You are on page 1of 32

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Wed, Dec. 9th 2015

DESIGN AND ANALYSIS OF A


SIMPLY SUPPORTED BEAM
FEA design and testing

Abstract
A simply supported beam designed in CAD, tested and optimized using FEA, and manufactured using CAM

Jeroen Van leeuwen


Student #:100528319

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Table of Contents
Table of Contents ........................................................................................................................... 1
List of Figures ................................................................................................................................. 2
List of Tables .................................................................................................................................. 3
Introduction and Justification........................................................................................................ 4
Conceptual Design ......................................................................................................................... 5
Purpose ....................................................................................................................................... 5
Specifications ............................................................................................................................. 5
Theory ......................................................................................................................................... 6
Finite Element Method .............................................................................................................. 7
Theory vs FEA results comparison: ......................................................................................... 10
Procedure ................................................................................................................................. 10
Beam 1.0 .................................................................................................................................. 11
Beam 1.5 .................................................................................................................................. 12
Beam 1.75................................................................................................................................ 13
Beam 2.0 .................................................................................................................................. 14
Beam Load Indentation ........................................................................................................... 16
Summary of Results ................................................................................................................. 16
Discussion ................................................................................................................................ 17
Manufacturing .............................................................................................................................. 18
Introduction .............................................................................................................................. 18
Purpose ..................................................................................................................................... 18
Procedure ................................................................................................................................. 19
Toolpaths .................................................................................................................................. 19
Verification................................................................................................................................ 21
Discussion ................................................................................................................................ 22
Testing and Validation ................................................................................................................. 23
Introduction .............................................................................................................................. 23
Purpose ..................................................................................................................................... 23
Procedure ................................................................................................................................. 23
Data .......................................................................................................................................... 28
Test Results .............................................................................................................................. 28
Discussion ................................................................................................................................ 30
Conclusions .................................................................................................................................. 31

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

List of Figures
Figure 1: Diagram of Loads and Constraints ............................................................................... 5
Figure 2: NX simulation constraints.............................................................................................. 8
Figure 3: FEA Stress of given beam .............................................................................................. 9
Figure 4: FEA deflection of given beam ........................................................................................ 9
Figure 5: Beam 1.0 model ........................................................................................................... 11
Figure 6: Beam 1.0 Stress .......................................................................................................... 11
Figure 7: Beam 1.0 Deflection .................................................................................................... 11
Figure 8: Beam 1.5 Model ........................................................................................................... 12
Figure 9: Beam 1.5 Stress .......................................................................................................... 12
Figure 10: Beam 1.5 Deflection .................................................................................................. 12
Figure 11: Beam 1.75 Model ...................................................................................................... 13
Figure 12: Beam 1.75 Deflection ............................................................................................... 13
Figure 13: Beam 1.75 stress ...................................................................................................... 13
Figure 14: Beam 2.0 - final model .............................................................................................. 14
Figure 15: Beam 2.0 - overall stress .......................................................................................... 14
Figure 16: Beam 2.0 - High stress areas ................................................................................... 15
Figure 17: Beam point load deformation ................................................................................... 16
Figure 18: Top view of CAD jig fixture ......................................................................................... 18
Figure 19: Right Side view of CAD jig fixture .............................................................................. 18
Figure 20: Top view of jig fixture - locating ................................................................................. 19
Figure 21: Drilled holes IPW ........................................................................................................ 20
Figure 22: 3/8 roughing IPW....................................................................................................... 20
Figure 23: cutter toolpaths .......................................................................................................... 21
Figure 24: Strain gauge location ................................................................................................. 23
Figure 25: Strain Gauge and wire support pad .......................................................................... 24
Figure 26: Strain gauge applied.................................................................................................. 24
Figure 27: beam testing setup dial indicator.......................................................................... 25
Figure 28: Strain gauge wire hookup ......................................................................................... 25
Figure 29: Overall setup ready for testing .................................................................................. 26
Figure 30: Weights in order ......................................................................................................... 27
Figure 31: Data from load testing ............................................................................................... 28

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

List of Tables
Table 1: Design Specifications ...................................................................................................... 5
Table 2: Calculated beam results ................................................................................................. 7
Table 3: Theory vs FEA comparison ............................................................................................ 10
Table 4: Design results summary ............................................................................................... 16
Table 5: Toolpath times ............................................................................................................... 22
Table 6: CAM Tooling ................................................................................................................... 22
Table 7: Calculated stress values ............................................................................................... 28
Table 8: Deflection comparison .................................................................................................. 28

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Introduction and Justification


Finite Element testing using computer software has been available to engineers
since the early 80s. With the improvements in computer hardware and cost, the strength,
availability and use has greatly increased. Current FE software has become user friendly in
most aspects of its use, aiding its popularity.
FE testing works by dividing the part into small, finite segments that are solved
mathematically. The higher the refinement of these segments, the more accurate the
solution becomes. It is this reason that computers are used to run FE testing, the sheer
volume of mathematical calculations possible is infinitely higher than if a human (or group of
humans) were to run through all the calculations.
Computer aided Design [CAD] software has allowed parts to be created virtually. Once a part
has been virtually created, it allows virtual testing through FE. These parts can be tested for
physical (load) stresses, thermal stresses, or both. Once a solution has been solved and any
weaknesses shown though exaggerated models, the designer can either modify the part, or
allow the software to create its own, optimized part. Either step requires the designer to be
able to understand and interpret what the simulation shows, and ensure the final model
accurately represents the solution. Optimization through analysis is called Computer Aided
Engineering [CAE].
Computer Aided Manufacturing [CAM] software ties the whole process of part design to part
production together. Using CAM, the virtual part can create molds, dies, and tool paths to
eliminate the need to create/interpret drawings for production of a part.
Together, all three have revolutionized the manufacturing industry. Virtual assemblies in
CAD can catch interferences not caught when hand drawn. CAE allows for testing and fault
detection before any tooling has been created. Finally CAM creates a direct link between the
optimized part and the manufacturing process, eliminating interpretation of critical
specifications and finding flaws in the manufacturing process (impossible procedures, tool
path violations, casting/ejection problems, etc). As a whole, this allows products that can be
lighter, smaller, stronger and cheaper.

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Conceptual Design
Purpose
The purpose of this project [Beam Design] is to use CAD, CAE, and CAM software on a
simple part to understand all steps of the process. By giving us the freedom to design the
beam while meeting specified criteria, our ability to understand the information presented
by the FEA is critical to achieving the aforementioned specifications. Manufacturing of the
beam ties our CAD and CAM courses together, and the report prepares us for documenting
and writing about our work.

Specifications
Beam
Weight (lbs)
As Given
.4897
Limits
<.2938

Change
60%

Deflection (in)
.00613
.0100

Change
163%

Stress (psi)
6,870
10,000

Change
146%

Table 1: Design Specifications

Design Constraints:

3/8" border thickness, min


Full extrusion type cut-outs only
Min radii of tool radius + .030 for a non-stop toolpath (tool dia = .250")
.125 minimum rib thickness
10.00in wide, 2.00in tall, nominal 1/4 thickness

Material: 6061 Aluminum


As seen in figure 1, a point load of 500lbf in the middle of the beam is distributed to simple
supports mirrored on either end of the beam.

Figure 1: Diagram of Loads and Constraints

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Theory
The 500lbf load as applied to the top center of the beam creates a bending moment,
which can then be used to calculate the maximum stresses within the beam. This bending
moment occurs along a centroidal axis (horizontally through the Moment of Inertia) where
the top half of the beam is put into compression, and the lower portion is put into tension.
The formula to find the max stress, known as the flexure formula is as follows:
=

Where:
M=Bending Moment at point of interest
c=Distance from centroidal axis
I=Moment of inertia
To calculate M, we need to figure out the max shearing force. As this is a simple beam with
equal constraint spacing it can be deducted that the loads on each support are equal at:

2
or 250 each in the vertical direction.
=
= 250 4.75
= .
Moment of inertia is based on the cross sectional shape, which for a simple rectangle is:
3
=
12
= .25
= 2

= 4

= /
(. )(1)
=

=
What this formula shows is that the closer to the centroidal axis the distance becomes, the
lower the stress until the centroidal axis is reached, where zero distance results in a stress
of zero. Therefore the ideal location to remove material in terms of stress would be along the
centroidal axis. This stress can be seen visually in pictures of FEA testing, fig 3(page 8).

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM


Maximum deflection at bottom center can also be calculated, using the formula for a simply
supported beam:
=

3
48

With P=force
L= Length between Points
E=Modulus of Elasticity
I=Moment of Inertia
Therefore:
=

500 9.53

[48 10,000,000 ( 4 )]

= .

Theoretical beam

Stress
7125 psi

Deflection
-5.36 x10-3 in

Table 2: Calculated beam results

Another important property for the purposes of evaluation is Youngs modulus. Youngs
modulus (E) defines the relationship between stress () and strain ().
=

What Youngs modulus allows is the calculation of deflection based on the stress of the part,
provided with the modulus of elasticity for the material used.

Finite Element Method


Von-Mises Stresses
Von-Mises stress calculations are used to approximate to the level of stress on each
element. It uses the distortion of each element from three dimensions (x,y,z) to calculate the
strain, and, using the materials modulus of elasticity, can calculate an equivalent stress
value. Von-Mises stress calculations are best used with ductile materials like most metals
(which also happens to be the most common material used). When brittle materials, like
ceramics, are tested, other methods such as Coulombs Law of Shear stress are typically
more appropriate.

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Areas of High Stress in Bending


Loads which result in bending moments create high stresses on the faces furthest away
from the centroidal axis. As well, the further a point is from a supporting constraint toward
the load also results in higher stresses. This is extremely evident in fig 3: FEA analysis of
given beam.
Other areas of high stress are holes which interfere with the flow of stresses, narrow
columns with limited cross sectional area (CSA), and at any change in CSA normal to the
stress. (ex. where a column meets the perimeter, or another column)

Mesh Size
Mesh size in NX can be an automatically calculated value or a user defined value. Choosing
the correct mesh size is important. Mesh size, or element size, determines the number of
mathematical equations that must be solved to generate an output. A large mesh size
requires few calculations are therefor can be processed quickly. However, this also produces
large steps in stress values from one element to the next, and these large steps lower the
accuracy of the simulation. In the other direction, small mesh size can create more accurate
representations, but results in long/failed compute cycles. The solution to this is to refine
(shrink) the mesh size only in areas of high stress or areas of interest, lowering the compute
time but maintaining the accuracy.
For this project, mesh size was important to creating an equal playing field for comparing
individual beam performance and a Ranking Value. Therefore, a mesh size of .15in was
used except in areas of high stress where refinement was necessary.

Constraints
3 constraints were used in the analysis of the beam (see fig 2).
Two reacting constraints on the lower face of the beam replicating a sliding constraint
these allow the lower face to grow as the beam deflects, as the real test stand allows. (Y + Z
in NX)
The third constraint is purely to limit lateral movement that may be predicted within the FEA
calculations. It is placed horizontally to the vertical force on the top of the beam. (X direction)
This creates a warning in NX software since the constraint could possibly be nullifying the
forces if the constraint was applied incorrectly.
Subdivided faces were used to simplify the creation of the above constraints and loads.
Other methods would also be possible with identical outcomes.

Figure 2: NX simulation constraints

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Visual FEA Representation of given beam:

Figure 3: FEA Stress of given beam

Figure 4: FEA deflection of given beam

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Theory vs FEA results comparison:


Beam
Stress
Deflection

Calculated
7125psi
.00536in

Modelled
6870psi
.006128in

Difference
3.7%
15%

Table 3: Theory vs FEA comparison

The difference in stress between the calculated beam results and the modelled results is
close (3.7%), however the modelled results come in slightly lower. This is possibly due to the
fact FEA errs on the lower deflection values per element calculation, which results in a
slightly lower overall strain, and thus stress.
Deflection, however, has a larger difference (15%). Changing from a CTETRA10 to a
CTETRA4 mesh brings the FEA deflection to .0060in, and increasing to the largest mesh
possible for a .25in beam at .25in mesh, lowers the deflection value to .00585in for a 8.5%
difference.
The hand calculations use the Euler-Bernoulli theory, whereas NX defaults to the
Timoshenko Beam Theory, which takes into account shear deformation and the effects of
rotational inertia.
Based on the direction needed to achieve lower FEA deflection by changing to less accurate
mesh type and size, I would conclude the CTETRA10 mesh with .15 mesh size values to be
more accurate than the hand calculated values for this size of beam.

Procedure
To simulate stresses (forces, pressures, thermal, etc) using NX, the part should be modelled
and open. Once modelled, create new simulation is selected to create the files necessary for
FEA testing. As we use in CADIII, Nastran Design simulation type is selected (for physical
loads like point loads and pressures). Once the files (.fem & .sim) have been created, an
initial mesh is assigned to the part, as well as a material (both in .fem). From there, the
loads and constraints (both in .sim) that best describe how the part will be stressed are set
up. Finally, NX is told to solve the simulation (.sim). After the computer has processed the
FEA requirements, the simulation files of the part can be viewed and interpreted. Colour
coded legends show high deflection, stress, strain and more. Displacement is also
exaggerated for better visual interpretation of the results.

10

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Beam 1.0
Beam 1.0 up to beam 1.75 were not meant to be competitive designs, they were a learning
attempt by trial and error, as well as familiarizing myself with both the FEA solver and
interpreting the results presented.

Figure 5: Beam 1.0 model

11

Figure 6: Beam 1.0 Stress

As can be seen easily in fig 6, stresses under the point load and midway around the arch are
overstressed due to lack of support. Upper arch deflection also high, as fig 7 will show. Base
deflection is minimal. This starts the theory that the wide center triangle provides less
deflection where measured. Part is clearly still overweight.

Figure 7: Beam 1.0 Deflection

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Beam 1.5
Beam 1.5 took lessons from 1.0 and narrowed the center arc, while adding outside triangles
to lower the overall weight. Both weight and stresses lowered but still beyond limits with
beam 1.5, deflection is still well within.

Figure 8: Beam 1.5 Model

12

Figure 9: Beam 1.5 Stress

Beam 1.5 has less red in high stressed areas (fig 9), indicating positive change. Overall,
there is less blue (non-stressed extra weight material) as well. Deflection still acceptable.
Top deflection down 4 thousandths.

Figure 10: Beam 1.5 Deflection

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Beam 1.75
The third iteration of the trial beam was an attempt to get the weight within limits. This beam
ended up having terrible deflection and stress, but gave me the indication to connect the top
outside corners to the lower legs of the arch as will be seen in beam 2.0.

Figure 11: Beam 1.75 Model

The extremely over max deflection as seen in fig 11 below:

13

Figure 12: Beam 1.75 Deflection

Stress values in this beam iteration were also well over limits in multiple locations (fig 13):

Figure 13: Beam 1.75 stress

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Beam 2.0
Beam 2.0 was designed using lessons from beam 1-1.75. The large arch from 1.0 created
two low stress points just away from the point load, followed by two overstressed areas
slightly farther away. To correct this, straight sides with a large joining radius were utilized
for what would now be the main triangle.
The outer areas of the beam were lightened in a manner similar to beam 1.5. The upper
triangles had the bottom raised and the whole profile brought in closer to the middle,
allowed by the change to straight sides in the center triangle. This created room for a lower,
outer triangle to remove unstressed material without causing the issues shown in beam
1.75. [Beam 1.75 with a support strap could also describe beam 2.0]

Figure 14: Beam 2.0 - final model

14

Figure 15: Beam 2.0 - overall stress

What figure 15 shows poorly due to refined mesh areas, is the high stresses on a few of the
internal radii. Figure 16 highlights these areas that took the majority of time to fine tune
below maximum stress levels. What can be seen well in fig 15 is the minimization of the
blue areas (low stress) on the beam, mostly relegated to untouchable geometry. Further
efforts to reduce this might yield positive results.

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

15
Figure 16: Beam 2.0 - High stress areas

As figure 16 shows, where many internal radii meet a flat surface is an area of high stress.
This area could utilize a 2 or more stage radius to reduce the stress concentration at the
intersection point of the two surfaces, while keeping the lower stressed areas at a small
radius to maintain weight. This would significantly improve the beam since stresses in these
locations raised both my weight and deflection to get the stresses within limits.
Other notable observations would be that the size of the center triangle played a major role
in the overall deflection the wider the triangle, the lower the deflection and stress at the
bottom of the beam.
As well, working with other students who achieved the same design but solved from
opposite directions (stress under control but weight was over) really emphasized what
changes made to the outer triangles affected which priority. It was a pleasure discussing
and theorizing the affects changes would make to get a certain over-limit back within
specifications without putting another limit over.

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Beam Load Indentation

Figure 17: Beam point load deformation

As can be seen in figure 17, the model had visible deformation and exceeded the allowable
stress limits at the point of contact. This is due to the infinitely narrow area of load the force
was applied on (a line). This type of deformation should result in a permanent indent when
physically tested, as the elastic limit of the material would be surpassed. Each of the
reactive constraints shows similar but smaller indents from the bottom.

Summary of Results
Beam

Weight
(difference)

Deflection
(difference)

Stress
(difference)

1.0

.3653lb
(24%) over
.3147lb
(7.1%) over
.2680lb
(-8.8%) under
.2922lb
(-0.16%) under

.00470in
(-53%) under
.00608in
(-39%) under
.01260in
(26%) over
.008918in
(-11%) under

12584psi
(26%) over
15940psi
(59%) over
25200psi
(152%) over
9955psi
(-0.45%) under

1.25
1.75
2.0
Table 4: Design results summary

16

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Discussion
The progression of the beam testing went from overweight, which was to be expected
given no initial effort was made to reach the required value, to under. The large arch
that compromised beams 1-1.75 did not show satisfactory results, and as such the
center arch was replaced with straight sides for beam 2.0.
The deflection progression shows that as the center triangle base narrows, the
deflection in the center increases.
A high beam weight limits the amount of stress when a poor design is utilized. The
progression of weight loss had an inverse relationship with stresses in the beams 1.0 to
1.75, with the given beam maintaining this trend by having the lowest stress.
Stress was the hardest value for me to reduce in this project. As can be seen in my
beam progression, only my final design managed to maintain the stress within limits. All
previous beams (except beam 1.75 disaster) met deflection values, mainly due to the
wide main triangle all had. Beam 2.0 sacrificed great deflection values to harness the
excessive stress values.
Beam 2.0 could be improved more by using alternate modelling techniques in the
transition zone from radii to straight faces, allowing a wider center triangle and
decreasing deflection for a better ranking value. As well, there were still remnants of
blue showing in the final FEA results, which could be minimized further for lower
weight/improved material use.

17

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Manufacturing
Introduction
To ensure dimensional accuracy between our manufactured beam and our designed
beam, the NX CAM package was used to create toolpaths for the Matsuura 5-axis CNC. By
using the 3d model in the creation of the code for the CNC machine, any interpretation of
dimensions and operator skill have been removed from affecting the final part for testing. As
this project was to compare FEA results with tested results, cut-out dimensional accuracy
was of extreme importance. Using NX CAM to machine the beam designed in CAD also
highlighted the overlap between our CAD and CAM courses.

Purpose
Each final beam was manufactured to give us the ability to physically test our beam
design. To manufacture the beams, we were given a modelled NX jig for our beam blank to
rest on (see fig 18-20). This was a CAD representation of the jig used in the Matsuura mill
our beams were to be machined on, allowing our toolpath avoidance moves to accurately
avoid the jig and holding fixtures when being machined. By using jigs and clamps during
toolpath creation, it reminds us to always consider how a blank will be constrained and
located in the mill for dimensional accuracy.

18

Figure 18: Top view of CAD jig fixture

Figure 19: Right Side view of CAD jig fixture

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Figure 20: Top view of jig fixture - locating

Procedure
In NX manufacturing mode, the designed beam is placed into the setup by replacing
the demo beam that occupies the jig, and properly located against each datum locator(s).
This ensures the MCS (Machining Co-ordinate System) is the same for all beams placed in
the jig. Once the beam is properly located, the new beam is selected as part within
workpiece selection for toolpath generation.

Toolpaths
The toolpath order to machine the beam goes in the following order:

Center drill
Access drilling for the 3/8 HSS end mill using a 3/8 HSS twist drill
Rough mill using a 3/8 end mill
2 finish passes using end mill

Starter holes locations are chosen within each pocket that will allow the 3/8 drill and end
mill to enter the pocket without notching the beam upon entry. For the majority of pockets,
this is roughly half of the height of each triangle. The locations can be seen in fig 21 (next
page) with the IPW (In-Process-Workpiece) showing the remaining material.
After the access holes have been drilled the 3/8 end mill is used to rough out the pockets.
The cutter pattern uses follow periphery starting from the drilled access holes. While
minimum radii were allowed to be smaller than the 3/8 end mill, the large cutter does not
fully enter the corners. Rather, the large cutter is used for faster material removal and
stronger shank. The amount of remaining material in the small radii can be seen in fig 22
(next page) in the secondary and tertiary triangles.

19

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Figure 21: Drilled holes IPW

20

Figure 22: 3/8 roughing IPW

Once the 3/8 roughing is complete, an initial finish pass is made with the specified end
mill. However, due to the amount of material remaining in the highlighted corners of fig 22,
and the small diameter of the 1/4 cutter shaft - which may allow deflection - the finish pass
is repeated to increase the dimensional accuracy of the machining process. Each of these
finish passes only does one follow profile cut due to the majority of the material already
removed by the 3/8 roughing procedure.

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM


Toolpath Visuals:

21

Figure 23: cutter toolpaths (consolidated holders for visual)

Verification
Once all the toolpaths have been created, the verify toolpath option is run, which can
be set to check for any tool, tool holder collisions with the part or check geometry, as well as
possible gouges from the tool engaging the work part on toolpath entry or exit. Inspection of
the final IPW can also expose any issues with gouging.
After all the toolpaths have been verified, a post-processor is used to convert the NX
CAM toolpaths, cutter information, speeds and feeds, into useable G-code for the machine
in use - for this beam, the post-processor used was matsuura_mill_3_in.
For the machined beam, visual inspection and measurement of known dimensions
such as border thickness and web thickness if the beam looked accurate. Measuring weight
and volume can also be done to compare the machined beam to the CAD model.

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Discussion
Operation
Drill - 3/8
Mill rough - 3/8 cutter
Mill finish - cutter path 1
Mill finish - cutter path 2
Overall

Time NX (min)
0:35
5:28
3:04
3:02
12:45

Time Actual(min)
~2
~8
~6
~4
~20

Table 5: Toolpath times

The significant difference between the projected completion time and the actual completion
time is split between two factors - the time for the machine to change tooling between
toolpaths, and the ability for the mill operator to slow the feed down when approaching the
work piece. As all our beams were unique, and the G-code was student generated, the
operator's cautiousness was easily justified. Once the program was underway and cutting in
the appropriate locations, the speed was set to 100% for the remainder of that operation. By
slowing down the tool on approach, it is possible to save cutter and/or machine damage if
the toolpath has collisions and the machine is stopped on time.
Cutter direction for all milling operations used in the manufacture of this beam was climb
milling. There are many advantages to using climb milling if the machine is capable, with the
major restriction being table backlash. Excessive backlash can allow a climb cut direction to
suddenly pull more part into the cutter than it is capable of removing - this results in broken
tools. However, when backlash is not an issue - which most CNC machines would not have the benefits include better surface finish, lower power requirements, less re-cutting of chips
(swarf that receives multiple cuts), and better tool life. However, deflection can be higher on
long or small tools, and is part of the reason for the second finish pass used with the 1/4in
cutter.
Tool
.375" drill
.375" end mill
.25" end mill
Table 6: CAM Tooling

Description

RPM

Cut Direction

Feedrate

twist drill, HSS


2 flute, HSS
2 flute, HSS

3200
3200
4800

climb
climb

10 IPM
20 IPM
20 IPM

22

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Testing and Validation


Introduction
Physical testing of a CAD component is extremely useful in allowing confirmation of
material properties, design properties, and FEA setup/simulation accuracy. Other factors
such as machining stress risers and material inclusions can show up during physical testing
that would not be factors in computer modelling and simulation.

Purpose
The purpose of testing our designed and manufactured beam is twofold. The first, to
confirm that all of our CAD modelling and FEA was accurately set-up to replicate real world
materials, loads and constraints. The second, to demonstrate an accurate manner of
physically testing our beam to compare to our FEA results, using both direct and indirect
measuring tools.

Procedure
Strain Gauge application procedure
After the beam has been manufactured, a strain gauge needs to be placed in a specific
location on the side face near the bottom as seen in figure 24 (the bottom face would also
work if the dial indicator did not need to measure deflection there). This same location was
inspected in the FEA simulation to compare the stress levels.
23

Figure 24: Strain gauge location

The strain gauge and supportive wire pad are attached to the beam using a cyanoacrylate
based adhesive. To ensure the strain gauge can accurately measure the beam strain, proper
prep work is needed. The aluminum has an acid (M-prep conditioner) applied and any
surface oxidation is removed using sandpaper and manual sanding. The remaining
conditioner is wiped off and fresh conditioner and a Q-tip are used to best clean the
remaining oxidation off the beam.
To cleanse the surface of the conditioner, fresh neutralizer is used in combination with Qtips until any remaining acid/oxidation is removed (Q-tip remains clean).
The strain gauge and wire pad are placed where required and tape is used to replicate the
location after adhesive is applied.
Once the location has been prepped and the gauge/wire pad are ready to be re-applied, a

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM


small amount of catalyst is applied to the gauge and wire pad, followed by the adhesive
application and quickly locating the gauge and wire pad onto the beam. Pressure is used to
maintain contact until the adhesive has bonded.

24
Figure 25: Strain Gauge and wire support pad

After both pads have been applied, wires are soldered to connect the strain gauge to the
wire support pad, and three connection wires soldered to the support pad that connect to
the strain indicator for testing. As a strain gauge is a resistance based device, the
orientation of the wires does not matter, only that the black/white wires are connected to
the same pad for the strain indicator to function properly. The resistance across the red and
black/white wires is used to confirm proper connections a resistance of approximately
120.5 is the desired result.

Figure 26: Strain gauge applied

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Testing Procedure
The beam is placed inside the testing setup with the dial indicator at the center bottom
face, unweighted load arm resting on the beam, and the strain gauge wiring connected to
the top post for each respective colour, as seen in fig 27-29:

25
Figure 27: beam testing setup dial indicator

Figure 28: Strain gauge wire hookup

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

26

Figure 29: Overall setup ready for testing

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM


Once both measuring devices were set to 0, the first weight with hook (see fig 30) is slipped
into the eyelet at the end of the load arm as gently as possible to minimize variation and
load spikes. Measurements from the strain gauge and dial indicator is recorded for future
use. Once the measurement is recorded, the 2nd weight is slipped onto the hooked 1st
weight and the process is repeated 2 more times, for a final effective value of approximately
500lbs. The 4th weight includes the small 1 lb to achieve the effective 500lbs.

27

Figure 30: Weights in order

Once the final weight has been added and the data recorded, each weight is removed and
the measurements at each unloading step compared to the loading step for accuracy.

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Data
Load
(effective)
0
160
270
380
500

Deflection (in x 10-3)


0
3.5
5.5
7.75
10.0

Strain
(in/in)
0
148
295
480
695

NX
Deflection
0
.0028
.00475
.00675
.009

NX
Stress
0
2280
3900
5500
7125

Figure 31: Data from load testing

Test Results
To calculate the stress from the strain units, multiply by the modulus of elasticity; for NX, this
value is 10.0047 x 106 psi (68.9MPa).
For the stress at the 500lb effective load:
= 695 10.0047
=
For the remaining load points, see table 7 below:
Load (lbs)

Strain (in)

0
160
270
380
500

0
148
295
480
695

Stress (psi)
Calculated
0
1481
2951
4802
6953

Stress (psi)
NX Values
0
2280
3900
5500
7125

Percent Difference
To NX values
0%
-35%
-24%
-13%
-2.4%

NX deflection (in)

Percent difference
to NX values
0%
+20%
+14%
+13%
+10%

Table 7: Calculated stress values

For the deflection differences to NX, refer to table 8:


Load (lbs)
0
160
270
380
500

Deflection
Measured (in)
.000
.0035
.0055
.00775
.010

0
.0028
.00475
.00675
.009

Table 8: Deflection comparison

For a visual representation of these measurements, see graph 1 and graph 2 on the
following pages.

28

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

STRESS VS LOAD
8000

Tested

NX stress
7125
6950

7000

6000
5500

4800

STRESS (PSI)

5000

3900

4000

2950
3000
2280
2000
1480

29
1000

0
0
0

160

270

EFFECTIVE LOAD (LBS)

Graph 1: Stress vs Load

380

500

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

DEFLECTION VS LOAD
Deflection

NX deflection

0.0095
0.009

0.00775

0.00675

0.0055
0.00475

0.0035

0.0028

0.0000
0

Graph 2: Deflection vs Load

Discussion
Physically testing the beam was an interesting if not slightly inconsistent process.
Experience using the machine/setup seemed to improve the consistency of the results. The
largest variance came in the middle of the curves, which may have a basis on how the strain
gauge is calibrated (load beam already weighting the beam) and/or how the actual gauge
correction factor was achieved for our beam testing. The deflection numbers were
consistent across the board, with the only varying factor being the placement within the test
rig affecting the leverage ratio of the load arm.
The dimensions for the load arm and weighting were:

25 from pivot to weight eyelet


2.65 from pivot to beam load application
~8kg for load 1 with hook
5kg for load 2 & 3
5kg + a 1lb weight for the final load

The leverage ratio works out to just under 9.5:1, requiring 53lb of load to replicate the NX
tested 500lb.

30

DESIGN AND ANALYSIS OF A SIMPLY SUPPORTED BEAM

Conclusions
This beam project was an excellent introduction into the combined world of
CAD/CAE/CAM and product testing. While the CAD modelling of the beam did not stress our
capabilities within NX, designing with the intention of future adjustments made a large
difference in how quickly a design tweak could be made, and then retested to see the
results. If there was one regret I had during this project, I was only that I did not have more
time to test my beam design further. I thoroughly enjoyed every aspect of this project, from
the design itself, the friendly competition with my peers, being able to CNC machine our final
design, as well as having more than one source of data for our physical testing. Having to
document each stage was useful in solidifying knowledge as it was acquired for future use
as well as being able to share the knowledge and experience to others.

31

You might also like