Lab Manual 2013

ME/CE 96: Mechanical Engineering
Laboratory
Spring 2013
CONTENTS
1 Course Information 1
1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Contact Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Location . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.4 Class Webpage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.5 Lectures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.6 Lab Notebooks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.7 Prelab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.8 The Lab Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.9 Grading Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.10 Time Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.11 Deadlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.12 Lab Partners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.13 Collaboration Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.14 Signing Up for Experiments . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.15 Operating Guides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.16 Starting an Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.17 Scheduling Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.18 What to Do if Something Breaks . . . . . . . . . . . . . . . . . . . . . . . 7
1.19 Lab Computers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.20 safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.21 Lab Safety Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Cantilever Beam 9
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Prelab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5 Optional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3 Turbomachinery 17
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
i
3.4 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.5 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
3.6 Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.7 Operating the System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.8 Precautions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
3.9 Prelab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4 Free and Forced Convection 29
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
4.2 Experimental Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.3 Data Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.4 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.5 Heat-Transfer Predictions . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.6 Lab Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.7 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.8 Prelab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5 Turbulent Air Jet 37
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.2 Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
5.4 Prelab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
5.5 Air Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
5.6 Appendix: Motion control . . . . . . . . . . . . . . . . . . . . . . . . . . 47
6 Mechanical Properties of Metallic Materials - The Tensile Test 51
6.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.2 Test Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.3 Test Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
6.4 Tests to Perform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
6.5 Fracture Surface Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.6 Lab Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
6.7 Pre-lab Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7 Dynamics of Coupled Mechanical Oscillators 57
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
7.2 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
7.3 Analysis of Experimental Model . . . . . . . . . . . . . . . . . . . . . . . 60
7.4 Lab Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
7.5 Advanced Experiment (Optional) . . . . . . . . . . . . . . . . . . . . . . 61
7.6 Prelab . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
7.7 Appendix: Making Accurate Period and Phase Measurements . . . . . . . 61
A Keeping a Lab Notebook 63
B LabVIEW 65
ii
CHAPTER
ONE
Course Information
1.1 Overview
ME/CE 96 is a laboratory course with experiments in mechanics, dynamics, control, uids,
heat transfer, turbomachinery, combustion, and fuel cells. During the course of the term,
you will do 4 experiments, selected from 9 that are available this year. Each experiment
lasts two weeks.
A new aspect of the course is an emphasis on data acquisition methods. As part of the
course, you will learn how to build Virtual Instruments using LabVIEW to acquire and
analyze experimental data and to control equipment.
1.2 Contact Information
Instructor: Monica Kohler kohler@caltech.edu x4142 229 Thomas.
TAs: Juan Cardenas cardenas@caltech.edu
Stephen Schwee sschwee@caltech.edu
1.3 Location
The experiments are located in the sub-basement of Thomas Laboratory, in rooms 0018 and
0021. Please see Chris Silva in Thomas 208 to get a key to the lab rooms. Keys must be
returned at the end of the term to Chris Silva to recieve your grade.
1.4 Class Webpage
There is a webpage for the course, which may be found at http://sites.google.
com/a/caltech.edu/me96spring/. This site contains the most up-to-date infor-
mation about the course, handouts, class noted, useful programs, and an experiment sched-
ule.
1
1.5 Lectures
A few lectures and demonstrations will be given early in the term. Each one will be 60 to
90 minutes, and is designed to introduce you to useful methods and tools. For several of the
lectures, there will be a homework assignment. The assignments will not be very difcult
or time-consuming, but are designed to reinforce concepts from the lecture. They will be
due one week after the lecture. All of the homework assignments together will count for
10% of your course grade.
Some of the planned lectures are:
1. Data acquisition with LabVIEW
2. Error analysis
The times and dates for these lectures/demos are to be determined. Consult the webpage
for more information.
1.6 Lab Notebooks
You will need two bound laboratory notebooks. The bookstore if still open, sells several
suitable styles. All of your written work will be done in these notebooks and you will
alternate from one to the other between labs, so that while one can be graded while the
other is used for doing experiments. See Appendix A for more information on how to keep
a lab notebook.
1.7 Prelab
Each lab has a set of questions, estimates to make and/or problems that must be solved
before beginning the experiment. These are designed to insure that you come to lab well
prepared. Write the answers to the prelab questions directly into your notebook. There is
no need to copy the problem statements. When you meet the TA to go over the experiment,
he or she will check that you and your lab partner have done the prelab appropriately.
You will not be allowed to begin the experiment until you have completed the prelab
satisfactorily.
1.8 The Lab Report
The lab report is the primary means for you to communicate your results, and to demon-
strate to the TA and the instructor your level of understanding and the quality of your pro-
cedures and results. Consequently, it is expected that the reports will be written thoroughly,
carefully, and thoughtfully.
The report should be written in your lab notebook. In most cases, your report will consist
of entries written directly into the notebook with pen, along with computer-generated text
2 Chapter 1. Course Information
or plots. Quick plots made in the lab for a particular purpose (e.g. to check the linearity
of a sensor) would typically be drawn by hand, but plots made after you have acquired the
data and have analyzed it (e.g. comparing measurements to theory) should be done using
the computer for faithful reproduction of quantitative results.
We expect the report to be reasonably clear and legible, in addition to being well-organized.
But we should emphasize that a lab report is different from the sort of reports you may have
written for other classes. We dont want a nished report like a term paper or a scientic
publication, but instead want to see the actual record of what you did in the lab, written in
your notebook.
A good lab report will contain crossed-out (but still readable) text where you made errors,
will usually contain hand-drawn sketches of equipment, may contain hand-drawn quick
plots done in the lab, and may have lots of things pasted into it, such as computer-generated
plots or photographs. Much of the report will be written by hand in pen (never pencil). It is
perfectly ne to write the entire report by hand, with glued-in computer plots or printouts
as necessary. But if you are writing a lengthy discussion of the results, and would prefer
to use a computer, that is ne. Simply print out the text and glue it page by page into the
notebook.
There are special rules that govern making entries in lab notebooks, including lab reports.
See Appendix A for more information.
The lab report should contain a complete
1
record of the data you took in the lab, along with
your analysis and conclusions.
The report should begin with a brief statement of the objective of the experiment. Dont
simply copy the objectives statement in the handout state the objectives in your own
words.
You should include a description of the apparatus, with a sketch with all signicant parts
labeled. Dont simply cut and paste the sketch (if there is one) from the lab handout. It is
recommended that you draw the sketch by hand in your notebook the act of drawing and
labeling each component makes you take notice of every part of the experiment, and gives
you a better physical feel for the experiment than can be obtained by simply passively
looking at sketch drawn by someone else.
The report should include a description of all procedures followed, a record of all measure-
ments made and a discussion and analysis section at the end, where the results are compared
to theory or other measurements, as appropriate, and conclusions are drawn. Your report
should also contain an error analysis, in which the uncertainties in the measured values and
in any functions derived from them are estimated. Any comparison of measured data to
theory should always include error bars on the data points.
The analysis and discussion portions should be thorough with attention to detail. If you
are asked to compare your measurements to theoretical or numerical predictions, and you
nd discrepancies larger than experimental uncertainties, discuss why this is so. If you
can, attempt to resolve the problem dont be satised with unexplained discrepancies.
If you conclude that an instrument was malfunctioning, or the measurements were done
incorrectly, you may want to go back into the lab to make more measurements or check the
rst ones (so dont leave the report writing until the last minute!). Or it may be that the
1
but see Appendix A for more on what constitutes complete.
1.8. The Lab Report 3
theoretical results are missing some important effects; can you account for them with better
theory? It is easy to claim the problem is with the theory and not with the measurements.
How do you know? What evidence supports your claim? Whatever you conclude, always
back up your statements with evidence.
As noted above, remember that a lab notebook is a lasting record of your experiment in
a research laboratory, they are often retained for years or decades, so that later researchers
can refer to the original data. (Millikans famed oil-drop experiments were recorded in
notebooks that are in the Caltech Archives.) It is just good policy to get into the habit of
making top quality lab notebooks right from the start. In addition, the best of your lab
reports will be used to train next years TAs.
1.9 Grading Policy
The course grade will be based on the lab reports for the four experiments, plus a few
homework assignments from the lectures early in the term. Grading will be done primarily
by the TAs and reviewed by the instructor. The grade for each experiment will be based
on a 100 point scale with the prelab counting for 20 points, the quality of the lab report
counting for 40 points and lab performance counting for 40 points. The lab performance
will be based on a number of factors that may include meeting the deadlines described
below, resourcefulness in nding needed information, preparation before coming to the lab,
safety practices, experimental technique and the quality of the resulting data. This portion
of the grade attempts to assess the quality of what you do in the lab and how you do it. In
most cases, it will be determined from your written lab report, although it may be based in
part on observations by the TA.
The lab report points will be based on the quality of the presentation, analysis, and discus-
sion in your report. The emphasis is not on supercial appearance, except that the report is
expected to be legible and well-organized.
1.10 Time Management
You should expect to put in approximately 9 hours each week on this course, ideally in
a few sessions of 2 or 3 hours each. Learning anything is most effective when you work
on a project or assignment for short periods every day (3 hours maximum), instead of in
one long session just before the assigment is due. When you get tired, or nd you are not
concentrating well, it is time to rest or do something else for a while, and come back to
the project when you are refreshed. This is particularly important with lab work, since
accidents are much more likely if you are tired or not mentally alert and fresh.
Note that 9 hours per week is not the same as 18 hours every other week, or 90 hours in the
last week of the term! The experiments are designed to be done over a two-week period,
and simply cannot be done in a marathon all-night session the night before the lab report
is due. It is fair neither to your lab partner nor to yourself to try to do the labs at the last
minute.
And remember, when doing experimental work, you have to expect the unexpected. Very
often things happen that require time to deal with. Maybe you cant get the oscilloscope to
work is it broken, or do you simply not understand how to use it, or is there some button
somewhere to reset it, or . . . ? Or maybe you have a bug in your LabVIEW VI that you need
to track down before you can make the measurements. Be sure to start the experiment early
enough that if something unexpected occurs, you have time to resolve it.
To help insure that everyone keeps up with the schedule, several intermediate deadlines
have been established that must be met, as described in the next section.
1.11 Deadlines
The 8-week period beginning with the second week of the term is divided into four two-
week blocks: Block A, Block B, Block C, and Block D. You will do one experiment in each
block, and the lab report for each experiment will be due on the rst day of the next Block.
In addition, there are some important intermediate deadlines.
All of the deadlines are listed on the course calendar, which is accessible from the ME96
class page. You can subscribe to this calendar from your own calendar program (Outlook,
or Google Calendar, for example) so that you can set up reminders for yourself.
The deadlines are as listed below. Here week n means the n
th
week as measured from
the start of a Block. So, for example, week 3 of Block B is the same as week 1 of Block C,
etc.
1. Thursday of week 1. By this day, you must have completed the prelab and have met
with the TA to go over the experiment. You cannot begin the experiment until you
have met with the TA and he or she gives you the go-ahead to start the experiment,
and he or she will not do so if your prelab is not commpleted.
2. Thursday of week 2. You must have completed the bulk of the measurements by
Thursday of week 2 at the latest, in order to allow sufcient time to analyze your
data and nish writing the report. You can still go back into the lab and take more
or different measurements after this date, since you may nd as you write the report
that additional measurements are needed. When grading your reports, the TAs will
note the dates you enter in your lab notebook to determine whether or not you met
this deadline.
3. Monday of week 3. On the Monday following the end of the two-week block for the
experiment, the nal lab report is due. Please turn your notebook in at my ofce (115
Thomas) before 5 pm sharp. After this date, late lab reports will be accepted for one
week only, with a 10 point reduction in the maximum possible points for each day
late. For example, if your report is 3 days late, and you would have received a score
of 80 if it had been turned in on time, your grade would be 80 0.7 = 56. No lab
reports will be accepted more than one week late.
4. Monday of week 4. Your graded lab report will be available in front of my ofce one
week after the deadline for turning in the lab reports.
5. Monday of week 5. If your lab report was submitted on time, you can correct and
resubmit anything you missed in the lab report, for 80% of the original point value.
1.11. Deadlines 5
Any corrections must be submittted by 5 pm on the Monday of week 5, which is also
the due date for the next lab report. Corrections will not be accepted after this date.
Corrections cannot be submitted for lab reports that were turned in late.
Note that no lab reports will be accepted more than one week late. Since your grade is
determined mostly by the four lab reports, it would be difcult at best to pass the course
with an acceptable grade missing even one lab report. It is impossible to pass if you miss
two. If you dont feel you can keep to this schedule, either due to Ditch Day, extracurricular
activities, a heavy course load, or any other reason, this is not the course for you. You
should drop it now.
The only exceptions to this policy will be for genuine, unexpected medical or other serious
events beyond your control. In such cases, please submit a note from the Dean or from a
doctor explaining the situation.
Note that preparations for Ditch Day do not constitute a valid reason for missing these
deadlines. If Ditch Day falls on a day with a deadline (a Monday or a Thursday), then the
deadline will be extended by one day.
1.12 Lab Partners
You will all work in teams of two. In most cases, you should arrange to have the same
partner over the whole term. You and your partner should both be present when taking data.
All lab reports must be written individually.
1.13 Collaboration Policy
You may freely discuss the experiment and your interpretation of the results with your
partner or other students. On the other hand, you are expected to do your own numerical
solutions, derivations, error analysis, etc.
1.14 Signing Up for Experiments
Contact the TA for the experiment you and you lab partner would like to perform. Each lab
group (of 2) should sign up for four experiments -one for each two week period. At most,
experiments can accommodate three groups at once (working in the lab at different times
during the week). To see what experiments in what experiment blocks are available, go to
the experiment schedule is listed on-line.
1.15 Operating Guides
In addition to this manual, each experiment there is an Operating Guide that describes in
detail the practical aspects of how to run the experiment. These are PowerPoint documents
with photos of the equipment, and show, for example, where the power switch is, what
knobs do what function, etc. You should take a look at the operating guide before meeting
with the TA. They are all available on the ME96 class page.
1.16 Starting an Experiment
At the beginning of each two-week period you should make an appointment with the TA
to meet in the lab at a mutually convenient time to have the TA go over the experiment.
You must do the prelab before meeting with the TA, and cannot begin the experiment until
starting the experiment. Grades will be reduced for those who are unprepared or for those
who miss their appointments with the TAs (remember, the TAs are busy students just like
you).
1.17 Scheduling Time
Each experiment will have a sign-up sheet where your group can reserve the experiment.
Each group can reserve one block of time (up to four hours in duration) in advance. Once
you have nished the lab time, you can sign up for an additional block of time. It is best
to space lab sessions by a day or two so that you can look over your results and check that
things are looking reasonable before going back into the lab.
1.18 What to Do if Something Breaks
It is an unfortunate reality of working with equipment of any kind that sometimes things
break, or simply fail to function as you expect or as the documentation describes. If some-
thing is broken or not working, rst see if you can diagnose what the problem is. Is the
problem in software? Is a connector loose? Is the problem constant or intermittent? Does
the manufacturers web site have any relevant information? An important part of becoming
a resourceful engineer is learning how to troubleshoot and x things.
But, having said that, dont spend hours tracking down the problem. If you cant gure
out the problem after 30 minutes of work, contact the TA for the experiment. He or she
may know immediately what the problem is. Also, if any component is actually broken
and needs to be replaced, contact the TA or an instructor as soon as possible. We may have
replacement parts stored elsewhere, or may need to order a replacement part.
1.19 Lab Computers
Each experiment has a PC that is used to acquire data, or to control the experiment. All but
one of the PCs run Windows XP. They should all have Microsoft Ofce, MATLAB R2006b
and LabVIEW 8.2, and are all on the Caltech network. In addition, other software tools that
allow you to transfer your data les to other machines are installed.
The PCs are all in Workgroup ME96. Each one has a folder entitled Shared that can be
accessed for reading and writing from any of the other PCs.
1.16. Starting an Experiment 7
Please keep your data les together in a folder . There is no way to protect your les from
accidental deletion or corruption by others, so you should transfer or backup all data les
at the end of each lab session. A memory stick is convenient for this purpose.
Note that the folders belonging to other students should be regarded as their private property,
and should not be perused without their permission. When you are completely nished with
an experiment, please delete your folder containing your data (but retain a copy elsewhere,
or burn a CD).
All of the PCs can access the B/W laser printer in 0018 attached to the airjet PC, the color
inkjet printer attached to the stress/strain PC, and the network printer in 0021.
Please do not install games or pirated software of any kind on these PCs. If you need any
software package installed that Caltech has a site license for, let the TA know and we will
see if we can install it.
1.20 safety
Please use caution when dealing with very hot or very cold entities. For example the ame
burner in the fuel cell experiment or the liquid nitrogen in the boiling heat transfer exper-
iment. Safety goggles and heavy gloves are available in the lab. Please use them when
necessary esp with these experiments. There is also a rst aid kit for minor injuries and a
re extinguisher. Hearing protection is available for the turbomachinery experiment.
1.21 Lab Safety Rules
To maintain a safe lab environment, there are a few rules that must be adhered to:
1. No working alone at any time. You can work in the lab whenever it suits your sched-
ule best, but you must always have at least one other person present in the room with
you if you are doing anything involving lab equipment. This rule does not apply if
you are only in the lab to use a computer. This should not present any problem, since
normally you will work together with your lab partner.
2. No food or drink is allowed in the lab at any time.
3. Shoes must be worn in the lab at all times.
4. Do not wear dangling jewlery or clothing that could be caught when working on
experiments involving rapidly moving mechanical components (inverted pendulum,
vibrations, turbomachines).
Also, please use common sense. It is not a good idea, for example, to work in any lab (or
shop) when you are sleep-deprived, or are for any other reason less than 100 %.
CHAPTER
TWO
Cantilever Beam
2.1 Introduction
The experiment involves the bending and vibration of an aluminum beam. Measurements
are made of the deection, strain rates, fundamental frequency, and damping constant. The
student is exposed to measurement techniques, data acquisition, and analysis. The experi-
mental results are also compared with theory.
2.1.1 Analysis of beam strain and deection
This section briey reviews the material necessary to make calculations of beam deection
and strain for an imposed load. The notation corresponds with the Figure 2.1.1 (a). Recall
for a linearly elastic beam that the strain in the x-direction,
x
, is linearly related to the
imposed normal stress in the x-direction,
x
, by the following relation:
x
=

x
E
, (2.1)
where E is the modulus of elasticity.
In pure bending, the strain on a beam can be expressed in terms of the radius of curvature
of the beam R and the distance from the neutral axis of the beam y:
x
=
y
R
(2.2)
Hence, the strain is zero along the neutral axis, and its magnitude increases with distance
from the neutral axis. The strain is negative in the +y direction, which corresponds to a
negative stress or compression in the +y part of the beam. Equations (1) and (2) can be
rearranged to give the stress at some position within the beam:
x
=
Ey
R
. (2.3)
The moment M(x) acting on the beam can be calculated from the stress:
M(x) =
_
A
y
x
dA. (2.4)
9
Figure 2.1: Geometry for analyzing beam deection.
Hence, a positive moment produces compression in the +y bers of the beam. Recalling
the denition for the moment of inertia I:
I =
_
A
y
2
dA, (2.5)
the moment-curvature relation can be found for a homogeneous beam:
M =
EI
R
(2.6)
and the exure formula follows as:
x
=
My
I
. (2.7)
The radius of curvature of the beam can be related to the displacement v of the neutral axis
of the beam due to bending. For small deections of the beam compared to the length of
the beam, the radius of curvature can be determined from the following equation:
1
R
=
d
2
v/dx
2
[1 + (dv/dx)
2
]
3/2
. (2.8)
If the slope dv/dx of the curve describing the loaded beam is at all points small relative to
unity then one can set the denominator of the expression for curvature equal to one. If this
is the case, then this equation can be approximated as:
1
R
= d
2
v/dx
2
. (2.9)
10 Chapter 2. Cantilever Beam
2.1.2 Analysis of beam vibration
Consider a beam of uniform cross section A and density , as shown in Figure 1(b). At
some position x within the beam, a balance of forces and moments can be drawn using
f(x) as some distributed load on the beam. The resulting equations for the balance of the
force and moments are as follows (assuming that dx approaches zero):
f(x) =
dV
dx
(2.10)
and
V =
dM
dx
. (2.11)
These equations can be combined with the moment-deection equation:
M = EI
2
v
x
2
. (2.12)
These equations can be combined to give:
EI
4
v
x
4
= f(x) (2.13)
The distributed load is the inertial load due to vibration and can be represented as the prod-
uct of the mass per unit length and the cross sectional area in the direction opposite of the
acceleration:
f(x) = A
2
v
t
2
. (2.14)
Combining equations (2.13) and (2.14) results in the following fourth order partial differ-
ential equation:
EI
4
v
x
4
+ A
2
v
t
2
= 0. (2.15)
This differential equation can be solved subject to specied boundary and initial conditions.
The above analysis assumes that there is no damping of the vibrating beam, and hence the
vibration of the beam would be of constant amplitude. However, in real systems damping
is usually always present due to friction. In a vibrating beam the friction is internal to the
medium, and is associated with the energy being dissipated randomly in the crystal lattice.
The energy dissipation manifests itself as an internal heating of the beam. The frictional
damping reduces the amplitude of vibration over time.
For a vibrating system with a single degree of freedom (harmonic oscillator), the motion of
the system is governed by the following equation:
y = a
1
cos pt + a
2
sin pt, (2.16)
where p is the frequency of vibration and is related to the period of vibration by = 2/p,
and a
1
and a
2
are unknown constants. If the system is damped and the damping force
is proportional to velocity (as is the case for internal friction), the motion of the damping
system is governed by the following equation:
y = e
nt
(a
1
sin qt + a
2
cos qt) , (2.17)
2.1. Introduction 11
Figure 2.2: The cantilever beam experiment.
where n is the damping constant and vibratory motion now has the period:
=
2
q
=
2
_
(p
2
n
2
)
(2.18)
Note that if n is small relative to p, the period of vibration is close to the value obtained
without damping.
2.2 Experiment
2.2.1 Data Acquisition
In this experiment, you will measure the strain at ve locations on a beam subjected to
various static and dynamic loads. The measurements will be done using a strain indica-
tor, which uses an electronic bridge circuit to accurately measure very small changes in
resistance.
For static measurements, the strain may be read directly on the front panel. The strain is
displayed in units of microstrain. For example, a reading of 5000 would correspond to a
strain of 0.005, or 0.5%.
For dynamic measurements, the strain indicator unit provides an analog output on the back.
Connect this to channel 1 of the digital oscilloscope. The oscilloscope can be used directly
to view the dynamic strain waveform, although it does not have the capability to store data.
Therefore, a better approach is to control the oscilloscope from the computer.
A LabVIEW VI is provided that controls the digital scope, and displays the waveform
vs. time. You will need to modify it to also display the frequency spectrum of the signal.
2.2.2 Measurements
1. Make a sketch in your notebook of the experimental setup, and give a short descrip-
tion of the major components. Include any relevant dimensions you will need for the
analysis.
2. Apply different loads on the end of the beam and record the resulting strain and tip
deection. Do the measurements for both axes of the beam and for all ve strain
gauges for a single repetition (2 axes X 5 gauges X # of loads measurements). Plot
your results. Note that the strain indicator must be balanced separately for each strain
gauge. Therefore, it is simplest to take measurements for the full range of loads with
one strain gauge, then select another strain gauge with the switch, re-balance the
strain indicator, and repeat the set of loads.
3. Investigate the repeatability of these measurements. For each axis, choose one strain
gauge (which one will give the biggest output?) apply a load multiple times to the
beam, each time beginning from an unloaded state. Record the strain (for the one
gauge) and deection readings each time. Do at least six repetitions. Repeat for at
least two loads. Repeat for the other beam axis (2 axes X 1 gauge X repetitions X #
of loads).
4. Familiarize yourself with the oscilloscope and the LabVIEW VI that runs the oscil-
loscope from the computer. Record the bandwidth, sampling rate, and resolution of
the oscilloscope.
5. Using the oscilloscope (either directly, or from LabVIEW), select one strain gauge
and determine the fundamental vibrational frequency of the beam on each axis, by
striking the beam with the supplied hammer and examining the resulting dynamic
strain signal. Adjust the scope timebase and voltage scale so that you see several
(ten or more) periods on the screen. You may need to adjust the trigger to get a
stable display. The easiest way to measure the fundamental frequency is to count the
number of zero crossings in the voltage vs. time signal.
6. Now using LabVIEW to run the oscilloscope with the supplied VI, edit the VI to
add a graph of the power spectrum (you can nd a convenient VI for this on the
Express function menu). Now hit the beam with the hammer and examine the
frequency content of the signal. Some things to explore:
(a) What are the frequencies of the harmonics of the beam vibration?
(b) Do the amplitudes of the harmonics depend on where you strike the beam?
Make several measurements of the harmonics, so that you can estimate the repeata-
bility. Do the same for the other axis.
2.2. Experiment 13
7. For one of the harmonics, for example the rst overtone, examine the effect of sam-
pling rate. Include at least one measurement with a sampling rate less than twice the
frequency of oscillation. Make a plot of the measured frequency vs. sampling rate.
2.3 Data Analysis
1. Compare your measured strain vs. load data to theoretical predictions for a linearly
elastic beam in pure bending. Show in a plot how the two compare, and include error
bars on the measured values. Discuss any discrepancies found, including possible
causes.(Were there unaccounted-for systematic errors in the measurements? Does
the theory make simplifying assumptions that arent valid? How do you know? What
additional measurements could you do to determine the problem?
2. Plot the tip deection vs. load, and compare to the solution determined by integrating
the differential equation. As in (1), discuss any discrepancies, their possible causes,
and the steps that could be taken to resolve the discrepancies.
3. If the mass of the beam were taken into account, how would it affect the deection
vs. load data? (The density of aluminum is 2.77 g/cm3.)
4. Compare your measured fundamental frequencies, and the ratio of the harmonic fre-
quencies to the fundamental frequency, to theoretical predictions. As in (1), discuss
discrepancies, etc.
5. Determine the damping constant for the fundamental frequency. This is most easily
done if you examine the signal with the scope timebase set to a long tim, so that you
can see on the screen the decaying-exponential envelope of the signal.
2.4 Prelab
Read the tutorial on measuring strain on the wiki.
1. What is the Gauge Factor (GF) of a strain gauge?
2. Suppose a strain gauge with GF = 2.5 experiences a strain of 0.1%. What is the
percent change in resistance?
3. Describe how bridge circuits can be used to accurately measure the very small
changes in resistance that result when a strain gauge is subjected to strain.
4. Determine the fundamental frequencies,
n
of vibration in terms of E, I, , and A
by solving equation 2.15. Assume that the beam is undamped.
Use separation of variables to obtain the solution v(x, t) = X(x)T(t).
Note that the solution of

T
2
T = 0 is T(t) = a
1
cos(t)+a
2
sin(t) where
the as are constants determined from the initial conditions.
Note that the solution of X
IV
+
4
X = 0 is X(x) = b
1
sin x + b
2
cos x +
b
3
sinh x + b
4
cosh x where the bs are unknown constants determined from
the boundary conditions. This solution can be rewritten as
c
1
(cos x+cosh x)+c
2
(cos xcosh x)+c
3
(sin x+sinh x)+c
4
(sin xsinh x),
(2.19)
The zeros of the transcendental equation cos Lcosh L = 1 are L =
(1.875, 4.694, 7.855, 10.996, . . . ).
relates the frequencies to the beam properties and the initial conditions.
5. Modify the spectralCant.vi to perform a frequency analysis on a signal generated
using the function generator.
Connect the function generator to channel 1 of the DAQ.
Generate a sinusoidal signal at roughly 2000 Hz.
Modify the VI to perform a spectral analysis on the signal and plot the output.
Include a printout of the modied front-panel and block-diagram (screenshot).
Verify that the spectral analysis gives reasonable results (since we know the
input frequency and amplitude).
Experiment with the sampling frequency and number of samples to see effect
these have on the measured frequency. Include at least one sampling rate that is
below 4000 Hz. Plot this data.
2.5 Optional
1. Time permitting, explore other effects of your own choosing. Some interesting ques-
tions: What is the effect of combined static and dynamic loading? Does it change the
vibration frequency? What would the dynamic response look like if you were to hang
a mass from a spring or a pendulum attached to the end? How would you analyze
such a problem theoretically?
2. Measure the damping constants of the harmonic frequencies. How do they compare
to that of the fundamental?
3. To appreciate how much care must be taken with strain gauges (or for that matter
with many types of delicate instruments), take one of the uninstrumented beams and
apply strain gauges to it. Each student should apply at least one strain gauge. Repeat
selected measurements of the static and dynamic response and compare the results.
Do your results agree to within previously-determined uncertainties? If not, can you
diagnose what went wrong with your strain gauge application? What evidence can
you give that the discrepancy is due to one cause, and not some other?
2.5. Optional 15
16
CHAPTER
THREE
Turbomachinery
3.1 Introduction
By turbomachine we mean a fan, pump, compressor, or turbine that changes the energy con-
tent of a owing uid by means of momentum exchange. Sometimes the term rotordynamic
machine is used for this class of machines to distinguish them from positive displacement
devices. Turbomachines are second only to electric motors in their number and are wide-
spread in practically all industries, ranging in power levels from a few watts to more than
100 MW. Because of the requirements of aircraft jet propulsion, new power plants (both
steam and gas turbine) and rocket propulsion research, development continues actively in
this eld today.
The present experimental turbomachine is an axial-ow fan powered by a three phase in-
duction motor. The fan is used on the MD-11 aircraft to cool electronic equipment in the
cockpit. The fan consists of a rotor which receives axial non-swirling air from the inlet
tube; the rotor imparts an angular velocity V
to this oncoming ow, which is subsequently

straightened out to be purely axial in a stationary row of vanes termed a stator. These stator
vanes also serve as heat-transfer surfaces needed to cool the drive motor that is mounted
internally inside the hub of the stator. The rotor consists of a cylindrical hub structure on
which equally spaced airfoil-shaped blades are mounted. The centroids of each radial blade
section are aligned on a radial line (this is the stack line) and are set at various angles to the
tangential direction to provide the requisite pressure rise (or enthalpy rise) of each radial
section of the rotor.
The present facility includes a motor controller that provides a variable voltage and fre-
quency (up to 400 Hz) for the compressor motor, permitting operation up to a synchronous
speed of 12,000 rpm. The motor is a four pole, 3 phase, Y wound induction motor. The
synchronous frequency in revolutions per minute (rpm) of the rotating eld is given by
rpm =
_
f(cycles/s)
(P/2)
_
60s
min
(3.1)
where f is the frequency in Hertz of the supply (read from the LED readout on the wall-
mounted motor controller) and P is the number of poles of the motor. Directions for oper-
ating the supply are given in the Operation section 3.7. An induction motor must operate at
a speed slightly less than the synchronous speed; the difference or slip is about 3% in the
present case. For the purpose of the experiment we will assume the fan operates at 97%
of the synchronous speed since we have no independent measurement of the actual rotor
17
Figure 3.1: The MD-11 aircraft. The turbomachine in this experiment is an axial ow fan
used to cool electronic equipment in the cockpit. Of course, the engines are also turboma-
chines only much larger and more powerful!
speed.
A schematic of the apparatus is shown in the Appendix. It consists of a semi-elliptical inlet
nozzle, upstream and downstream static pressure ports, the fan, and a manually adjustable
throttle plate. The inlet nozzle provides a smooth inlet ow, the pressure ports enable
determination of inlet ow and fan compression, and the throttle plate provides a means
to adjust the load on tthe fan. This test apparatus is built to the AMCA 210-85 American
National Standard Laboratory Methods of Testing Fans for Rating.?
3.2 Background
The energy level change brought about by a turbomachine is expressed as the total enthalpy
rise (for a fan, compressor or pump) per unit mass of the uid. In most turbomachines, this
work is accomplished by momentum exchange with the rotor and the total enthalpy rise is
given by a famous formula, the Euler Turbine Equation
h
t
= (UV
) (3.2)
where h
t
is the enthalpy increase in J/kg, U is the rotative speed of the impeller tip, i.e. r
at a particular radius r ( is the angular velocity), and V
is the component of the absolute

velocity in the direction of U. For an isentropic, incompressible ow,
h
t
=
p
t
(3.3)
where p
t
is the total pressure increase.
Closely associated with the Euler equation is the notion of geometric ow similarity; the
velocity vector triangle formed by the tip speed U, the absolute velocity V , and relative
velocity W (see ?, g. 12.18 for example) remains similar when ow rate and speed are
18 Chapter 3. Turbomachinery
Figure 3.2: The turbomachinery experiment in 0021 Thomas.
changed. Then it follows that
V
U
Q UA
m
where A
m
is the meridional cross section of the discharge ow. What follows, then are
p
t
D
2
2
Q D
3
Power D
5
3
where D is a reference length dimension such as the diameter of the pipe containing the
compressor ow. By means of these relations, the ow rate, pressure rise, and power can be
scaled to a standard reference speed and density. The scaling of these relations implies that
the effect of Reynolds number is weak; this is approximately so if the Reynolds number
(Re = WL/) based on blade chord length L, relative ow velocity W, and kinematic
viscosity of air is about 10
5
.
3.3 Experiment
There are several aims of this experiment: (1) to gain some knowledge of the typical behav-
ior of a heavily loaded axial fan as a function of ow and operating speed and to relate this
behavior to basic uid mechanics; (2) to learn how to measure fan performance; (3) to learn
basic similarity laws of fan performance; and (4) to make some qualitative dynamic mea-
surements of ow phenomena in stall a subject of much current turbomachine research.
3.3. Experiment 19
3.4 Instrumentation
This experiment is equiped with a water-equivalent manometer, two differential pressure
transducers, and a linear gauge. The manometer is used to calibrate the pressure transducers
that are in turn used to measure the pressure drop across the inlet nozzle and the pressure
rise across the fan. The pressure transducers are also capable of resolving time-varying
changes in the static pressures. The linear gauge is used to measure the throttle opening.
3.4.1 Experimental Variables
The following are the variables which the operator of this system is free to adjust:
Throttle setting The throttle plate is adjusted by rotating the plate on threaded lead
screw. Throttle displacements can be measured using the provided linear gauge, to
adjust the operating point of the fan. Always start the fan with the throttle plate at
least two inches open. The fan should not be operated at shut-off conditions (i.e.
throttle closed) for more than ve seconds maximum since ow through the fan is
used to cool it; operation at shut-off can lead to thermal overload and damage to the
fan.
Fan speed The maximum fan speed is about 12,000 rpm (400 Hz drive frequency).
For the purpose of this experiment the maximum speed has been limited to 10,500
rpm (350 Hz drive frequency). Operation at frequencies above about 200 Hz is very
noisy and the use of provided ear protection is mandatory. The minimum frequency is
limited by the ability to measure the pressure differences with the provided pressure
transducers.
to adjust the operating point of the fan. Always start the fan with the throttle plate at
least two inches open. The fan should not be operated at shut-off conditions for more
than ve seconds maximum.
Fan speed The maximum fan speed is about 12,000 rpm; this results in a very
noisy operation. For the purpose of this experiment the maximum frequency should
not exceed 250 Hz. The minimum frequency is limited by the ability to measure
the pressure differences with the provided pressure transducers. The directions for
starting the fan and adjusting the speed are given in the Appendix under Operating
the System.
3.4.2 Inlet Flow
The ow rate entering the compressor is determined by measuring the pressure drop across
the inlet nozzle. From Bernoullis equation:
(V
2
2
V
2
1
)
2
= p
1
p
2
, (3.4)
where V represents the ow velocity, p represents the static pressure, and represents
the density. The subscripts 1 and 2 represent the conditions before and after the nozzle,
respectively. Since condition 1 is the ambient air of the laboratory, V
1
is zero; therefore,
the velocity downstream of the nozzle (upstream of the fan) is determined by the pressure
difference across the nozzle:
V
2
=
2(p
1
p
2
)
. (3.5)
So, the theoretical volumetric ow rate Q
ideal
, assuming uniform ow across the pipe, is:
Q
ideal
= AV
2
= A
2(p
1
p
2
)
, (3.6)
where A is the cross-sectional area of the pipe. However, the real rate is actually less than
this theoretical value due to boundary-layer effects in the nozzle and other uid losses. To
accommodate the difference between the actual and the ideal ow rates, it is customary to
dene a discharge coefcient
c
d
=
Q
actual
Q
ideal
. (3.7)
The value of the discharge coefcient is determined by making measurements of the actual
velocity prole across the pipe just downstream of the nozzle. This was done before for
a very similar facility with the same inlet nozzle and 0.98 was the value found for the
discharge coefcient. Therefore, the actual ow rate for this facility is approximately:
Q
actual
= c
d
A
2(p
1
p
2
)
, (3.8)
So, at each throttle position, the ow rate through the compressor can be determined from
the measurement of the pressure drop across the inlet nozzle.
3.4.3 Fan Performance Curve
Incompressible turbomachines, fans and pumps, are tested at constant speed and uid den-
sity; the performance curve is a plot of total pressure rise and input power vs. ow rate all
at constant rotative speed and density. We are not able to measure the power in the present
experiment. In American industrial practice the pressure rise is given in inches of water and
the ow rate in cubic feet/minute and the pressure rise and power are corrected to a standard
density for air (0.075 lbm/ft
3
or 1.205 kg/m
3
) corresponding to 0.1013 MPa pressure and
293 K temperature. In this experiment we will use SI units: pressure rise in Pascals and
ow rate in cubic meters/second.
The performance curve for the present experiment is to be measured for three different rota-
tive speeds. At each speed, make measurements at a sufcient number of ow rates to map
the performance curve; start at a wide-open throttle setting and closing the throttle toward
the shut off then increase the throttle opening until the fan returns to normal operation. Use
the inlet nozzle pressure difference to measure the ow rate. As the throttle plate is closed,
the fan will eventually reach a point at which the head rise abruptly and drastically falls.
This is a result of rotating stall, an instability in compressors. The stall point should be
determined for decreasing ow rates, and the bottom of the stall region should be found as
well as shut-off point. As the throttle plate is reopened, the compressor will recover from
stall (stall-recovery point), signaled by a rapid pressure rise. It is useful to keep track of the
throttle opening to enable careful mapping of the stall point and the stall-recovery point.
3.4. Instrumentation 21
3.4.4 Unsteady Behavior
At some operating conditions the performance of the fan becomes unsteady and the mea-
sured pressures will oscillate. In particular, at the peak of the performance curve (just
before stall) and in deep stall (the bottom). These pressure oscillations are indicative of a
phenomenon known as rotating stall.
The onset of this condition is best visualized in the blade-xed frame of reference. As the
throttle opening is reduced, the ow rate through the compressor is decreased, so the axial
velocity v is reduced. However, the tip speed of the compressor blades remains constant,
since the compressor is running at constant speed. Therefore, the angle of incidence of the
air ow with respect to the blades is reduced translating to an increase in the angle of attack
of the blades. At larger angles of attack the blades will stall (ow will separate from the
upper side of the airfoil shape) just as an aircraft wing will stall at large angles of attack.
When a blade stalls, the ow that normally passes through the now-stalled region is diverted
to the neighboring blades. The angle of attack of the preceding blade then falls and the angle
of attack of the following blade rises. This decreases the tendency of the preceding blade to
stall while increasing the tendency of the following blade to stall. Therefore, in the blade-
xed frame of reference, the stalled region of the blade row (stall cell) appears to move
in the opposite direction of the compressor rotation. In the lab-xed frame of reference
the stall cell rotates in the same direction as the compressor, but at a lower speed. As the
stall cell rotates around the inlet tube it creates an oscillation in the pressures measured at
the pressure ports. From a measurement of the oscillation frequency of the pressure, it is
possible to determine the speed of the rotating stall cell.
3.5 Measurements
These are the measurements that are to be made on the apparatus:
Air density For the purpose of this experiment we may assume the air to be a
perfect dry gas with a gas constant of 287 J/kgK (in practice we measure the abso-
lute humidity and so determine the effect of water vapor on the density). The local
barometric pressure and temperature are measured on the wall-mounted barometer to
calculate the density.
The pressure rise across the fan is small compared to the absolute pressure (less than
2 percent) so that the ow may be considered as effectively incompressible. The
best effective density is the average of the inlet and discharge densities. For the
purpose of this experiment, the inlet stagnation density may be used as the reference
density.
Calibrate the Differential Pressure Transducers Both pressure transducers should
be calibrated using the water-equivalent differential manometer. A lab View VI (Vir-
tual Instrument) is provided on the data acquisition system to facilitate this process.
The throttle plate is set to a small opening (about 1 cm), the number of desired points
for the calibration is entered into the appropriate eld on the VI control panel, and
the color code of the transducer to be calibrated is selected. Upon running the VI,
the user will be prompted to enter the pressure reading from the manometer into the
appropriate eld and click the continue button. The compressor speed will be ramped
under computer control and the user will be prompted for pressure readings until the
desired number of points is acquired. The VI will determine the best t line to the data
and prompt the user for a pressure that encompasses your expected measurements.
Performance Curves Measure performance curves of the compressor at three op-
erating speeds. Be sure to carefully resolve the stall onset, stall bottom, and stall-
recovery points. A Lab View program may be available to aid you in this process.
Repeatability For one of the compressor speeds used above, repeat the perfor-
mance curve measurements two more times to assess the repeatability of the mea-
surements. Base this repeatability on how well you can reposition the throttle plate.
It is not necessary to repeat every point on the original performance curve, choose
about 5 points that are distributed throughout various portions of the performance
curve.
Unsteady Behavior During the measurement of one of the performance curves,
record time series of oscillatory pressure near stall onset and in the bottom of stall to
determine the frequency content of these signals. A Lab View VI may be available to
aid you in this measurement. If not, temporary use of the oscilloscope to determine
the frequency content is recommended. Look at both the frequency and amplitude.
Speed Curves (Optional) If time and group interest permits, modify a copy of
the LabView VI used for calibration to make measurements of performance along a
speed curve. At a xed throttle setting (you can repeat for several throttle settings
if you like), step through a range of compressor speeds and measure the differential
pressure. Be sure to wait long enough at each speed setting before measuring the
pressures. Take note of any apparent stall phenomena and the throttle opening(s)
used . . . it may be necessary to take longer time averages of pressure under oscillating
conditions to get a measure of mean performance. This exercise is intended to give
you an opportunity to explore the use of LabView in a data acquisition and control
situation. If you are unable to achieve measured results in a reasonable time, dont
fret, view this as an opportunity to learn more about LabView. Remember this is
optional and it is up to you whether you include any of this in your lab report. If you
do include speed curves in your report, print out your LabView VI for the appendix.
3.6 Report
Your results should include:
1. For each of the three speeds, plot p vs. Q; indicate by error bars the uncertainty of
the measurements. Use estimates of measurement uncertainties of the instrumenta-
tion along with results from your repeatability measurements.
2. Using the scaling laws outlined in Section 2, non-dimensionalize the performance
curves plotted above to test the similarity of the compressor performance.
3. Carefully identify the stall, stall-bottom, and stall-recovery points on your perfor-
mance curves.
3.6. Report 23
4. Speed Curves, if you decide to measure and include them. Identify points where stall
or rotating stall were observed.
Your discussion should include:
1. Describe the performance curves noting any features of special interest.
2. Discuss the results of the scaling calculations.
3. The design ow rate for the fan is 0.566 m
3
/s (1200 ft
3
/min) when the exciting fre-
quency is 400 Hz. Predict the performance of this compressor, based on a similarity
discussion, using your acquired data.
4. If you decided to measure and include speed curves, describe these curves and note
any features of special interest.
Figure 3.3: Test Set-up Schematic
3.7 Operating the System
There are several details concerning the actual operation of the fan rig which require atten-
tion before actually attempting to operate the system.
The axial fan, an Able 29680, is driven by an electric motor. It is designed to run continu-
ously at 11700 rpm, to draw 2700 Watts of power, and 10 Amps of current. Its commercial
use is in the cooling of the avionics (the control circuitry) of an MD-11 Aircraft.
In the ME 96 Experiment, you will be running the fan at speeds slower than this. (When
running full speed, the fan produces a great deal of noise). The power supply, since it
is variable frequency, can run the fan at most any speed. This is done by changing the
frequency of the input power, which (since the fan uses an induction motor) will change the
rotor speed. In fact, the fan impeller will run at a speed half the speed of the input power
frequency. Standard airplane power frequency is 400 Hz, which corresponds to a fan speed
of 12000 rpm (nominal) which is 200 Hz. The power supply outputs power up to 400 Hz.
To actually operate the system, follow these steps:
1. Turn on the LARGE wall switch to allow current to ow to the power supply (motor
controller). Then turn on the power toggle switch on the tethered remote control box
mounted to the same table as the axial fan.
2. Use the switch on the side of the remote control box to select manual or BNC (re-
mote) control of the fan speed. Under manual (knob) control turn the potentiometer
3.7. Operating the System 25
knob until the desired frequency appears on the LED readout on the power supply
(motor controller). Under BNC control, a voltage (0 8 Vdc) is supplied to the BNC
connector (usually from a D/A channel of the data acquisition system) to set the drive
frequency in the range of 0 to 350 Hz.
3. Turn off the power toggle switch on the tethered remote control to shut down the fan.
This will signal the power supply to automatically ramp down the drive frequency.
4. Be sure to turn the LARGE wall switch off only AFTER the fan has come to a com-
plete stop.
Note that the fan speed may be changed while the fan is running by simply following steps
2 and 3. The power supply will automatically ramp the voltage in the appropriate way.
Please see the precautions section for cautionary information.
3.8 Precautions
The fan rig uses a signicant amount of power and includes some very high speed rotating
machinery. Therefore it is very important to handle the entire system with care to avoid
damaging the setup and, more importantly, to ensure operator safety. There are several
precautions to take whenever using the setup which are described in the following list.
1. Do not let anything come close to the inlet of the nozzle. Loose items could be sucked
into the impeller, damaging it. A ow straightener is located just inside the nozzle
to prevent large items from entering, but these large items may become lodged in the
inlet. Items small enough to t through an 8 mm square hole could pass through the
straightener and damage the impeller.
2. Do not turn off the LARGE switch on the wall that provides power to the power
supply (motor controller) before shutting the fan off from the tethered remote control.
Turning off the power switch on the tethered remote control causes the power supply
to ramp down and safely shut off the fan.
3. There are several wires and hoses which run from the facility to the power supply,
monometer, and computer. Use caution when moving around the rig to avoid disturb-
ing these connections.
4. Do not run the fan with the throttle closed for any signicant amount of time. The fan
requires ow through it to expel the generated heat. Running with the throttle closed
can cause the fan to overheat.
5. Avoid running the fan at too low a speed (say lower than 66 Hz drive frequency or
2000 rpm). This may not produce enough ow for the fan to cool properly.
6. Do not start the fan in a throttled condition. Starting up with high losses in the system
(e.g. a closed throttle) can cause undue stress.
7. Avoid leaning on, sitting on, or manipulating the fan. This may cause misalignment
of the assembly.
8. When adjusting the throttle plate, be sure to avoid positioning your head (especially
your eyes) in the ow. The ow velocity is high and the potential for injury exists.
9. Wear hearing protection while operating the experiment. Headsets are provided
in the room.
3.9 Prelab
It may be helpfuul to refer to a uid mechanics textbook that has a chapter on turbomachin-
ery, such as ?. (There are some in 0018).
1. Derive the Euler Turbine Equation.
2. Discuss what is meant by geometric ow similarity, and draw a sketch showing the
velocity vectors.
3.9. Prelab 27
28
CHAPTER
FOUR
Free and Forced Convection
4.1 Introduction
The term forced convection refers to heat transport that results from uid motion caused by
external means such as a pump, a fan, or atmospheric winds. The ow velocity depends on
the uid mechanical properties of the system and not (ideally) on the heat transfer processes
that occur in the system. Typical examples of forced convection include forced-air heating
of homes, liquid lled active solar heating systems, and nuclear reactor cooling (during
normal operation).
For free or natural convection, the ow velocity depends on both the uid mechanical
properties of the system and the heat transfer processes that occur. Free convection is
driven by buoyant forces that result from density differences in the convecting uid. In
most situations, the density gradients are caused by temperature variations in the uid.
Examples of free convection include a single-phase closed loop thermosyphone (roughly
the condition in a nuclear reactor immediately after pump failure), a Trombe wall (a passive
solar heating device), and the shimmering visible above a paved highway on a hot summer
day.
The essential difference between free and forced convection manifests itself in the govern-
ing equations of the two convection modes. The boundary layer equations for laminar free
convection over a heated vertical surface are (neglecting viscous dissipation and pressure
gradients):
Energy: u
T
x
+ v
T
y
=
2
T
y
2
Momentum: u
u
x
+ v
u
y
=
2
u
y
2
+ g(T T
)
(4.1)
where is the coefcient of thermal expansion, = 1/(/T)p. The important
point to be noted about these equations is that they are coupled. For forced convection the
momentum equation does not include the buoyancy term, and hence does not depend on the
temperature gradients within the ow.
29
Figure 4.1: The free and forced convection experiment.
Figure 4.2: Experimental apparatus.
30 Chapter 4. Free and Forced Convection
Figure 4.3: Plate and cylinder thermocouple and heater layout.
4.2 Experimental Apparatus
This experiment is designed to study both free and forced convection heat transfer using a
heated plate and a heated cylinder. A schematic of the experimental apparatus is shown in
Fig. 1. The plate and the cylinder are shown in Fig. 2. Either one can be mounted above the
fan and the honeycomb ow straightener. The plate can be oriented at different angles to
the oncoming ow. The plexiglass side walls are to minimize disturbances from the room.
The heat input is regulated by separate control of three nominally 1 in. 5 in. kapton-
insulated exible heaters. The power dissipated through each heater is controlled using a
ten-turn potentiometer mounted in the control panel. The power can be calculated from the
voltage input as displayed on the control panel, and the heater resistance (601%). The
maximum input voltage is 21 volts. The central heater serves as the experimental heater,
and the two side heaters act as guard heaters. The power input to the guard heaters should
be adjusted so that there is no lateral temperature gradient across the central heater, and
hence no heat loss from the central heater to the sides and the metal supports. Typically,
the voltage input to the side heaters should be 3 to 7 volts higher than to the central heater.
The temperature gradient across the central heater can be monitored using the three aligned
thermocouples.
Chromel-alumel (type K) thermocouples are attached to the heated surfaces as shown in
Fig. 2. There is an additional thermocouple to measure the ambient temperature. The ther-
mocouples are connected to a USB data acquisition unit designed specically for acquiring
thermal data from thermocouples and other temperature sensors. Thermocouples 1 through
6 are connected, in numerical order, to channels 0 through 5 of this unit. A thermocouple
to measure the ambient air temperature is connected on channel 6.
In addition to the thermocoupls, a heat-ux gauge is mounted on one surface of the plate.
The calibration for the heat ux gauge is in the lab folder. The output for the heat ux gauge
is read on the analog microvolt meter. In the range of 03 mV, the meter has an accuracy of
5%.
The fan is controlled by a DC power supply. The ow velocity is measured using a hot-wire
anemometer. The anemometer can be inserted through the collar mounted on the plexiglass
wall. Be careful not to break the wire, which is very fragile!
4.3 Data Acquisition
The acquisition of the thermocouple data for this experiment is to be done using LabVIEW.
In addition to the primary objetive of investigating convective heat transfer phenomena, a
secondary objective of this experiment is to teach you something about modern, computer-
ized data acquisition methods.
You must have worked through the LabVIEW tutorial or completed the LabVIEW home-
work assignment before beginning this lab. If you have not yet done so, do this before
meeting with the TA to begin this experiment.
You are provided with a small USB data acquisition unit that is designed specically for
acquiring thermal data from thermocouples, thermistors, or other types of temperature sen-
sors.
But you should congure it in software. Run the InstaCal program (under Measurement
Computing on the Windows Start menu) on the lab PC to congure, calibrate, and test
this unit. Make sure the channels you will use are congured for Type K thermocouples.
Within InstaCal, you can access from the Help menu the manual for this unit (model USB-
TEMP). Once you are condent the unit is operating as expected, you can start writing a
LabVIEW VI to use it to make the desired measurements for this experiment.
Start LabVIEW 8.2, create a new, blank VI, and view the block diagram window. On
the Functions palette, select User Functions, where you will nd the Measurement
Computing drivers. Under Analog Input, you will nd one called Tin. This function
will acquire a temperature reading from the USB-TEMP unit. This should be all you need
from this palette to build your VI.
Your LabVIEW VI should do the following:
1. Continuously monitor the temperature for all thermocouples, and alert you in some
way when the temperatures have stabilized sufciently to begin measurements. (Per-
haps it could generate a sound, or send you an e-mail message, or...)
2. Provide a convenient display of the three horizontally-spaced thermocouples to watch
while adjusting the guard heater voltages to minimize the T in this direction.
Optionally, you can acquire data from the heat ux sensor also using LabVIEW.
4.4 Experiments
4.4.1 Free Convection
Begin by orienting the plate in the vertical position, and do not use the fan. Set the power
to each of the heaters and monitor the surface temperature until it reaches a steady state.
Use power settings close to the maximum value. This warm up process may take about 40
minutes.
After the heater reaches a steady state, record the surface temperatures and the ambient
temperature. Record the voltage input to the heaters, and the output from the heat ux
sensor. From the measurements of the heat input one can calculate the surface heat ux, q
in W/m
2
by dividing the input power by the surface area. Remember that the heater has two
sides. Check to see if the two methods for determining q
agree. If the difference is more

than 10% and the measurement from the heat ux gauge is lower, reduce the heat input to
the central heater.
Without changing the input power, change the plate orientation. Again allow the surface to
equilibrate. Record the temperatures and the voltage output from the heat ux gauge. Try
at least two different orientations.
Natural convection experiments should be performed in quiescent environments. Room
currents will affect the results.
4.4. Experiments 33
4.4.2 Forced Convection
Keep the input power at the same value. With the plate in the vertical position, turn on
the fan to a low speed. Use the anemometer to record the incoming ow velocity over the
central heater. Allow the plate to equilibrate, and record the temperatures and the heat ux.
Record the conditions for at least three different fan speeds. The warm-up time should be
faster than for the free-convection conditions.
Also try at least one experiment with the plate at a different inclination.
4.4.3 Flow Over a Cylinder
Disconnect the plate and connect the cylinder. Orient the cylinder so that there is a thermo-
couple directly at the bottomand at the top. Repeat the measurements for natural convection
and for forced convection using several different ow speeds.
There is no heat ux sensor in the cylinder. Start with the same heat input settings used for
the forced convection experiment; however, the settings may need to be adjusted.
4.5 Heat-Transfer Predictions
The Nusselt number, Nu is dened as follows
Nu =
h
c
k
=
q
c
(T
w
T
a
) k
(4.2)
where h is the heat transfer coefcient,
c
is a characteristic length, T
w
and T
a
are the wall
and ambient temperatures, and k is the thermal conductivity of the uid. The Nusselt num-
ber can be either a local value, Nu
x
where
c
is the distance from the beginning of the plate
to the location of interest (
c
= x) and T
w
is the temperature at that location T
w
= T(x), or
the average value, Nu
a
, using the total plate length (
c
= L) and the average temperature.
The average temperature is often difcult to dene, and so the midpoint temperature (T
w
at
x = L/2) is often used.
For forced convection, the Nusselt number is presented as a function of Reynolds number,
Re = u
c
/, where u
is the approach velocity, and is the kinematic viscosity. For free

convection, the appropriate number is the Rayleigh number, Ra = g(T
w
T
a
)
3
c
/. The
coefcient of thermal expansion for an ideal gas is equal to the inverse of the absolute uid
temperature. All properties should be evaluated at the lm temperature, T
f
= (T
a
+T
w
)/2.
For laminar forced convection (Re
L
< 5 10
5
), over a constant heat-ux surface, the
theoretical local Nusselt number is
Nu
x
= 0.453Re
1/2
x
Pr
1/3
(4.3)
and the average value is
Nu
a
= 0.679Re
1/2
L
Pr
1/3
(4.4)
Pr is the uid Prandtl number, Pr = 0.7 for gases.
For laminar free convection of air over a vertical plate (0 < Ra
L
< 10
9
), the average value
is
Nu
a
= 0.68 + 0.52Ra
1/4
L
(4.5)
For ow over a heated cylinder, the characteristic length is the cylinder diameter. The
Nusselt number based on the cylinder diameter for air ow is (10
5
< Ra
D
< 10
12
)
Nu
a
=
_
0.6 + 0.32Ra
1/6
D
_
2
(4.6)
For forced convection, the Nusselt number for air ow is (Re
D
Pr > 2)
Nu
a
= 0.3 + 0.54Re
1/2
D
Pr
1/3
_
1 +
_
Re
D
28200
_
5/8
_
4/5
(4.7)
Remember that the heated surfaces lose heat by conduction, convection and radiation. The
effects of conduction are minimized using the guard heaters. The radiation loss, however,
should be accounted for. The radiation loss from a surface can be estimated from the fol-
lowing equation:
q
= (T
4
w
T
4
a
) (4.8)
where is the surface emissivity, and is the Stefan-Boltzmann constant.
4.6 Lab Report
1. For the natural convection results, plot the average Nusselt number as a function of
inclination angle. For each inclination you can obtain two Nusselt numbers corre-
sponding with positive and negative angles from the vertical. Also show on the graph
the theoretical value for a vertical plate. Why are there differences?
2. For the forced convection results, plot the average Nusselt number as a function of
Reynolds number. Show theoretical values. Also indicate results from the different
orientations.
3. For the cylinder, plot the Nusselt number as a function of Reynolds number. Show the
theoretical values. Determine the Nusselt number for natural convection and compare
it to the theoretical value.
4.7 Discussion
In the discussion section of the lab report, the following issues should be addressed.
1. How large is the radiation contribution for each of the ows?
2. How well do the theoretical and experimental values compare? What are possible
reasons for the differences?
3. How does the angle of attack of the plate affect the heat transfer?
4.6. Lab Report 35
4.8 Prelab
Before beginning the lab, you need to have worked through the LabVIEW tutorial on the
wiki, and have read the thermocouple tutorial, also on the wiki.
Answer the following questions in your lab notebook before the rst lab session.
1. What is the Prandtl number of air at 300 K? What is it for water?
2. Estimate the Rayleigh number for air at 1 atm pressure, 300 K, for a length scale of
1 cm, and a temperature difference of 10

C. Repeat the calculation for water.
3. Using the parameters in the last question for air, calculate the heat loss by free con-
vection from a 1 cm radius heated cylinder. How large an air speed transverse to the
cylinder would be required to achieve the same heat ux by forced convection?
4. Describe how a hot-wire anemometer measures air speed.
5. How do you think the heat ux gauge works?
It may be helpful to refer to a heat transfer textbook, such as ? or ?.
CHAPTER
FIVE
Turbulent Air Jet
5.1 Introduction
If you have taken a course in uid mechanics, you probably learned all about laminar ows
laminar boundary layers, laminar ow in tubes, etc. Unfortunately, in most real engineering
applications this is not at all the sort of ow that is observed. Instead, the ow is turbulent,
consisting of a cascade of eddies, resulting in a chaotic, unsteady ow.
Turbulent ow is much more difcult to characterize than laminar ow, and although a
great deal has been learned about many of the properties of turbulent ows, we still cannot
predict from theory some features of turbulence that are easy to observe. For example,
the spreading angle of a turbulent jet (see Figure 5.1) is easy to measure experimentally,
but cannot be predicted from theory, even though we know the governing equations of
the jet (the Navier-Stokes equations)! One problem is that a direct simulation of the ow,
resolving even the smallest eddies, is still too large a calculation even for the fastest parallel
computers today, at least for Reynolds numbers of practical interest. So much effort has
gone into developing approximate, statistical theories of turbulence, but so far there is no
approximation that is good for all turbulent ows.
Much of what we know about turbulence comes from careful study of model ow problems.
One of the classic model problems is that if a round jet, issuing into still air. This is the ow
that you will investigate in this lab.
Figure 5.1: A turbulent jet.
37
Figure 5.2: Notation for the jet experiment.
5.1.1 Background
For any turbulent ow, the instantaneous velocity uctuates in time at every point in the
ow. But for steady turbulent ows, the time-averaged mean velocity is independent of
time:
u(x, r) =
1
T
_
T
0
u(x, r, t
)dt
. (5.1)
Here the averaging period T is chosen to be long enough to capture the motions of very
many eddies. For a truly steady turbulent ow, u is independent of T, as long as it is
sufciently large.
We dene the uctuating or turbulent velocity by
u
(x, r, t) = u(x, r, t) u(x, r). (5.2)

Of course, by this denition
_
T
0
u
(x, r, t
)dt
= 0. (5.3)
We will use the notation shown in Fig. 5.1.1. All mean ow quantities are assumed to be
axisymmetric, so that all mean quantities depend only on the axial distance downstream of
the nozzle x as the radial distance r from the jet centerline. We will use (x) to denote the
jet width at a given axial distance. While this could be dened in many ways, it is usually
taken as the radial distance from the jet centerline to the point where the mean axial velocity
has dropped to half its value on the centerline (half-velocity point).
For a narrow jet issuing into still air, we can make the following assumptions:
1. The ow is nearly parallel and the inclination of the (mean) streamlines is small.
This assumption requires v u, or, equivalently, a negligible transverse pressure
gradient.
2. The jet is issuing into innite space, and therefore there is a negligible axial pressure
gradient. (If, however, we were to place an object into the jet, the pressure would rise
38 Chapter 5. Turbulent Air Jet
as the stagnation point on the leading edge is approached, in order to decelerate the
ow and bring it to rest there.)
These assumptions clearly dene a uniform pressure eld. Further, we assume the jet mo-
mentum ux remains constant at all x:
J = 2
_

0
u
2
rdr = constant = J
0
(5.4)
where is the density.
We want to determine the jet growth law that is, how and u(x, 0) depend on x. We
can determine the functional form, just from dimensional analysis, without even solving
the governing equations (which are too hard to solve anyway).
An important non-dimensional parameter in any ow is the Reynolds number. By consid-
ering the momentum ux at the nozzle exit, it is easy to show that the Reynolds number for
a round jet can be written as
Re =
J
2
(5.5)
(to within a factor of /4, which for the present analysis is close enough to 1 to neglect).
The Reynolds number is the only non-dimensional parameter that we can form from molec-
ular parameters (, ) and conserved quantities (J).
Now consider how to non-dimensionalize . The only available length scale to use as a
reference is x itself (we assume we are far enough downstream of the nozzle that the nozzle
width is not an important parameter), so we conclude that
x
= f
n
(Re). (5.6)
(Here f
n
means a function of).
Dimensional analysis tells us that /x can be written as some function of J/
2
, but it cant
tell us what the form of that function is. For a turbulent ow, we are particularly interested
in the limiting behavior for J/
2
1. There are three possibilities /x might diverge
as J/
2
, or it might go to zero, or it might asymptotically approach a constant.
We can rule out the second possibility right away, since we know the jet spreads, it does
not contract. We can also rule out a diverging function, since in that case very small values
of viscosity would have a large effect on /x. In reality, most aspects of a turbulent ow
become independent of , since the churning eddy motion of turbulence is a much more ef-
fective way to disperse uid momentum than is molecular viscosity. So the only possibility
that is consistent with observations is that the function approaches a constant. Therefore, in
a turbulent ow,
x
= C, (5.7)
where C must be determined from experiment.
Now to non-dimensionalize u(x, 0), we need to form a quantity with units of m/s from the
available independent variables and parameters. We could choose /x, which would be
appropriate for a laminar ow, but this would be a poor choice for a turbulent ow, since
we know is not an important parameter for most aspects of a turbulent ow. We can form
a quantity with velocity units that is independent of viscosity using the Reynolds number:
(/x)
_
(J/
2
) =
_
J//x.
Therefore, we construct the non-dimensional velocity
u(0)x
_
J/
, (5.8)
and require that it too, like /x be a function only of J/
2
. And, following the same
arguments as above, we require that this approach a constant in the limit of high Reynolds
number, with the result
u(0)x
_
J/
= B, or u(0) =
B
x
. (5.9)
The constant B must also be determined from experiment.
Equations (5.7) and (5.9) showthat the jet width increases linearly with x, and the centerline
velocity decreases with x
1
. The constants of proportionality, C and B, can be determined
experimentally.
The same arguments we used for u(x, 0) could have been applied to the mean velocity
at any other radial location u(x, r), with the only difference that we would have had to
introduce another non-dimensional parameter x/r.
As a result, we can write
u(x, r)
u(x, 0)
= f
1
_
r
x
_
or u
= f
1
(r
) (5.10)
which denes u
and r
.
The concept of self-preservation in turbulent ow requires the turbulent time scales be much
smaller than the mean ow time scales. Again we normalize with the local conditions to
give:
u
RMS
u
m
= f
2
_
r
x
_
(5.11)
where u
is the RMS value of the uctuating velocity component, dened as

u
RMS
=
_
1
T
_
T
0
(u
)
2
_
1/2
(5.12)
and u
m
is the centerline velocity.
Note that since we neglected the nozzle diameter in this analysis, the jet velocity prole
close to the nozzle will not be similar to proles far downstream, because near the nozzle
the diameter is an important parameter. The ow does not become self-preserving until we
reach the far eld of the jet.
5.2 Instrumentation
In the experiments, a pitot tube and a hot wire anemometer will be used to measure the
velocity. The pitot tube is robust and inexpensive, and the anemometer is fragile and ex-
pensive. The benet of the anemometer is that the transient characteristics of the ow can
Figure 5.3: Airjet experiment.
be measured, which cannot be obtained with the pitot tube. In the experiments take great
care in handling the anemometer. If the wire is broken, it must be sent back to the vendor
for repair, which is expensive and time consuming.
5.2.1 The Pitot Tube
The pitot tube measures the total or dynamic pressure of the ow. At a point in the ow,
the relation between the total pressure, P
t
, the static pressure, P, and the velocity, u, can be
determined from Bernoullis equation:
P
t
= P +
1
2
u
2
(5.13)
where is the density of the uid. The static pressure is the pressure measured for no-ow
conditions. The total pressure is measured using a manometer.
5.2.2 The Hot-Wire Anemometer
The hot wire anemometer is based on principles of heat transfer. An energy balance on the
electrically-heated wire yields,
V
2
R
= h(u)A(T
w
T
a
) (5.14)
where V is the voltage, R is the wire resistance, h is the heat transfer coefcient (which
depends on the ow speed u), A is the wire surface area, and T
w
and T
a
are the wire and
air temperatures.
For a constant-temperature anemometer (CTA), the CTA electronics use closed-loop con-
trol to maintain the wire resistance at a xed value (and thereby maintain the wire tem-
perature xed) by continually adjusting the electrical current to the wire is continuously
5.2. Instrumentation 41
adjusted to maintain the resistance of the wire at a xed value. Therefore, if the ow is
isothermal, then the temperature difference between the wire and the ow is a constant.
The CTA voltage is then given by
V
2
= Rh(u)AT (5.15)
The dependence of the heat transfer coefcient on ow speed can usually be represented by
h = c
1
+ c
2
u
n
(5.16)
where n 1/2. In this case, the speed is related to the measured voltage by
u = A
1
_
V
2
V
2
0
_
1/n
(5.17)
where A
1
is a constant, and V
0
is the voltage reading in still air. If n = 1/2, then this rela-
tionship is a fourth-order polynomial. In practice, a CTA should be calibrated by measuring
the voltage for a few different known speeds (measued, for example, with a pitot probe),
and t to a polynomial.
5.2.3 Data Acquisition and Stage Control
LabView will be used to acquire the data from the hot-wire anemometer and from the Pitot
probe. The probes are mounted on two orthogonal motorized translation stages, allowing
them to be positioned vertically and horizontally under computer control. The data acqui-
sitioning and stage control are accomplished with the same VI (provided).
The main parameters of the DAQ that can be changed are the sampling frequency and the
number of samples. Note that two channels are connected, so the number of samples for
each channel is different.
The VI controls the horizontal and vertical stages on which the probes are mounted. The
stages can be controlled in 3 modes:
1. Set the horizontal spacing, vertical spacing, and dwell time, and let the VI control the
system taking measurements at each location automatically
2. Control the stages manually using buttons
3. Control the stages manually using command-line inputs
The speed of the stages can be set, with an available range of 0.025-20 rev/s. This can
be controlled either from the System Conguration tab, or from the command line on the
Command tab. The set of available commands for command line control is given in table.
Note that the encoders measure relative to some home location. Thus, to command the
stages to a specic location, the stages must be commanded to the Home location.
5.3 Experiment
In this experiment, you will use a hot-wire anemometer and pitot tube to measure velocity of
ow in a turbulent jet. The hot-wire anemometer will be used to measure the both transient
and mean velocity prole, while the pitot tube is used to measure the mean velocity. The
pitot tube and anemometer need to be calibrated. The anemometer is used to measure the
shedding frequency from a bluff body immersed in the ow. The owvelocity prole (mean
and uctuating velocities) are measured at different distances from the nozzle.
5.3.1 The Pitot Tube Calibration
The pitot tube is connected to a pressure transducer which converts the pressure to a volt-
age. This transducer is in turn connected to the DAQ board. The conversion factor from
transducer voltage output to ow velocity, u(V ), needs to be measured. To do this, manu-
ally control the stages to position the pitot tube in the center of the jet just above the nozzle.
Take care that the probe do not hit the nozzle! Take total pressure measurements, P
t
, at
different ow speeds (controlled by throttling the air intake). One measurement should be
taken with the fan off (to get the static pressure, P). The total pressure should be read off
the manometer (the manual is the the lab-folder). At the same time, the voltage should be
recorded. The ow velocity can be calculated as:
u =
_
2(Pt P)/ (5.18)
1. Record data.
2. Using Excel, MATLAB, or other means of your choice, generate an appropriate poly-
nomial curve t for u(V ).
3. Include a plot of this t and your velocity vs. voltage data in your lab report. Make
sure the curve is a good t to the data, and is monotonic.
5.3.2 The Anemometer Calibration
The objective is now to use the calibrated pitot tube to calibrate the anemometer. Carefully
position the anemometer just above the nozzle by traversing the stages vertically and hor-
izontally. Measure the time-average voltage from the anemometer. Move the pitot tube to
the same position and measure the ow velocity. The Pitot tube has a slow time response,
since the tube is small. Make sure you wait long enough that the pressure signal has stabi-
lized. Make sure to take measurements at several different intake valve throttle positions,
ranging from wide-open, to nearly closed. Also, take one CTA voltage measurement with
the pump off.
1. Record data.
2. Using Excel, MATLAB, or other means of your choice, generate an appropriate poly-
nomial curve t for u(V ).
3. Include a plot of this t and your velocity vs. voltage data in your lab report. Make
sure the curve is a good t to the data, and is monotonic. If you nd you need more
data points to determine the polynomial, take more readings until you are satised
with the accuracy of your curve t.
4. Enter the polynomial coefcients for your t into the appropriate place in the VI
(System Conguration tab).
5.3. Experiment 43
5.3.3 Vortex Shedding
Vortex shedding from a bluff body immersed in a cross ow leads to a periodic disturbance
at the vortex shedding frequency f. The Strouhal Number S
r
is dened as
St =
fD
U
(5.19)
where f is the vortex shedding frequency, D is a characteristic length, and U the ow speed.
For a long cylinder, with D equal to the cylinder diameter, the Strouhal number is given by
the empirical formula
St = 0.198(1 19.7/Re) (5.20)
At some height in the ow, place a stretched-out paperclip or other wire that can be modeled
as a cylinder across the ow. Two mounted wires are provided for this purpose. Position
the anemometer probe approximately 5 cylinder diameters behind the cylinder using the
command-line control of the probe. With the anemometer you should be able to measure
the turbulent uctuations caused by the cylinder, and by examining the power spectrum you
can determine the shedding frequencies. Theoretically, the frequency should correspond to
a Strouhal number of approximately 0.2 where S = fD/u and f is the shedding frequency,
D is the cylinder diameter, and u is the approach velocity of the air stream.
The provided VI has a frequency analysis tab to do this analysis. .
1. Experiment with the averaging in the frequency analysis to see the effect on the plot.
2. Record the list of command-line inputs to move from the home location to the appro-
priate location.
3. Compare the power spectrum with the cylinder in the ow to that obtained without
the cylinder.
4. What is the experimental Strouhal number?
5. Repeat frequency measurement for different ow velocities and plot the measured
shedding frequency vs ow velocity.
5.3.4 Mean and Fluctuating Velocity Proles
Adjust the pump inlet to an intermediate value and leave it at that position for the remainder
of the experiment. Using LabVIEW, traverse the probe across the jet at a xed height and
measure proles of the mean and uctuating velocity. Repeat the measurements for several
different heights, from the near-eld just above the nozzle, out into the far eld (or as high
as you can go with the translations stage).
Make the measurements at a sufcient number of heights so that you can generate a 2D
map of mean velocity and uctuating velocity. The scans do not need to be equally spaced
in height you will probably need more closely spaced ones near the nozzle.
1. Record mean velocity vs position data for both the anemometer and the pitot tube.
2. Record uctuating velocity vs position data for the anemometer.
Compare your results to expectations based on theory discussed above. Your discussion
should include, but not necessarily be limited to, the following points:
1. Test whether self-similar mean velocity proles are obtained: make a plot of
u(x, r)/ u(x, 0) as a function of r/x, where u(x, 0) is the centerline velocity at height
x. Include the results for multiple heights on the same plot. If the proles are self-
similar, they will collapse to a single curve.
2. Do a similar analysis of the uctuating component u
. Does the uctuating velocity

become self-similar at the same distance from the nozzle as does the mean velocity?
3. Is the jet momentum actually constant, as assumed in the theory?
4. Does the far-eld jet width grow with distance as expected?
5. How far downstream (measured in number of nozzle diameters) do you need to go to
reach the far-eld?
6. What is C? What is B? Compare to values from Ref. 1: C

= 0.084 (Fig. 1), B

= 6.1
(Fig. 3). To nd B, use J
0
= u
2
0
r
2
0
where r
0
is the jet radius and u
0
is a constant
velocity across the diameter.
5.4 Prelab
1. Describe how a constant-temperature hot-wire anemometer works. Why is the wire
so ne?
2. Calculate the Reynolds number based on jet diameter of an air jet with nozzle diam-
eter = 2 cm and velocity = 40 m/s.
3. Calculate the expected vortex shedding frequency for a 1 mm diameter cylinder im-
mersed in a 30 m/s ow.
4. Modify the spectralTurbAirJet.vi to perform a frequency analysis on a signal gener-
ated using the function generator.
Connect the function generator to channel 1 of the NI Scope DAQ on the can-
talever beam experiment.
Generate a sinusoidal signal at roughly 2000 Hz.
Modify the VI to perform a spectral analysis on the signal and plot the output.
Include a printout of the modied front-panel and block-diagram (screenshot).
Verify that the spectral analysis gives reasonable results (since we know the
input frequency and amplitude).
Experiment with the sampling frequency and number of samples to see effect
these have on the measured frequency. Include at least one sampling rate that is
below 4000 Hz. Plot this data.
5.4. Prelab 45
5.5 Air Properties
T (
C) Density (kg/m
3
) Viscosity (Pa-s)
0 1.294 1.72 10
5
50 1.093 1.95 10
5
100 0.947 2.17 10
5
References
1. Wyganski, I. and Fielder, H. (1969) Some Measurements in the Self-Preserving Jet,
JFM, Vol. 38, Part 3, pp. 577612.
2. Sabersky, R. H., Acosta, A. J., and Hauptmann, E. G. (1989) Fluid Flow, 3rd edition,
MacMillan Publishing Co.
3. Dantec literature on hot-wire anemometry.
4. Schlichting, H. (1979) Boundary-Layer Theory, 7th edition, McGraw-Hill, pp. 21
32.
5. White, F. M. (1991) Viscous Flud Flow, 2nd edition, McGraw-Hill, pp. 470476.
5.6 Appendix: Motion control
A goal of this lab is to measure the air velocity as a function of position above the nozzle
i.e. the velocity eld. To start, we can assume that the jet is axially symmetric so that we
only need to measure in single plane above the edge of the nozzle. As you would expect,
we will dene the r direction as perpendicular to the symmetry axis of the nozzle (the ow
direction) and passing through the center of the nozzle. We will dene the x direction as
parallel to the symmetry axis and also passing through the center of the nozzle. These axis
will form the r-x plane in which we make our measurement.
We will measure the air velocity using two instruments. The rst probe is a hot-wire
anomometer, which is a direct measure of air speed, and the second probe is the Pitot
tube, which is measure of the local air pressure and is an indirect method of determining
the local air velocity. The Pitot tube is connected in parallel to an electronic pressure trans-
ducer and an old-fashioned water manometer. Dont be fooled. The latter pressure gage is
much more sensitive than the electronic gage! Be aware that the dynamic response of the
Pitot is much less than than that of the hot-wire anomomter. Both the hot-wire anonometer
and the electronic pressure transducer are fed to the USB DAQ card on channels 1 and 0,
respectively.
Since both the hot-wire anonometer and the Pito tube measurements are point-wise i.e. the
velocity in the immediate vicinity of each the probe, we will need to sample the r-x plane to
fully determine how the velocity eld and how it changes with the addition of disturbances.
Sampling of the velocity eld could of course be done in an arbitrary point by point manner.
However, it is interesting to measure a velocity prole of the jet. That is to say, the change in
the velocity in the r direction as the position of x is held constant. An example of a velocity
prole is shown in Figure ??. This velocity prole was taken close to the nozzle, i.e. small
x, and the ow region is well dened and uniform. Note that in this case we measured
the air velocity with time as the stage moved in the r direction; the velocity proles you
measure will of course need to be measured with respect to position. As you will see later,
velocity proles are a powerful tool for characterizing the air jet. By combining velocity
proles for a range of x, we can construct the entire velocity eld of the jet.
In past ME96 labs, velocity proles were measured by attaching the probes to a translation
stage that moved in the r direction. The stage was driven by a hand crank and the position
was read from a mechanical indicator. The x position could only be set much less pre-
cisely using a coarse rail guide. This year we have replaced both the manual r-x translation
mechanisms with computer controlled stepper motor linear translation stages. These will
allow you to make very accurate measurements of the eld, and because they are computer
controlled the entire measurement process can be automated using LabView.
5.6.1 Motion Control
Stepper Motor Linear Translation Stage
A stepper motor linear translation stages is shown in Figure ??. Conceptually, the stage
consists of a carriage (i.e. where the probes are mounted) attached to a pair of rails via
linear bearings. The linear bearings allow the carriage to slide along the rails with minimal
friction. (A very important feature of linear bearing is that there is little side-to-side play
5.6. Appendix: Motion control 47
or slop so that accuracy is guaranteed.) The motion of the plate is controlled by ball screw
that passes through a threaded mount that is attached to underside of the carriage. The screw
in turn is driven by a stepper motor. As the motor turns, the plate moves precisely along the
rails.
As the name implies, a stepper motor rotates in discrete steps i.e. angular increments. This
is unlike a simple DC motor where rotation occurs in a continuous fashion. The discrete
steps are controlled by current pulses sent to multiple electromagnets inside the motor from
an external controller. Each step is some small fraction of a revolution, and the amount of
rotation will depend on parameters set by the controller for individual motor. The pitch of
the lead screw i.e. the threads per inch, attached to the stepper motor determines how far
the stage will move with each revolution. Using this and the angular rotation per step, you
can determine how far the stage is translated per step. For our stages the screw pitch is 2
mm. The controller can be set between 2000 and 50,800 steps per revolution. However the
absolute resolution of the stage is limited mostly by play in the ball screw to about 50 m.
Stepper Motor Controller
A special controller is needed to activate and manage all the necessary hardware settings
and activation required to control each stage. The stepper controller we are using, the
Si3540, can itself be controlled using commands sent from the computer using the RS232
serial protocol. Si3540 controller, Figure ??, can be thought of as the mediator between the
computer and the stepper motor stage both receiving commands as well as responding to
queries. The controller comes with two canned software packages. The Si Programmer and
the SCL program. The Si Programmer package is best used for setting motor parameters
such as the current and the phase. The motion program is included with this package is
bizarre and should be avoided if possible. The second software package is SCL. This is
simply a command terminal for sending and receiving ASCII commands and responses.
We have written an equivalent LabView VI that you can use for your experiments.
Motion Commands
The Si3540 controller can receive commands as well as respond to queries from the com-
puter. For example the set position command SP can be used to arbitrarily set the position
of the stage in steps. By sending SP0. The step count is set to 0 at that stage position. How-
ever, by sending just SP, the controller will read and send the current position from the
stage to the computer. As mentioned above, commands are sent as ASCII characters. the
return character \r. The communication uses the RS-232 protocal i.e. serial lines. There is
a separate Si3540 controller for each translation stage, and in our setup COM4 and COM5
are used for the x and r axis, respectively.
There are two main modes for operating the motors. Steps can be taken in discrete amounts,
which is in the Si3540 command language is referred to as feed length or steps can be
fed continuously until, which is universally referred to as jogging. To move the stage
a discrete number of steps you will need two commands, the distance index DI and feed
length FL. The FL command moves the stepper motor a number of steps and direction
as dened by the DI command. For example by sending DI-1000. The stage will move
to the left by 1000 steps when the FL command is issued. By sending DI1000 the stage
will move the right by 1000 steps. Finally, by sending DI to the controller, the controller
will send back the current index distance which is this case is 1000. Jogging and more
importantly stopping the jogging requires CJ commence jog to start jogging and then SJ
stop jogging to stop the jogging. It is important to note that the direction of the jogging
will be determined by the last DI command. If DI=-1000, the stage will jog to the right, and
if DI= 1000 the stage will jog to the right. The feed to position FP will jog to an absolute
position. For example if you dene your starting point for your measurements as 0 (SP0)
you can move back to this point by commanding FP0 PROVIDED that the DI is set in the
proper direction. Shown in Table 5.1 are the basic commands you will need to command
the Si3540 controller.
It is important to note that the stages are equipped with hardware limit switches so
that if the stage moves to end of either travel range, the motors will automatically stop.
This is very important to prevent damage to the stage! The switches have a red diode
that lights when limit has been reached.
Computer control of the stage
Although almost any programming language could be used to communicate with the Si3540
controller, we are going to use LabView because it will allow us to easily integrate both the
motion and the measurements into one VI. The VI you will uses is a stand alone executable
i.e. you cannot edit it. It is located in the Windows Run menu. This VI allows you to
create a full 2-Dscan of the velocity eld automatically, record data to text le, congure the
DAQ, and change the stepper motor setting. It also allows you to enter motion commands
to move the stage manually. Your TA or the instructor will cover the program with you. A
screen shot of this Vi shown in Figure ??.
5.6.2 Troubleshooting
Stage does not move after limit stop This can happen if motion of the stage is stopped
by the limit switch and the stage sits idle for more than a few minutes. The x is to
cycle the power on the Si3540. NOTEThe plug must pulled until the red power light
extinguishes.
Constant voltage from the hot-wire anomometer If the voltage reads a constant output
of a few volts and does not change values when you unplug the BNC connection.
The hot-wire is broken. You can verify using a 60X objective microscope.
5.6. Appendix: Motion control 49
Cmmnd Descr. Param. W Only R Only Units Range Def
DI distance or position dstnc steps 16,000,000 20,000
FL feed to length x
FP feed to position x
VE velocity setting speed rev/sec 0.025-50 1
MR microstep resolution 3-16 8
SJ stop jogging x
SP set abs position 16,000,000
ST stop x x 16,000,000
Table 5.1: Brief table of Si3540 commands. Additional commands can be found in the
bound lab folder
MR Code Steps/rev MR Code Steps/rev MR Code Steps/rev
3 2000 8 20000 13 36000
4 5000 9 21600 14 50000
5 10000 10 25000 15 50800
6 12800 11 25400
7 18000 12 25600
Table 5.2: MR code and the actual steps/rev represented
CHAPTER
SIX
Mechanical Properties of Metallic
Materials - The Tensile Test
6.1 Objectives
The tensile test is one of the most important methods for characterizing the mechanical
properties of a metallic material. Several important engineering characteristics of a mate-
rial can be deduced from this test. Examples include: Elastic modulus, yield and tensile
strengths, energy absorbed to failure, ductility, etc. These mechanical properties allow the
engineer to predict how a part will react when placed under various working loads.
The major objectives of this laboratory exercise are:
1. Understand the basic process of deformation due to tensile loading
2. Characterize the mechanical properties of various metals from their stress-strain
curves
3. Determine how annealing a metal affects its mechanical behavior
4. Examine the fractured surfaces using a Scanning Electron Microscope (SEM) and
determine the mechanism of failure.
In this lab you will use the available apparatus (this lab manual will provide you only with
general guidelines on how to operate the equipment. There is also a user manual provided
by the manufacturer of the equipment in the lab that is very useful, along with some tutorial
CDs).
6.2 Test Equipment
The available axial loading system consists of a PASCO AP-8216 Stress Strain Appa-
ratus, a PASCO CI-6746 Force Sensor and a PASCO CI-6538 Rotary Motion Sensor
(www.pasco.com). The combined system gathers the necessary data to analyze stress strain
behavior of material samples using the accompanying software.
The furnace used for annealing is a Lindberg Tube Furnace, which produces heated environ-
ments up to 1150

C. It is next the experiment. The tube is made of quartz. It has stainless
51
Figure 6.1: The stress/strain analysis equipment and test samples.
steel end caps with Swagelok ttings for the nitrogen purge. The exit of the purge is passed
through water to prevent backow. The furnace is controlled by a Eurotherm temperature
controller, and instructions for a temperature ramp are on the class web page.
The materials to be tested are: steel, brass and aluminum
6.3 Test Procedures
1. Get Started.
Plug in the apparatus (there is no on/off switch) and turn on the accompanying com-
puter. Double-click the DataStudio icon, and select create experiment.
2. Set Start/Stop Parameters.
Its important that the forces applied to the PASCO Force sensor not exceed the limit
of the components within the sensor. Click the Setup button near the top of the
screen, and within the experiment setup window click options. In the sampling op-
tions window, go to the delayed start tab. Select Data measurement with Force,
push Positive as the limiting parameter. Set the value so that recording begins when
the force is above 1 N.
Next go to the Automatic Stop tab in the sampling options window. Again select
Data measurement with Force, push positive as the limiting parameter. Set the
limit at 45 N. Take care to watch the software during any testing so that you can stop
cranking immediately after the software stops taking data (when the force registered
reaches your set limit of 45 N).
3. Familiarize Yourself with the Program and Apparatus.
52 Chapter 6. Mechanical Properties of Metallic Materials - The Tensile Test
Learn how to load and operate the testing apparatus (the user manual is a good
place to start). You can also use some of the plastic samples for practice.
Determine what measurements the program can extract from the testing appa-
ratus.
Explore the various ways to manipulate data provided by the program (creating
graphs, determining slopes, calculating functions of measured properties).
Remember:
The extension of the samples can be determined from the angular rotation of the knob
by manually measuring the relationship between them.
As you test samples, the tensile force will cause the samples to stretch. However, the
applied force also causes the apparatus platform and the force sensor to bend. The
displacement being calculated by the software will be the combination of the sample
stretching and the bending of the rest of the apparatus. Regardless of how the sample
stretches, the bending within the apparatus is constant for a given force. You can
measure this deformation directly by using the calibration bar, which does not stretch
signicantly.
The force measured by the force sensor is applied by a moment arm. The force acting
on the sample being pulled is not equal to the force measured by the sensor, but rather
related according to the physics of how levers work and the distance from the pivot
to the places where the forces act.
6.4 Tests to Perform
You will measure stress-strain curves for three different metals steel, brass, and aluminum.
6.4.1 Steels
Obtain 10 tensile samples from the TA. Divide them into two batches ( 5 samples/batch).
Conduct the tensile tests on the rst batch as described above, ,and plot the stress-strain
curve. From these curves, determine:
1. Youngs Modulus
2. Yield Strength
3. Tensile Strength
4. 0.2% Yield Stress
5. Fracture Stress
6. Toughness
7. Ductility.
6.4. Tests to Perform 53
Report each one of the above properties as a mean value, along with the standard deviation,
and the Coefcient of Variation (100* SD/ mean).
Place the second batch of 5 samples in the tube furnace and follow the heating procedure.
The samples with ramp up to the desired maxium, hold for a preset period of time, and
then down to room temperature. The entire cycle takes plance under a nitrogen purge. See
the class webpage for details on the oven and purge. Test these samples as a batch, and
determine the same 7 parameters that you tested on the rst batch of 5.
6.4.2 Brass and Aluminum
Obtain 5 samples of brass, and 5 samples of aluminum from the TA and conduct the exper-
iments as in the case of the rst batch for steel.
6.5 Fracture Surface Imaging
Once your test samples have fractured, remove them carefully from the test apparatus with-
out touching the fracture surface. Tape them directly into your lab notebook (it doesnt
matter which partners notebook) to save them for later SEM imaging. Give each sample
an ID number or name, and record any relevant notes about how it was processed or tested
that you might want to know later.
We will arrange for all lab groups in each block to go together to the scanning electron
microscope (SEM) in the Geology Division. Selected samples from each group will be
imaged, focusing in particular on the fracture surface. Include the images of your samples
in your lab report.
6.6 Lab Report
You lab report should include a complete description of all measurements made, including
stress strain curves measured for all samples. It should also contain a table of the parameters
you determined from the stress-strain curves, and SEM images of the fracture surfaces.
In your discussion section, compare the results for the steel samples that you heat-treated
with the results for those you did not. Also compare the results for steel, brass, and copper.
Compare a few of the parameters you measured to literature values for these materials
(Youngs modulus, yield strength, and possibly others if you nd them). What factors
might explain any differences between your results and those in the literature?
For the heat-treated steels, include a TTT plot with the time-temperature history of your
samples sketched in. Discuss in terms of this plot and the iron-carbon phase diagram the
transformation in microstructure produced by the heat-treatingn process.
From the images of the fracture surfaces, discuss what you can conclude about the failure
mode. Was the failure brittle or ductile?
54 Chapter 6. Mechanical Properties of Metallic Materials - The Tensile Test
6.7 Pre-lab Questions
1. Give a brief physical description of stress, and strain, including relevant equations.
2. Draw a typical stress strain curve for a metal specimen that underwent tensile loading
until fracture. Label and describe the different stages of deformation.
3. Dene each of the following terms. Also describe how you can determine each of
them from a stress-strain graph: (include diagrams if useful)
(a) Youngs Modulus
(b) .2
(c) Ultimate Tensile Stress
(d) Fracture Stress
6.7. Pre-lab Questions 55
56
CHAPTER
SEVEN
Dynamics of Coupled Mechanical
Oscillators
7.1 Introduction
In many engineering elds, the input signal to a dynamical system is unknown, but the
output signal can be observed and measured. An example of such a system is a seismo-
graph used to measure the strength of an earthquake. Often the characteristics of the input
signal can be determined by constructing an analytical model of the dynamical system and
examining the response of the system to different input signals. In addition, the dynamic
characteristics of various mechanical systems such as an automobile suspension are evalu-
ated by observing the output response to various input test signals.
This experiment is designed to illustrate the dynamic response of a mass-spring-damper
system. The input signal is sinusoidal with variable frequency. The movement of two
masses is measured with linear variable-displacement transformers (LVDT).
7.1.1 Description of the Experiment
Figure 7.1.1 shows a simplied dynamic model of the experiment. The system is composed
of two large masses of mass m (approximately 700 grams) on sliding tracks. The masses
are separated by three linear springs with spring constant k and two air dashpots with vis-
cous friction coefcients b
1
and b
2
. The system is driven by a variable speed motor with
an eccentric shaft to provide a sinusoidally oscillating input. There are three LVDTs to
measure the displacement of the mass from the stationary reference frame: one on each of
the two masses and one on the driver plate. The output from the sensors is sent through a
LabJack to the PC via USB. The LabVIEW VI Coupled Oscillators can be used to monitor
the signals.
Simply navigate to the Data Acquisition tab and click Enable Stream to start data ac-
quisition from the LVDTs. Please refer to the Documentation tab for further instructions
on how to use the calibration and data analysis functions.
57
Figure 7.1: The coupled oscillator experiment.
7.1.2 System Model
By doing a force balance on each of the masses, two coupled linear ordinary differential
equations are determined that we use to model the system. Let X = [x
1
x
1
x
2
x
2
]
T
be
the state vector of this system, where x
1
and x
2
are the respective centers of mass. The
dynamic model has the form

X = AX+ Bu, where:
A =
_
_
0 1 0 0
2k/m b
1
/m k/m 0
0 0 0 1
k/m 0 2k/m b
2
/m
_
_
B =
_
_
0
0
0
k/m
_
_
(7.1)
The transfer function can be derived as: C(sI A)
1
B where A and B are as above, and
C is the output matrix. In this case, if one is interested in measuring the response of both
x
1
and x
2
then C takes the form:
C =
_
1 0 0 0
0 0 1 0
_
(7.2)
Figure 7.2: Simplied dynamic model of experiment. (Note b1 is incorrectly labeled as b
and b2 is not shown.
58 Chapter 7. Dynamics of Coupled Mechanical Oscillators
The characteristic polynomial (e.g., D(s) = det(sI A)) of this system has the form:
D(s) = s
4
+
1
s
3
+
2
s
3
+ 4
2
n
s
2
+
1
2
s
2
+ 2
1
2
n
s + 2
2
2
n
s + 3
4
n
(7.3)
where
1
= b
1
/m,
2
= b
2
/m, and
2
n
= k/m. Then the transfer functions from u(s) to
x
1
(s) and x
2
(s) are of the form:
_
x
1
(s)
u(s)
x
2
(s)
u(s)
_
=
_
2
n
(s
2
+
2
s+2
2
n
)
D(s)
4
n
D(s)
_
. (7.4)
An alternate approach to nding the transfer function is to use Laplace transforms. The
transfer function is dened as the ratio of the Laplace transform of the output function to
the Laplace transform of the input or driving function. Using this approach the transfer
function is the same as given in equation (7.4).
7.2 Experiments
7.2.1 Calibration of LVDT transducers
Calibrate the LVDT transducer by determining the displacement versus transducer voltage
output curve. The calibration can be done by simply moving the mass a known distance.
Check the linearity of the device and be sure to take measurements over the entire range of
motion of the masses.
7.2.2 Response of System to Step Input
Disconnect the two masses from each other. For the single-mass, single-spring system,
record the response of the damped and undamped system to a step input. This can be
obtained by moving the uncoupled masses with your hand and recording the output. This
information will be used in determining k, b
1
, and b
2
.
7.2.3 Frequency Response (Bode Plot) of the Undamped System
Using the results of Sections 7.2.1 and 7.2.2 you should construct plots of the theoretical fre-
quency response. Using these plots as a guide, you should measure the frequency response
of the undamped system. The term frequency response refers to the steady-state response
to a sinusoidal input. The frequency response information is presented in two separate g-
ures, one giving the magnitude of the ratio of the output to input signals as a function of
frequency, and the other presenting the phase angle between the input and output signals
versus frequency. The gures are usually presented on a log-log scale for the magnitude and
a log-linear scale for the phase. This representation of the frequency response is referred to
as a Bode plot.
Disconnect and remove the damper from both masses. To nd the frequency response,
you want to examine several different input frequencies and record the displacement of the
input signal and the displacement of mass 1 and 2. Start with a low frequency and cover
a frequency range that includes the rst two resonant frequencies. Be sure to take enough
data to characterize each of the resonant frequencies.
7.2. Experiments 59
7.2.4 Frequency Response of the Damped System
Reconnect the dashpot to both masses. Set the damper to a relatively low damping coef-
cient and do not readjust the knob during the course of your experiments. As done in
Section 7.2.3, record the displacement of the input signal and the response of mass 1 and 2.
(Note: be sure to take the response of the damped system with the same damping settings
as you had for the step response. The damper has a small knob which controls the amount
of damping. If this knob has been moved, you will have to take another step response at
this time to compute the correct damping factor.)
7.3 Analysis of Experimental Model
To match the experimental data to the analytical model, we make a number of simplifying
assumptions:
1. Assume that the friction of the LVDT transducer is very small and neglect stiction
in the sliding bearings. Thus the friction will be dominated by the air dash pots.
These dashpots can nominally be modeled as viscous friction devices, with viscous
damping coefcients b
1
and b
2
.
2. Neglect the mass of the structure connected to the eccentric input shaft.
3. Assume that both larges masses have uniform mass, m.
4. Assume that all 3 springs are linear and have uniform spring constant k.
7.4 Lab Report
1. Sketch and briey discuss the experimental setup.
2. Record the calibration data, and estimate the uncertainty in the amplitude measure-
ment. Can the transducer be modeled as linear?
3. Measure the step response and the undamped and damped frequency response, as
described in the text. Record all data.
4. From your measurements, estimate as accurately as you can the model parameters
(spring constants, damping coefcients). There are two ways you might do this. You
could determine them from the vibration period and decay rate of the step-response
time series data. Alternatively, you could work in the frequency domain, by adjusting
the parameters in the theoretical amplitude and/or phase frequency response func-
tions (using the program you wrote) to achieve the best t to the measured frequency
response data. However you do it, make sure your conclusions are self-consistent -
i.e., that both the measured step response and frequency response are consistent with
the same model parameters.
5. Plot the theoretical phase response vs. frequency for both masses. At what frequency
or frequencies do the two masses move in phase? Out of phase? Verify these predic-
tions experimentally.
6. Note that there is a frequency between the two resonances where mass 2 is predicted
to remain motionless, even though mass 1 and the input are both moving. Verify this
experimentally, and check that the frequency where this occurs is where it is predicted
to occur. (This frequency location might be useful in part 4 too)
7. Discuss any problems, discrepancies with theory, possible causes, etc.
8. Time permitting, explore other topics of your own choosing. Can you replace a spring
by a nonlinear one, for example? What new phenomena, if any, result from nonlin-
earity?
7.5 Advanced Experiment (Optional)
Find or make a nonlinear spring(s). Insert this spring(s) in place of an existing spring,
and try to observe the non-linear jump resonance phenomenon which occurs as you slowly
increase the frequency of excitation. Alternatively, the springs currently installed in the
system are slightly nonlinear. With the proper initial conditions and frequency of excitation,
they will exhibit a steady-state nonlinear response. Have the TA show you the conditions
which lead to this phenomena.
7.6 Prelab
1. Derive Equation (7.1).
2. Explain what a transfer function is, and derive Equation (7.4).
3. Write a program or spreadsheet to make Bode plots of the amplitude and phase fre-
quency response from Equation (7.4).
4. Using the program, explore how damping affects the frequency response functions.
Print out and put into your notebook an amplitude Bode plot showing the frequency
response for a few damping coefcients ranging from very small to large.
7.7 Appendix: Making Accurate Period and Phase Measure-
ments
1. You should measure period and phase information by using zero crossing times in-
stead of peaks in data. The data near the peak is very noisy while zero crossing data
is much cleaner (due to slope crossing zero).
2. To measure zero crossing times you must calculate the mean of the data from the min
and max of the waveform and not the mean reported by the data capture program.
7.5. Advanced Experiment (Optional) 61
The reason for this is that the data capture program calculates the mean for all of the
captured samples which usually dont correspond to an integral number of cycles.
3. Getting error information for stiffness and damping is easy using different cycles of
the data and also using multiple runs.
APPENDIX
A
Keeping a Lab Notebook
Learning how to keep a lab notebook is critical to success as an experimental scientist or
engineer. The lab notebook is the primary record of what was done in an experiment, and
should be written in a manner so that someone reading it years later (perhaps you) could
understand exactly what was done, and if necessary replicate the experiments.
Lab notebooks are legal documents, and are often called upon in legal disputes over patents
or scientic priority. Dont let your patent rights for your brilliant idea, or your Nobel prize
for your brilliant discovery, slip away due to sloppy lab notebook practices!
There is no single way to keep a lab notebook, but there are some generally-agreed-upon
points. These are summarized below.
1. Use a lab notebook with pages bound securely, not a spiral-bound notebook or one
with easily-removed pages. The pages should be numbered; if they are not, then
number them yourself.
2. Never remove a page from the notebook.
3. Write in ink, preferably blue or black. Never make any entry in pencil.
4. Do not make notes or write down measured values on scrap paper, to later be entered
into the lab notebook. Write your observations and measurements directly into the
lab notebook.
5. Write legibly. It is not necessary to have a beautiful notebook, but it should at least
be possible to read what you write.
6. Never erase anything, or use white-out, or cover anything with scribbled lines. If
something is incorrect, simply draw a single, horizontal line through it, so that the
text can still be read. You may decide later that the values were correct after all, so it
is important that they be still readable. If a large section of material is in error, simply
draw an X covering it.
7. Whenever you begin making entries into the notebook, rst date the entry. It should
be possible years later to determine the exact date on which each entry was recorded.
When you start a new page, put the date at the top.
8. Never leave a blank space to be lled in later. All entries must be in chronological
order.
63
9. When you need to include an image or a graph, sometimes a sketch or quick plot by
hand is sufcient. For computer-generated images or plots, print it out and fasten it
into the notebook. Glue is best for this.
10. Write down how the measurements are made as you make them; if you wait until
later, you will forget the details. Was the length of a beam measured from its end, or
from its support post? Was the voltmeter zeroed before making the reading?
11. In many experiments, it is not practical to include all data generated by the exper-
iment in the notebook. This is particularly true if your primary measurements are
made using a computer. In this case, there is usually little point in fastening into the
notebook many pages of printouts listing numbers. Instead, save the data in a le on
the computer, and write in the notebook the name of the le(s) containing the data.
Summarize the salient features of the data in the notebook, possibly with a sketch or
a graph. What is worth noting about the data? Why did you acquire it? What does it
show? The interested reader (perhaps you years later, or the patent examiner, or the
Nobel committee) should be able to rst read the notebook to see the record of what
you did and were thinking at the time, with the raw data location clearly identied so
that he/she can access all of it (and possibly re-analyze it) if necessary.
12. Make sure your data les are secure; you may want to transfer them to a memory
stick at the end of each lab session. When you are nished with the experiment (or
with the course), burn a CD containing all raw data les, and attach it to your lab
notebook.
13. If you are acquiring data by computer, save all data les. Just like erasing entries in
a lab notebook, if you delete data les it is all too tempting to delete the ones that
dont show the phenomena you expect to see. If there is a good reason to think that
some data les are in error (for example, due to equipment that was not functioning
properly), then note this in the notebook, but retain the data les anyway.
14. Never falsify data or dates in a lab notebook. Either one is serious scientic miscon-
duct. Intentionally entering an incorrect date may affect claims of scientic priority
or patent claims. And recording measurments that were never made can lead (and
has led many times in the past) to fraudulent claims of new physical phenomena.
Inevitably the fraud is discovered when no one else can repeat the purported exper-
iments, and the result is always the sudden end of a scientic career in disgrace. In
this course, false entries in a lab notebook will be taken as seriously as any other form
of academic Honor Code violation.
64 Appendix A. Keeping a Lab Notebook
APPENDIX
B
LabVIEW
A new feature of ME 96 is the emphasis on teaching modern data acquisition methods, and
in particular, data acquisition using LabVIEW. LabVIEW is a product of National Instru-
ments, Inc. and is the leading software package by far for measurement and test systems.
Like many other tools we use routinely (Mathematica, MATLAB, Excel, . . . ), LabVIEW is
commercial software, and is not free. But an inexpensive student editition is available that
costs no more than a textbook ( $100). See the wiki for more information.
Most of the experiments use LabVIEW to acquire data and/or to control the equipment. In
several of the experiments, you are asked to write your own VI, or modify an existing one
to add new functionality. These experiments are:
1. Boiling Heat Transfer. Write a VI to acquire thermocouple data and plot temperature
vs. time and the boiling curve (heat ux vs. T).
2. Convection. Write a VI to monitor thermocouples and notify you when steady-state
conditions are achieved.
3. Cantilever Beam. Modify a supplied VI to plot the frequency spectrum of a time-
varying signal.
4. Flame Fuel Cell. Write a VI to display the temperature read by the thermocouple
probe and the voltage across the fuel cell.
5. Turbulent Air Jet. (To be determined.)
In addition, the Coupled Oscillators and Turbomachine experiments use LabVIEW to ac-
quire data. Currently, these run using supplied VIs that do not require modication.
65
66
BIBLIOGRAPHY
67

Lab Manual 2013

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Lab Manual 2013

Hochgeladen von

Copyright:

Verfügbare Formate

ME/CE 96: Mechanical Engineering

to this oncoming ow, which is subsequently

is the component of the absolute

agree. If the difference is more

is the approach velocity, and is the kinematic viscosity. For free

(x, r, t) = u(x, r, t) u(x, r). (5.2)

is the RMS value of the uctuating velocity component, dened as

. Does the uctuating velocity

Das könnte Ihnen auch gefallen