Escolar Documentos
Profissional Documentos
Cultura Documentos
Daniel Rose
S1203681
16/03/2015
Problem 1
A frame structure was created in GSA software suit by mapping out nodes connected with
beam elements. The dimensions of the structure can be seen in figure 1 alongside an
isometric view of the frame modelled in GSA. The beams are colour coded in figure one to
highlight the different geometries of the beam sections. The first section of this report will
display the analysis results and comment on how they can be used in the design process.
Figure 1. Dimensions of the frame (left), pictorial view of GSA model (right)
1a.
Table 1 shows the global stiffness matrix for the whole structure, K, and its partitions Kff,
Krf, Kfr and Krr. By inspection of the frame there are 15 degrees of freedom and 12 reaction
forces.
Matrix
K
Kff
Krf
Kfr
Krr
size
27 x 27
15 x 15
12 x 15
15 x 12
12 x 12
K is an nxn matrix where n equals the total possible reaction forces on the structure
if the structure was fully constrained.
Kff is an nxn matrix where n represents the total number of DOFs
Krf is the matrix used to find the reaction forces from the DOFs
Kfr is the matrix used to find the displacements from the reaction forces
Krr is an nxn matrix where n represents the total number of reaction forces
The local stiffness matrix for member four was calculated and can be seen in table 2.
V44
P4
M44
P44
V45
M45
P45
V4
0
12
3
6
2
0
12
3
6
2
M4
0
6
2
4
6
2
2
P5
V5
0
12
3
6
2
0
12
3
6
2
M5
0
6
2
2
6
2
4
1b.
The points of 0 moment were found on the bending moment diagram by inspection. It was
simply where the moment graph crossed the beams local x axis. These points were
highlighted with black circles and shown in figure 2 for visualisation. A more accurate
location of the points of zero moment is shown in table 3. This was achieved using GSAs
moment output function. Halfway between the points where the bending moment changed
sign were approximated as the points of zero moment. The percentage distance along the
beam in the output was converted to meters.
Table 3. Points of zero moment across the frame
Member
1
3 (i)
3 (ii)
4
6 (i)
6 (ii)
7
X coordinates (m)
0
0.36
6.376
8
1.99
5.73
34.30
Y coordinates (m)
2.128
5
5
1.955
5
5
2.635
% of individual member
28.2
4.4
79.7
60.9
33.2
95.5
58.9
Figure 3. Full bending moment diagram for full factored load with maximum values shown in kNm
To obtain the locations of the maximum moments along the span, elements 3 and 6 were
analysed. This was achieved through GSAs output function. The desired output was set to
Beam and Spring Forces and Moments to show the results on beams 3 and 6 for the full
factored load case. The locations of diplacement were initially given as a percentage
distance along a member, these were converted to a displacement using the beam lengths
which were known. Table 4 sumarises the results. Because the programe could reference a
maximum of 100 points the error in the displacement value was 0.5%. This was considered
reasonable since a 4cm can easily be analysed for defects if required.
Element
3 +ve
3 ve
6 +ve
6 -ve
Max Moments
417.6
-315.4
387.5
-120.9
1c.
Beam 3
Beam 6
Figure 4. Deflected shape diagram showing the maximum displacements for beams 3 and 6.
Element
3
6
Max displacement
0.002796
0.011520
Location
3.684m from node 5
3.640m from node 2
The exaggerated deflected shape show in figure 5 has been annotated with black circles to
show the frames points of contraflexure. A point of contraflexure is a point in a beam where
no bending occurs. This will be when the rate of change of gradient transitions from positive
to negative. These points are at the same locations as the points of zero moment on the
bending moment diagram so the locations are also shown in table 3. To summarise, when
the moment diagram is positive the rate of change of the displacement diagram curve will be
negative. When the moment diagram is negative the rate of change of the displacement
curve will be positive.
Beam 3
Beam 6
1d.
15.50E-6
3.100E-6
-9.714
-14.46
Values shown are the shear maxima for each member (N)
Figure 6. Full shear force for the full factored load case to show maximum member values.
Figure 6 shows that the sign of the shear force diagram is directly related to the sign of the
bending moment diagrams gradient. The shear force diagram for beams 1, 4 and 7 are
straight lines and do not change in sign or magnitude. Looking at the bending moment
diagram in figure 5 the bending moment gradient is constant. Beams 3 and 6 have uniformly
distributed loads so the bending moments gradient changes in sign and magnitude along
the length. We can see that the shear force diagram has a constant gradient but its
magnitude changes along the length with a sign change corresponding to a peak on the
moment diagram.
Values shown are the axial force maxima for each member
(N)
Figure 7. Full axial force diagram for the full factored load case
The axial force along the length of all beams is constant since when a static system is at
equilibrium the applied forces and reaction forces do not change. The applied forces are
transferred as a constant down to the supports.
1e.
To analyse the beam for the worst possible loading scenarios three cases additional to the
full factored load were created. These are summarised in figure 9. The dead load and full
factored load was given in the problem description, the active load was simply the full minus
the dead load.
Dead load
DL= 25.8kN/m
Active load
AL= 42.32kN/m
Beam 3
Beam 6
Figure 8. Beam and node location across the span with associated numbering
Full Factored
Partial Left
Partial Right
Dead Load
Each of the cases were analysed using GSA. The output function was utilised to extract
results from the analysis. Results for the maximum midspan moments, support moments
and support shear stresses were noted then tabulated in tables 6, 7 and 8. The numbers in
bold in these tables represent the maximum value for each study.
Table 6. Showing Max moment at midspan of 3 and 6 for full factored load
i.
Load Case
Dead
Beam 3 (kNm)
-114.1
Beam 6 (kNm)
-36.06
Max. moment at
midspan
Full Factored
-301.4
-95.21
Partial Left
-335.1
-38.39
Partial Right
-80.41
-169.7
ii.
Load case
Full Factored
Node 2 (kNm)
-69.21
Node 5 (kNm)
29.60
-70.98
51.60
Partial Right
-24.45
-10.8
Load case
Full Factored
Node 5 (kNm)
-69.21
Node 8 (kNm)
29.60
-70.98
51.60
-24.45
-10.8
iii.
Load case
Full Factored
Max shear at
supports (Right)
Partial left
-237.8
307.2
Partial Right
-77.81
128.6
Load case
Full Factored
Node 5 (kN)
-263.4
Max shear at
supports (Left)
Node 2 (kN)
-228.9
Node 5 (kN)
316.1
Node 8 (kN)
-145.3
Partial left
-124.9
29.85
Partial Right
-238.2
170.5
A bending moment envelope can be easily created in GSA for two or more loading cases.
The envelope gives a quick visualisation of the maximum and minimum bending moments at
every point along the beam. The envelope diagram shown in figure 10 gives this
representation and is annotated with maximum ve/+ve moments and moments at the end
points. The maximum values from this diagram are the maximum from all cases considered
in analysis so for the purposes of bridge design it is a very useful tool. That is if the bridge is
designed to cope with these maximum bending moments it should not fail.
418.0
227.5
388.4
238.3
35.44
70.98
24.45
-5.4
-92.46
-323.7
11.87
-177.7
-343.5
1f.
A fifth scenario of loading during an earthquake was considered for the final part of the
analysis. This loading case is shown in figure 11. There were UDL loads across beams 3
and 6 and unlike the other cases there were Point loads acting along the x axis. These point
loads will cause greater horizontal reaction forces at the base and a different distribution and
magnitude of shear forces and bending moments across the frame. These will be discussed
in the following section.
The bending diagram in figure 2 shows the different magnitude and distribution when
compared to figure 3 in section 1b. We can see that the maximum bending moments along
the span and at the end points have increased. The moments along the vertical components
have also increase dramatically. Particularly in beams 2, 5, and 8 where they were
approximately 0Nm before. This changes the strength requirements of the top beams since
for the vertical loads they did not need to be considered as load bearing elements. For the
bottom vertical beams 1, 4 and 7 the maximum bending moment is increased by as much as
1600%. This means a cross section with larger outer dimensions would be required for this
case than for previous loading if the bridge was being designed efficiently.
581.6
539.2
200.0
15.43
200.0
200.0
70.98
-381.6
-468.9 -526.3
429.4
385.8
315.1
Figure 12. Full bending moment diagram for earthquake load case
The shear force diagram shown in figure 12 shows that the shear force in the vertical section
also increases, by as much as 1400%. The greater shear force means that there would be a
greater cross sectional area in the vertical beams since the shear force varies inversely with
area, equation 1.
, =
-100.0
1.
-100.0
-100.0
-138.9
10.1
164.9
406.1
-162.3
-199.8
-137.9
Figure 13. Full shear force diagram for earthquake load case
Problem 2
To approach the problem a set of geometric data had to be calculated for the bridge
structures. The basic design of the bridge is shown in figure 14. The calculation procedure
can be found in the Appendix.
There were three different curve forms to compute and for each, three different basic
parameters. That gives nine cases in total and these are highlighted below in table 9.
Table 9. Highlighting the different geometric curves
L = 50m, constant
Sin curve
Parabola
Circle arc
H1
L/2
L/2
L/2
H2
L/4
L/4
L/4
H3
L/8
L/8
L/8
The bridge was designed to be a pedestrian/cycle bridge going over a road or river - many
bridges similar to this are currently in operation, for example a bridge in Berkley [1]. The
Bridge deck length, L, is 50m and width is 3m in accordance with the Highways Agency [2].
The deck would be 0.3m thick of concrete. Concrete is widely used as a deck material due to
its wide availability, low cost and ease of manufacture.
The load used in designing the bridge was considered to be worst case. Firstly the dead load
was calculated using the volume multiplied by the density of standard density concrete,
2400kg/m2 [3].
The maximum load case was considered to be a crowd of people bunched together with
very little space for walking, a common scene after a football game in a large city. The
dimensions of the pedestrian were taken as the 95 th percentile man to allow some room. The
weight was taken as a 50th percentile man as an average. From Anthropometric tables each
person would occupy 0.1484m2 and weigh 90kg [4]. The bridge deck area is known so the
live load can be calculated as the number of people that would fit on the deck multiplied by
the weight of each person.
There was assumed to be two arches supporting the deck using 17 loading points on each.
This meant the total load from the dead load and live load could be halved. This was
calculated as 53.72kN, calculations can be seen in the appendix.
The material for the cross section was chosen in line with BS EN 10025 which is a British
standard code for practice in the design of bridges and use of materials [5]. S275 steel
(275MPa yield strength [6]) sheet welded into a hollow rectangular section was chosen in
accordance with the standard. To match this design decision a user designed material was
created in GSA, see appendix for a screenshot.
The cross section dimensions were chosen by a method of trial and error. The case of the
L/2 was found by testing to give the largest displacement and stress for constant load, cross
section, materials. This was used as a test case to find an appropriate cross section, a
dimensioning reference is shown in figure 15. The cross section was initially taken as 0.5mW
x 0.5mD x 5mmt x 5mmT. The W and D values where take as an estimation from looking at
similar bridges and the 5mm thickness is the minimum required by the standards for
highways.
Figure 15. Dimensioning parameters for the beam cross section from GSA
The analysis was started with an aim of getting a safety factor of 1.5. This is recommended
where structures are not under extreme loading conditions and the materials are reliable.
The loading for a footbridge is predictable and I have chosen a worst case scenario for this.
The materials used are in line with British standards so by definition will be reliable if
responsibly sourced. The safety factor was calculated using equation 2.
2.
After several trials for the L/2 circular arc case a cross section of 0.8mW x 0.6mD x 12mmt x
12mmT was chosen. This gave a stress safety factor of 1.55 which was suitable for the
required 1.5.
The above cross section gave a deflection of L/312.5 where L is the length of the bridge.
The width and depth dimensioning lends itself to minimising deflection since when the
moment of area is calculated there is a cubed term which relates directly to the width or
depth, depending on the inertial axis.
I have kept the cross section constant for all three curve geometry cases (sin, parabola and
circle) since the analysis on each case for the three bridge heights is purely relative. That is
that every case for curve geometry does not have to be optimised in order to give an
accurate comparison. The final outcome being which curve geometry is the most efficient for
each of the cases of H.
Efficiency Criteria
For judging the efficiency of the bridge types a standard set of criteria were adopted. The
criteria aimed to analyse three different desirable features of the bridge; the structural
efficiency, strength, and its stiffness. These would then be combined together to give an
overall Efficiency rating. This section will take each of these in turn and explain how and why
they were analysed.
Strength
The main criteria for strength was how much load the arch could take without breaking. This
value was computed to compare the three geometric curve cases for each bridge height.
The cross section a material of the arch was held constant. A safety factor calculation
method was adopted using the yield strength of S275 steel, (275MPa) and the max stress
from the case being analysed. The Von Mises Stress is often taken as the maximum stress
that will occur in a cross section under load. This is true providing the material has not
become brittle. For this study it was assumed that if the Yield strength is larger than the Von
Mises stress the material would not fail [7]. The output tool in GSA was used to obtain values
for the max Von Mises stress for each case. Equation 3 could then be used to calculate a
safety factor.
=
3.
Stiffness
The beam stiffness is defined as a beams resistance to bending. The equation used to
calculate this is stiffness (k) = force/displacement. Since the force on the structures were all
the same a different method was adopted using the maximum allowable displacement of
160mm - as stated in section 2a. This analysis used the difference between allowable
displacement and the maximum displacement as a percentage of the allowable
displacement for each case under loading. This gave an easily workable factor which had a
significant influence when multiplied with the other coefficients to achieve the overall
efficiency rating. This was important since the deflection would be the second most
important factor after strength. Once these requirements are satisfied the design can then be
optimised for structural efficiency.
=
4.
5.
Results
It was considered that the results should be presented in table form to keep them concise
and easily comparable. Table 10 and 11 show a complete set of data found using GSA and
calculated efficiency data using the methods described in the previous section. An example
of the bending moment, shear force and axial force diagram is shown in figure 16. For
completeness the diagrams for all cases are shown in the subsequent section.
L/2
Safety
factor
Efficiency Ratings
Stiffness
Structural
Factor
Factor
5.11
14.60
1.47
0.7867
0.9815
-0.0475
18.957
19.155
20.276
L/4
Sin
Parabola
Circle
581.54
583.07
585.23
18.790
4.724
25.66
-58.26
-23.01
-59.25
4.72
11.95
4.64
0.8826
0.9705
0.8396
14.769
14.872
15.015
L/8
Sin
Parabola
Circle
561.70
562.22
562.40
19.80
13.57
17.32
-73.51
-36.76
-43.48
3.74
7.48
6.32
0.8763
0.9152
0.8918
13.447
13.481
13.493
Table 11. Summary of main results table showing combined efficiency ratings
Bending Moment
Max = -8.532kNm
Curve
L/2
Sin
Parabola
Circle
Combined Efficiency
Rating
11.39
40.20
-0.19
L/4
Sin
Parabola
Circle
15.15
41.90
13.94
L/8
Sin
Parabola
Circle
13.10
27.28
22.46
Shear Force
Max = -0.4668kN
Axial Force
Max = -751.7kN
Figure 16. GSA diagrams for the L/4 Parabolic curve which gave the highest efficiency
Circle L/4
Circle L/8
Parabola L/2
Parabola L/4
Parabola L/8
Sin L/2
Sin L/4
Sin L/8
The circle arch first appeared very poor but as the height of the arch decrease it
performed better. It is possible that for smaller bridge heights this trend would
continue and it may be the best case for these, this could be a topic of further
research.
The sinusoidal curve was the most consistent in its efficiency rating but was never
particularly high.
In conclusion the parabolic curve was found to be much more suitable for bridge arch
geometry design with it having performed the best for all three cases of height. It also
has a more predictable deflection shape and shear force distribution. That is that it
only bends one way and the shear force is unidirectional. This may make designing
of the cross section easier since one side would always be in compression and one
in tension.
References List
Appendix
Problem 1:
Output results for max bending moment
3 A2
-24.20
0.0
316.0
0.0
417.6
0.0
3 A2
41.6%
-24.20
0.0
-2.331
0.0
-315.4
0.0
6 A2
-14.49
0.0
-263.3
0.0
387.9
0.0
64.4%
-14.49
0.0
-0.2477
0.0
-120.9
0.0
61.4%
0.001875
0.0 -0.002075
0.002796
45.5% 0.001892
0.0 -0.01136 0.01152
0.0 0.001762
0.0 -0.001243
0.0 0.001762
0.0
0.001243
15 A1
90.1%
75.95E-6
1 A1
1
0.01083
15 A1
92.1%
223.5E-6
8 A1
9
0.0
1 A1
1
0.01083
4 A1
5
0.01108
1 A1
1
0.01083
4 A1
5
0.01108
0.09185
0.0
0.02285
0.09465
0.0
75.95E-6
0.0
0.0
0.0
0.0
0.0
-0.01083
0.09185
0.0
0.02286
0.09465
0.0
223.5E-6
0.0
0.0
-0.1343
0.1343
0.0
0.0
0.0
0.0
0.0
0.0
0.0
-0.01083
-0.02881
0.0
-0.04910
0.05693
0.0
0.01108
0.0
0.0
0.0
0.0
0.0
-0.01083
-0.02881
0.0
-0.04910
0.05693
0.0
0.01108
9.9%
-0.09185
0.0
0.02285
0.09465
0.0
-75.95E-6
0.0
0.0
0.0
0.0
0.0
-0.01083
0.0
0.0
-0.1343
0.1343
0.0
0.0
0.0
0.0
0.0
0.0
0.0
-0.01083
0.0
0.0
0.0
0.0
0.0
-0.01083
13
0.02881
0.0
-0.04910
0.05693
0.0
-0.01108
0.0
0.0
0.0
0.0
0.0
-0.01083
0.0
0.0
-0.1343
0.1343
0.0
0.0
Minima
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2 A1
75.95E-6
1 A1
0.01083
8 A1
0.0
1 A1
0.01083
1 A1
0.01083
12 A1
0.01108
1 A1
0.01083
8 A1
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
753.4E-6
0.0
-0.01062
0.01064
0.0
860.8E-6
0.0
0.0
0.0
0.0
0.0
184.4E-6
0.0
0.0
0.0
0.0
0.0
184.4E-6
0.0
0.0
-0.01540
0.01540
0.0
0.0
0.0
0.0
0.0
0.0
0.0
184.4E-6
591.4E-6
0.0
-0.007693
0.007715
0.0
910.6E-6
0.0
0.0
0.0
0.0
0.0
184.4E-6
591.4E-6
0.0
-0.007693
0.007715
0.0
910.6E-6
12
-753.4E-6
0.0
-0.01062
0.01064
0.0
-860.8E-6
0.0
0.0
0.0
0.0
0.0
184.4E-6
0.0
0.0
-0.01540
0.01540
0.0
0.0
0.0
0.0
0.0
0.0
0.0
184.4E-6
0.0
0.0
0.0
0.0
0.0
184.4E-6
13
-591.4E-6
0.0
-0.007693
0.007715
0.0
-910.6E-6
0.0
0.0
0.0
0.0
0.0
184.4E-6
0.0
0.0
-0.01540
0.01540
0.0
0.0
Minima
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
11 A1
860.8E-6
1 A1
184.4E-6
8 A1
0.0
1 A1
184.4E-6
1 A1
184.4E-6
12 A1
910.6E-6
1 A1
184.4E-6
9 A1
0.0
Problem 2:
Geometric calculations
The calculation of the bridge geometry was set up by finding the equation of the curve for
each case. These equations were manipulated to a form where y was being calculated as a
function of x, L and H where x is the horizontal coordinate along the beam, L is the total
length of the bridge and H is the height of the bridge. The x coordinates corresponded to the
locations of the 17 nodes which were distributed evenly along the length. Taking node 1 as
the origin coordinate (0, 0) the rest of the nodal coordinates could be calculated. From this
an excel spreadsheet was set up so that the values of H, L and x could be changed as an
input and y displacement was given as an output. Below details the geometric relationships
used for the 3 curve types.
= (
2
)
= (sin (
where = 2 and =
))
X coordinates of the nodes where plugged into this formula to get the y coordinates.
Parabolic equation
General equation: y = ax 2 + bx + c
Initial conditions:
L
2
1.
y = 0 x =
2.
y = 0 x = [0, L]
3.
x=
L
2
y=H
The constants were found by substituting the initial conditions into the equation. Three conditions and
three unknowns meant the equation was solvable.
a=
4H
L2
b=
4H
L
y=
c=0
4H 2
x
L2
Circle equation
h
L
d
R= d+h=
h L2
+
2 8h
Specific equation:
h L2
y= +
x2
2 8h
yrelative = y d
General equation:
2 + 2 = 2
4H
L
Max Loading
Took average weight of man to be = 90kg
Dimensions of a person = 530mm, 280mm = 0.1484m2
Bridge surface area = 50 x 3 = 150m 2
L/2
Sin
Parabola
Circle
Max Moment
-291.000
-2.532
132.100
L/4
Sin
Parabola
Circle
-288.800
-8.532
302.700
L/8
Sin
Parabola
Circle
-297.400
-31.660
-90.720