shear flow (shown above in Figure 8), is given by:
Where q is the shear flow, and t is the thickness of the material. For t, the thickness of the skin used
to give a critical value. With the shear flow distribution acquired from the boom approximation, wing
twist could be calculated using:
Because the wing root is fixed, and z is measured from the root in this case, θ is equal to zero at z is
0, so the integration constant, C, is equal to zero:
Again, to minimise human error in calculations, this was programmed into MATLAB for the final wing-
box design and analysis.
10
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
The maximum shear stress experienced by the wing-box was found by conducting a Mohr’s circle
analysis. This was relatively simple to do, as it can be assumed that the stresses acting on the wing-
box elements are one dimensional, in the y-direction only:
Where:
The τ value is taken from the equation for τ given by the shear-flow analysis.
For the equation above, it is observed that a minimum occurs when n is equal to one, hence the
critical load becomes:
Here, n is the number of half waves which occur laterally across the thin walled structure, and m is the
number of half waves that occur longitudinally along the z axis. The buckling load can be expressed
by introducing a buckling constant, k, which is defined as:
11
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
The number of half waves in the z axis is still an unknown. This number is an integer and is defined as
a function of the aspect ratio:
Table 1: Upper- and lower-limits of the aspect ratio related to each mode
The length of the wing box is 16m, with an average width of 2.5m. The aspect ratio is therefore 6.4,
which below the upper limit of m=6. From this, the waveform expected of the wing box is sketched
below:
Once this has been done, the buckling coefficient and critical load could be calculated with a value of
m=6:
Note here that D is known as the flexural rigidity of the material and is defined as:
The specified alloy has a Young’s modulus of 70GPa, a Poisson’s ratio of 0.39, and an initial
thickness of 1mm. Hence, the critical load is:
12
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
As the bending stress distribution through the wing is known, it was possible to design the
strengthening mechanism required to prevent a 1mm thick wing-skin from buckling. Because the
wing-skin takes up such a large area, it was decided that a thinner skin with strengthening
components would be much more weight efficient than a thicker skin.
With Microsoft Excel, the critical stress at points along the wing box were evaluated and compared to
the expected stress, as shown below in Figure 11. The full spread-sheet can be found in Appendix 2.
70
60
50
Stress (MPa)
40 Critical Stress Pa
20
10
0
0 5 10 15 20
Distance from tip (m)
Figure 11 - The Red Series Shows the Expected Stress along the wing-box
From the graph, it is clear that the expected stress is much higher than the critical stress, so a
longitudinal stiffness design is required. Due to the scale, it is difficult to see the variation in Critical
Stress, which has therefore been plotted on the figure below:
13
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
60
50
Critical Stress, KPa
40
30
20
10
0
0 2 4 6 8 10 12 14 16 18
Distance from wing tip (m)
The stiffened members only run 15m along each wing, as they are not required at the tip. The skin
needs to be broken up into sections:
Wing Chord
bsk
The critical stress was taken as the bending moment at the root during a worst case scenario gust.
Using this, the width of each “segment” was calculated by solving the above equation for b, which is
the distance between each stiffened member. The width of the wing-box was then divided by this
distance, to find the number of stiffened members required at the root to prevent the wing-skin from
buckling:
Applying this analysis at different points on the wing-box demonstrated that the number of stiffeners
required reduces, moving from the root to the tip. From this, it was deduced that some stiffened
14
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
members do not need to run the full 16m. The depth of each stiffened member was determined by:
With a skin thickness of 1mm, and the critical stress value found earlier, the stringer depth was
calculated as 20mm. To prevent buckling, 88 stiffened members are required at the root, each with a
thickness of 1mm and a depth of 20mm. This only adds 70kg to each wing-box, which is much more
mass-efficient than an increase in skin thickness. Again, Appendix 2 demonstrates the full buckling
analysis spread-sheet, giving the number of stringers and their dimensions.
To increase the reliability of the structural analysis carried out in MATLAB, Pro-Engineer was used to
conduct a finite-element analysis. As mentioned in Section 3.1, a simple deflection and bending
analysis was carried out on a square wing-box with a thickness of 20mm. This allowed time for
familiarisation with the software, and to help conduct a full wing-box design. The worst-case scenario
caused a deflection of 6.81m for a uniform load, and the Pro-Engineer analysis found a deflection of
6.78m, as shown below in Figure 14:
With a variation of less than 1% on this simple example, Pro-Engineer’s finite element suite was to be
used to verify our more complex analytical methods in MATLAB.
The composed program took the input values given in the design brief, and then calculated the
dimensions within the wing skin at both the wing root and tip. This allowed for a secondary function to
create C-section spar dimensions that fit would within this space. Values for the cross-sectional area
and second moment of area were also calculated, and added to each spars data. The use of
programming allowed for a vast number of varying spar shapes and sizes to be analysed – this would
not have been possible to do manually.
To begin with, the program worked using two identical spars that were mirrored and ran the length of
the wing box with a constant cross-section. The distance between the spars in this analysis varied
15
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
form root to tip, to match the taper of the wing. Spars were initially tested for both direct stress and
shear at the root, whilst the maximum amount of twist was tested at the wing-box tip. These
calculations follow the methods outlined in section 4. With the given criteria, the program omitted all
spars that did not conform to the maximum stress value, and then would output the lightest possible
solution. At this point the chosen spars were then analysed in more depth to calculate deflection,
shear stress and twist, at intervals from root to tip.
Figure 15 – Initial spar refinement stage using spars with a uniform section
As can be seen above in Figure 15, the chosen spars met all the initial criteria, and experienced very
little deflection and twist, but they are far from ideal in reality. To improve on this design, the spar
dimension would need to change along the wing-box to fit inside the wing. New restrictions limiting the
height of the spars were added to the programme, resulting in spar taper. The function used to create
the spar dimensions was also updated to create a separate set of c-section data based upon these
new restrictions. Once these additions had been made, the spars tapered in height and therefore
continued to fit within the skin.
With the spar cross-sections being generated independently from root to tip, the shape could vary
freely allowing for a much more efficient use of material in the wing box structure. Figure 16 shows
that in although a more light-weight and feasible design had been reached, the deflection and twist
were larger. This refinement made the spars 46% lighter than at the previous stage, saving 1697kg
per wing.
There was still one last adjustment to make to the program which would allow for a complete spar set
that would fit truly inside the wing skin. The restriction of identical spars meant that the spar close to
the trailing edge would not fit inside the dimensions of the wing. Due to the airfoil shape of a wing,
there is far more height at the location of the leading edge spar than at the trailing edge and so using
the same-sized spar for both locations is illogical. By modifying the function that creates the spar data
sets, it was possible to allow for independent sizing of each spar whilst calculating area and second
16
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
By making this final modification to the program, completely independent cross section combinations
could be analysed to make for the best possible solution. Again, this reduced the weight, by 400kg,
but caused 24% more deflection and 7% more twist. Despite this increase, both are still of an
acceptable magnitude. A maximum twist of 1.5°, which only occurs in the absolute worst-case
scenario and will therefore not be prolonged, would not have a large effect on the aerodynamic
performance of the wing.
Figure 17 – Final modification to the spars showing wing deflection and twist
To further refine the wing-box design, the thickness of the wing skin was taken into consideration. The
calculations made so far were made using an arbitrary wing-skin thickness of 3mm, and show that the
skin mass is a large contributor to the overall wing-box mass. By reducing this value to 1mm, a large
weight loss could be achieved but this resulted in a large increase in twist as well as a moderate rise
in shear and direct stress. The deflection value also changed here but only by a small degree.
Figure 18 – Analysis of the wing box with alternate skin thickness of 1mm
The value for twist here is considerable so analysis for this wing box under straight and level
conditions was used to evaluate how the wing would act during flight. Figure 19 shows that the wing
twist returns to a very low value of 0.6° when under normal conditions.
17
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
Figure 19 – Analysis of the wing box under straight and level conditions
60 200
60 120 20
450 20
240
Root
30 100
30 80 10
250 10
120
Tip
Using the MATLAB programme, a set of spars have been designed for the wing-box, and the skin
thickness has been determined. The programme allowed a minimum-mass solution to be achieved
using numerical methods, and has output a set of dimensions for the spars. To validate this method,
as mentioned above, finite element analysis can now be done using the dimensions calculated.
Due to the size and shape of the wing box, the ratios between dimensions are very high, which
resulted in some difficulty running a finite-element approach. To create the mesh for the thin
structures, a small element size was required and this, when used along the entire length of the
structure, leads to the analysis calculating for such a large number of points that it becomes
unfeasible. Therefore, a shell idealisation technique was used within Pro-Engineer to solve for the
surfaces of the wing box. Once a material and thickness was applied to each surface, a much faster
and more feasible analytical solution was possible.
For the analysis, the root was constrained and the distributed force was added along the 25% chord
18
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
line. This force was interpolated from root to tip to replicate the triangular loading assumptions from
the numerical analysis.
The maximum deflection calculated by Pro-Engineer for the wing box design was 0.96 metres. In
comparison, the calculated value in MATLAB was 0.81m, giving a difference of 15.6%. The error
between these results can be explained by the assumptions made in using the shell idealisations
within the Pro-Engineer analysis. To create the shells, average thicknesses along the wing-box had to
be used for each case, rather than including the design’s varying dimensions. This simplification
resulted in smaller spars close to the root than the actual design, hence producing slightly larger
stress and wing deflection values.
The principle stress was also analysed by Pro-Engineer, and the values displayed in Figure 23 help to
reinforce the explanation of the variation between the deflection values. The maximum principle stress
is much higher at 370MPa compared to 300MPa, and this will cause the extra deflection.
A 15% error in deflection between these results is satisfactory to validate the numerical methods used
in MATLAB, when taking into consideration the approximations used in defining the mesh used in Pro-
Engineer. Although the stress in the finite-element analysis is considerably higher than that calculated
in the numerical analysis, and also slightly higher than the yield stress of the aluminium alloy, the fact
that these values only occur in areas that are underrepresented, by up to half the volume of material
through using the shell approximation method, restores confidence and reliability in the design.
19
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
Figure 23 – Principles stress of the wing box by FE analysis using Pro-E with the view from below
This wing-box design combines maximum weight-effectiveness with multiple analytical methods that
fully demonstrate the ability of the design to safely handle the aircraft’s worst-case loading scenario,
whilst retaining structural integrity and experiencing minimal distortion to the wing’s important
aerodynamic profile.
References
eFunda, I. (2011). Euler Bernoulli Beam Equation. Retrieved November 23, 2011, from
efunda: The Online Reference for Engineers:
http://www.efunda.com/formulae/solid_mechanics/beams/theory.cfm
Megson, T. H. (1999). Aircraft Structures for Engineering Students. Oxford: Butterworth-
Heinemann.
Vanovschi, V. (2006). Online Integral Calculator. Retrieved November 27, 2011, from
NumberEmpire.com: http://www.numberempire.com/integralcalculator.php
6.0 Appendices
Appendix 1 - MATLAB code
Appendix 2 - Excel spread-sheet
20
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
% * * * * Inputs * * * *
% Box materials
maxs=350E6; % Max stress MN
E = 70E9; % Youngs Modulus
G = 27E9;
dens=2700; % Density
% Wing data
RootC=4800; % Root chord
TipC=3600; % Tip chord
% Box dimensions
mDlr=10*floor(((RootC*RootT)/1.05)/10); % max spar hieght at
root/leading mm
mDtr=10*floor((mDlr/2.2)/10); % max spar hieght at
root/trailing mm
% Loads
m=74000; % mass
slf=4; % straight and level factor
w=m*slf*10; % total load at gust (n)
w2=(w/2); % loading per wing (n)
q=w2/L; % distributed load (n/m)
21
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
[tdata,rdata]=DataC( mDlr,mDtr,mDlt,mDtt,Wr,Wt,ts);
sds=size(dsdata,1);
p=0;
for i=1:1:sds
t=dsdata(i,3);
D=dsdata(i,1);
hav=(dsdata(i,1)+dsdata(i,4))/4;
d=D-(2*t);
b=dsdata(i,2);
B=b+t;
ixx=dsdata(i,8);
% Shear flow
Ba=ixx/(4*hav^2); % Point area for boom
dq1=(-Ba*q*hav)/ixx;
q0=q/(8*hav);
qmax=dq1+q0;
% Rate of twist
tw=((wr)+2*hav)/(2*(wr*2*hav)*G*ts) * (q/(8*(wr/2)*hav) *
(4*Ba*(wr/2)*hav^2)/ixx);
dsdata(i,10)=-qmax;
22
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
sdsdata=zeros(p,10);
p=0;
for i=1:1:sds
if dsdata(i,10)<=maxs
p=p+1;
sdsdata(p,:)=dsdata(i,:);
end
end
ssds=size(sdsdata,1);
rArea=100; % Random large number
for i=1:1:ssds
if sdsdata(i,7)<=rArea
rArea=sdsdata(i,7);
root=sdsdata(i,:);
end
end
sdt=size(tdata,1);
p=0;
for i=1:1:sdt
t=tdata(i,3);
D=tdata(i,1);
hav=(tdata(i,1)+tdata(i,4))/4;
d=D-(2*t);
b=tdata(i,2);
B=b+t;
ixx=tdata(i,8);
% Shear flow
Ba=ixx/(4*hav^2); % Point area for boom
dq1=(-Ba*q*hav)/ixx;
q0=q/(8*hav);
qmax=dq1+q0;
dtdata=zeros(p,10);
p=0;
for i=1:1:sdt
23
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
if tdata(i,9)<=maxs
if tdata(i,10)<=13 % Twist limit
p=p+1;
dtdata(p,:)=tdata(i,:);
end
end
end
sdt=size(dtdata,1);
rArea=100; % Random large number
for i=1:1:sdt
if dtdata(i,7)<=rArea
rArea=dtdata(i,7);
tip=dtdata(i,:);
end
end
% * * * * * * * * * * * * * * * * * * *
% Analysis/Plotting of selected spars
% * * * * * * * * * * * * * * * * * * *
iters=16;
dz=L/iters;
% Spar Dimensions
t1=root(3);
t2=tip(3);
dt1=(t1-t2)/iters;
t3=root(6);
t4=tip(6);
dt2=(t3-t4)/iters;
D1=root(1);
D2=tip(1);
dD1=(D1-D2)/iters;
D3=root(4);
D4=tip(4);
dD2=(D3-D4)/iters;
d_1=D1-2*t1;
d_2=D2-2*t2;
d_3=D3-2*t3;
d_4=D4-2*t4;
b1=root(2);
b2=tip(2);
db1=(b1-b2)/iters;
b3=root(5);
b4=tip(5);
db2=(b3-b4)/iters;
24
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
B_1=b1+t1;
B_2=b2+t2;
B_3=b3+t3;
B_4=b4+t4;
w1=Wr/1000/2;
w2=Wt/1000/2;
dw=(w1-w2)/iters;
uspar1=zeros(9,3,iters+1);
uspar2=zeros(9,3,iters+1);
for i=1:1:L+1
D=D1-(i-1)*dD1;
b=b1-(i-1)*db1;
t=t1-(i-1)*dt1;
x=w1-(i-1)*dw;
uspar1(1,1:2,i)=[x,0.5*(D1-D)];
uspar1(2,1:2,i)=uspar1(1,1:2,i)+[0,D];
uspar1(3,1:2,i)=uspar1(2,1:2,i)+[-(b+t),0];
uspar1(4,1:2,i)=uspar1(3,1:2,i)+[-0,-t];
uspar1(5,1:2,i)=uspar1(4,1:2,i)+[b,0];
uspar1(6,1:2,i)=uspar1(1,1:2,i)+[-t,t];
uspar1(7,1:2,i)=uspar1(6,1:2,i)+[-b,0];
uspar1(8,1:2,i)=uspar1(7,1:2,i)+[0,-t];
uspar1(9,1:2,i)=uspar1(1,1:2,i);
uspar1(:,3,i)=i-1;
end
for i=1:1:L+1
D=D3-(i-1)*dD2;
b=b3-(i-1)*db2;
t=t3-(i-1)*dt2;
x=w1-(i-1)*dw;
uspar2(1,1:2,i)=[-x,0.5*(D1-D)];
uspar2(2,1:2,i)=uspar2(1,1:2,i)+[0,D];
uspar2(3,1:2,i)=uspar2(2,1:2,i)+[(b+t),0];
uspar2(4,1:2,i)=uspar2(3,1:2,i)+[0,-t];
uspar2(5,1:2,i)=uspar2(4,1:2,i)+[-b,0];
uspar2(6,1:2,i)=uspar2(1,1:2,i)+[t,t];
uspar2(7,1:2,i)=uspar2(6,1:2,i)+[b,0];
uspar2(8,1:2,i)=uspar2(7,1:2,i)+[0,-t];
uspar2(9,1:2,i)=uspar2(1,1:2,i);
uspar2(:,3,i)=i-1;
end
% uspar1(:,2,:)=uspar1(:,2,:)D1-0.05;
% uspar2(:,2,:)=uspar2(:,2,:)D1-0.05;
hold on
25
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
set(gcf,'renderer','opengl')
fc=[0.4 0.4 0.4]; % colour
uspar1f=zeros(4,3,8);
uspar2f=zeros(4,3,8);
for i=1:1:L
for j=1:1:8
uspar1f(1,:,j)=uspar1(j,1:3,i);
uspar1f(2,:,j)=uspar1(j+1,1:3,i);
uspar1f(3,:,j)=uspar1(j+1,1:3,i+1);
uspar1f(4,:,j)=uspar1(j,1:3,i+1);
h1=fill3(uspar1f(:,3,j),uspar1f(:,1,j),uspar1f(:,2,j),fc*2);
set(h1,'edgecolor','none');
uspar2f(1,:,j)=uspar2(j,1:3,i);
uspar2f(2,:,j)=uspar2(j+1,1:3,i);
uspar2f(3,:,j)=uspar2(j+1,1:3,i+1);
uspar2f(4,:,j)=uspar2(j,1:3,i+1);
h2=fill3(uspar2f(:,3,j),uspar2f(:,1,j),uspar2f(:,2,j),fc*2);
set(h2,'edgecolor','none');
end
end
h1=fill3(uspar1(:,3,1),uspar1(:,1,1),uspar1(:,2,1),fc*2);
set(h1,'edgecolor','none');
h2=fill3(uspar1(:,3,L+1),uspar1(:,1,L+1),uspar1(:,2,L+1),fc*2);
set(h2,'edgecolor','none');
h3=fill3(uspar2(:,3,1),uspar2(:,1,1),uspar2(:,2,1),fc*2);
set(h3,'edgecolor','none');
h4=fill3(uspar2(:,3,L+1),uspar2(:,1,L+1),uspar2(:,2,L+1),fc*2);
set(h4,'edgecolor','none');
alpha(0.1)
ixxr=root(8);
ixxt=tip(8);
k=(ixxr-ixxt)/(iters);
% constant c
f=-q/(6*L*E*k) * ((-ixxt^3*log(16*k+ixxt)/(k^3)) + (16*ixxt^2)/(k^2) -
((16^2*ixxt)/(2*k)) + (16^3)/3);
fL=f*L;
% constant d
g=(-q*(((-6*ixxt^4)*log(ixxt+(16*k)))/k - (96*ixxt^3*log(ixxt+16*k)) +
(96*ixxt^3) + (3*ixxt^2*k*(16^2)) - (ixxt*(k^2)*(16^3)) +
((k^3)*(16^4))/2))/(36*L*E*k^4)-fL;
26
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
for i=0:1:L
z=i;
fz=f*z;
y1=(q/(36*L*E*k^4)) * (((-6*ixxt^4)*log(ixxt+k*z))/k -
(6*ixxt^3*z*log(ixxt+k*z)) + 6*ixxt^3 + 3*ixxt^2*k*z^2 - ixxt*k^2*z^3 +
(k^3*z^4)/2);
y=y1+fz+g;
yi(i+1)=y;
zi(i+1)=z;
xi(i+1)=0;
end
miny=min(yi);
if miny<0
yi=yi-miny;
end
% Shear flow
Ba=ixx/(4*hav^2); % Point area for boom
dq1=(-Ba*q*hav)/ixx;
q0=q/(8*hav);
tau(i+1)=-(dq1+q0)/ts;
end
theta=-theta*180/3.14;
27
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
% Add rotation
alpha(i)=atand(yi(L+1-i)/i);
Rx=[1,0,0; 0, cosd(alpha(i)),-sind(alpha(i)); 0,
sind(alpha(i)), cosd(alpha(i))]; % Rotational matrix
for j=1:9
lspar1(j,:,i+1)=lspar1(j,:,i+1)*Rx; % Add twist to leading edge
lspar2(j,:,i+1)=lspar2(j,:,i+1)*Rx; % Add twist to trailing
edge
end
end
h1=fill3(lspar1f(:,3,j),lspar1f(:,1,j),lspar1f(:,2,j),fc*2);
set(h1,'edgecolor',fc);
lspar2f(1,:,j)=lspar2(j,1:3,i);
lspar2f(2,:,j)=lspar2(j+1,1:3,i);
lspar2f(3,:,j)=lspar2(j+1,1:3,i+1);
lspar2f(4,:,j)=lspar2(j,1:3,i+1);
h2=fill3(lspar2f(:,3,j),lspar2f(:,1,j),lspar2f(:,2,j),fc*2);
set(h2,'edgecolor',fc);
end
end
h1=fill3(lspar1(:,3,1),lspar1(:,1,1),lspar1(:,2,1),fc*2);
set(h1,'edgecolor','none');
h2=fill3(lspar1(:,3,L+1),lspar1(:,1,L+1),lspar1(:,2,L+1),fc*2);
set(h2,'edgecolor','none');
h3=fill3(lspar2(:,3,1),lspar2(:,1,1),lspar2(:,2,1),fc*2);
set(h3,'edgecolor','none');
h4=fill3(lspar2(:,3,L+1),lspar2(:,1,L+1),lspar2(:,2,L+1),fc*2);
set(h4,'edgecolor','none');
% * * * * * * * * * *
% Wing overlay
% * * * * * * * * * *
wing=n12u;
for i=1:1:sn12-1
wing(i+sn12,:)=[wing(sn12-i,1),-wing(sn12-i,2)];
end
28
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
wingm=zeros(31,3,L+1);
wingm(:,:,1)=wingr;
plot3(wingm(:,3,1),wingm(:,1,1),wingm(:,2,1))
h7=fill3(wingm(:,3,1),wingm(:,1,1),wingm(:,2,1),fc*2);
set(h7,'edgecolor','none')
for i=2:1:L+1
wingm(:,:,i)=wingm(:,:,i-1)-dwp;
end
wings=wingm; % copy for unloaded
for i=2:1:L+1
% Add twist
Rz=[cosd(theta(i-1)),-sind(theta(i-1)),0; sind(theta(i-1)),
cosd(theta(i-1)), 0; 0, 0, 1]; % Rotational matrix
for j=1:31
wingm(j,:,i)=wingm(j,:,i)*Rz;
end
Rx=[1,0,0; 0, cosd(alpha(i-1)),-sind(alpha(i-1)); 0,
sind(alpha(i-1)), cosd(alpha(i-1))]; % Rotational matrix
for j=1:31
wingm(j,:,i)=wingm(j,:,i)*Rx;
end
end
h8=fill3(wingm(:,3,17),wingm(:,1,17),wingm(:,2,17),fc*2);
set(h8,'facealpha',1);
set(h8,'edgecolor','none')
h1=fill3(wingf(:,3,j),wingf(:,1,j),wingf(:,2,j),fc*2);
set(h1,'facealpha',0.5);
set(h1,'edgecolor','none');
end
29
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
end
h2=fill3(wingfs(:,3,j),wingfs(:,1,j),wingfs(:,2,j),fc*2);
set(h2,'facealpha',0.1);
set(h2,'edgecolor','none');
end
end
% Calculate wieghts
% Spar weight
s1a=((D1*B_1-d_1*b1)+(D2*B_2-d_2*b2))/2;
s1v=s1a*L;
s1m=s1v*dens;
s2a=((D3*B_3-d_3*b3)+(D4*B_4-d_4*b4))/2;
s2v=s2a*L;
s2m=s2v*dens;
sm=s1m+s2m;
% Skin weight
avC=(RootC+TipC)/2/1000;
skv=2*avC*L*ts;
skm=skv*dens*1.05; % +5% mass as estimation
sigma=root(1,9)/1000000;
taumax=max(tau)/1000000;
txt1=text(L*0.95,0,L/5,str1);
set(txt1,'HorizontalAlignment','center','FontName','calibri','color',0.75*f
c)
txt2=text(L*0.1,0,L/5,str2);
set(txt2,'HorizontalAlignment','center','FontName','calibri','color',0.75*f
c)
txt3=text(L*0.5,0,-L/7,str3);
set(txt3,'HorizontalAlignment','center','FontName','calibri','color',0.75*f
c)
30
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
axis equal
axis vis3d
axis off
grid on
camlight
lighting gouraud
i=0;
data1=zeros(titers*biters*diters,5);
for Dm=Dlm:deltad:mDlr
for tm=tmin:deltat:tmax
for bm=bmin:deltab:bmax
i=i+1;
D=Dm/1000;
t=tm/1000;
b=bm/1000;
B=(b+t);
d=(D-2*t);
A=B*D-b*d;
ix=2*(B*D^3-b*d^3)/12;
data1(i,1)=D;
data1(i,2)=t;
data1(i,3)=b;
data1(i,4)=A;
data1(i,5)=ix;
end
end
end
31
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
biters=(bmax-bmin)/deltab+1;
i=0;
data2=zeros(titers*biters*diters,5);
for Dm=Dtm:deltad:mDtr
for tm=tmin:deltat:tmax
for bm=bmin:deltab:bmax
i=i+1;
D=Dm/1000;
t=tm/1000;
b=bm/1000;
B=(b+t);
d=(D-2*t);
A=B*D-b*d;
ix=2*(B*D^3-b*d^3)/12;
data2(i,1)=D;
data2(i,2)=t;
data2(i,3)=b;
data2(i,4)=A;
data2(i,5)=ix;
end
end
end
i=0;
data3=zeros(titers*biters*diters,5);
for Dm=Dlm:deltad:mDlt
for tm=tmin:deltat:tmax
for bm=bmin:deltab:bmax
i=i+1;
D=Dm/1000;
t=tm/1000;
b=bm/1000;
B=(b+t);
d=(D-2*t);
A=B*D-b*d;
ix=2*(B*D^3-b*d^3)/12;
data3(i,1)=D;
data3(i,2)=t;
data3(i,3)=b;
data3(i,4)=A;
data3(i,5)=ix;
end
end
end
32
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
i=0;
data4=zeros(titers*biters*diters,5);
for Dm=Dtm:deltad:mDtt
for tm=tmin:deltat:tmax
for bm=bmin:deltab:bmax
i=i+1;
D=Dm/1000;
t=tm/1000;
b=bm/1000;
B=(b+t);
d=(D-2*t);
A=B*D-b*d;
ix=2*(B*D^3-b*d^3)/12;
data4(i,1)=D;
data4(i,2)=t;
data4(i,3)=b;
data4(i,4)=A;
data4(i,5)=ix;
end
end
end
sd1=size(data1,1);
sd2=size(data2,1);
sd3=size(data3,1);
sd4=size(data4,1);
rdata=zeros(sd1*sd2,7);
p=1;
for i=1:1:sd1
for j=1:1:sd2
rdata(p,1)=data1(i,1); %D1
rdata(p,2)=data1(i,3); %b1
rdata(p,3)=data1(i,2); %t1
D1=data1(i,1); %D1
b1=data1(i,3); %b1
t1=data1(i,2); %t1
rdata(p,4)=data2(j,1); %D2
rdata(p,5)=data2(j,3); %b2
rdata(p,6)=data2(j,2); %t2
rdata(p,7)=data1(i,1)+data2(j,2); %A
D2=data2(j,1); %D2
33
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
b2=data2(j,3); %b2
t2=data2(j,2); %t2
d1=t1;
d2=t2;
B1=t1;
B2=t2;
y1=(D1-t1)/2;
y2=(D2-t2)/2;
ixxsp=(B1*D1^3)/12+(B2*D2^3)/12+2*(b1*d1/12+b1*d1*y1^2)+2*(b2*d2/12+b2*d2*y
2^2);
ixxsk=2*(wr*ts*((D1+D2)/2)^2);
rdata(p,8)=ixxsp+ixxsk;
p=p+1;
end
end
tdata=zeros(sd3*sd4,8);
p=1;
for i=1:1:sd3
for j=1:1:sd4
tdata(p,1)=data3(i,1); %D1
tdata(p,2)=data3(i,3); %b1
tdata(p,3)=data3(i,2); %t1
D1=data3(i,1); %D1
b1=data3(i,3); %b1
t1=data3(i,2); %t1
tdata(p,4)=data4(j,1); %D2
tdata(p,5)=data4(j,3); %b2
tdata(p,6)=data4(j,2); %t2
tdata(p,7)=data3(i,1)+data4(j,2); %A
D2=data4(j,1); %D2
b2=data4(j,3); %b2
t2=data4(j,2); %t2
d1=t1;
d2=t2;
B1=t1;
B2=t2;
y1=(D1-t1)/2;
y2=(D2-t2)/2;
ixxsp=(B1*D1^3)/12+(B2*D2^3)/12+2*(b1*d1/12+b1*d1*y1^2)+2*(b2*d2/12+b2*d2*y
2^2);
ixxsk=2*(wt*ts*((D1+D2)/2)^2);
tdata(p,8)=ixxsp+ixxsk;
p=p+1;
end
end
34
AERO318 – DESIGN OF A MINIMUM MASS WING-BOX December 2011
35