ROAD VIBRATION SIMULATION EXPERIMENT: FINAL REPORT Report3.pdf · road vibration simulation...
Transcript of ROAD VIBRATION SIMULATION EXPERIMENT: FINAL REPORT Report3.pdf · road vibration simulation...
ROAD VIBRATION SIMULATION
EXPERIMENT: FINAL REPORT
DESIGN AND DEVELOPMENT OF A ROAD VIBRATION SIMULATOR FOR THE MIAMI UNIVERSITY VIBRATIONS LABORATORY
CAROLYN BASSETTI SARAH BRADY
MANDY FREDERICK JESSICA IRONS MATT WINDON
DR. SHUKLA
DECEMBER 15, 2003
M I A M I U N I V E R S I T Y S C H O O L O F E N G I N E E R I N G A N D A P P L I E D S C I E N C E
D E P A R T M E N T O F M A N U F A C T U R I N G A N D M E C H A N I C A L E N G I N E E R I N G
2
Road Vibrations Simulation EGR 448 SENIOR DESIGN FINAL REPORT
EXECUTIVE SUMMARY
The Road Vibration Simulation Team’s task is to design and develop a road vibration
simulator for the Vibration Laboratory at Miami University. This will test the vibration
response to various potential road conditions of a component of an automobile or an entire
automobile scaled down to 1/5 the size. The simulator must be able to generate input and
then measure the response of the specimen being analyzed. The vibration simulator will be
used primarily for educational purposes and three lab experiments will be developed to
coincide with the simulator. The final product will be posted to the internet so other smaller
schools can mimic the design created by the team here at Miami University. The design of
the vibration simulator includes many components. The simulator will be suitable for use
with a scale model of an entire car as well as individual car components. The sensors used
will be force and accelerometers. The actuator will be an electrodynamic shaker and the data
acquisition/analysis package used will be Matlab. The design team is closely following a
Gantt Chart written at the beginning of the project to stay on task. This vibration simulator
will be finished by mid-April of 2004. It will then be tested by EGR 315 students. Based on
their feedback and time remaining, changes will be made accordingly to improve the
experiments written for the simulator.
3
CONTENTS
1. Executive Summary 2
2. Introduction and Background 4
3. Research 8
4. Proposals 34
5. Conclusions and Future Work 50
6. References 53
7. About the Team 54
8. Appendix 55+
4
INTRODUCTION AND BACKGROUND
INTRODUCTION
Road vibrations simulators are very important in the automotive industry. Before
assembling any vehicles, the parts must be analyzed in order to determine their reaction to
road vibrations. Once constructed, the entire vehicle is often analyzed to determine
response to vibrations as well. There are many safety reasons and comfort factors to be
considered in road vibrations analysis. While there are theoretical ways to roughly estimate
the response these components will have to road vibrations, it is necessary to experimentally
measure the actual responses to get accurate information. This simulator should be able to
measure the responses as accurately as possible, and it should give students an intuitive,
hands-on experiment in which to learn about these vibrations.
PROBLEM DEFINITION
Problem definition is a critical step in engineering design. The problem statement
should include, “objectives and goals, the current state of affairs and the desired state, any
constraints placed on solution of the problem, and the definition of any special technical
terms [Dieter, 10].” The vibrations design team identified an overall objective, several
important constraints and some general tasks within their problem definition.
Objective
To design and implement a versatile device and corresponding classroom experiments
for use in the Miami University Vibration Laboratory to simulate road vibrations felt by
vehicles during normal operation.
5
Constraints
The constraints of the project include physical laboratory space, time, and funds. Kreger
is currently the home for the School of Engineering and Applied Science. Room 201 is used
as not only a classroom, but also as a laboratory for many engineering classes. Because of
this, the design team feels lack of space is a constraint. In initial meetings it was determined
that the space available for building, storing and running the experimental setup could only
be the size of a tabletop. The next constraint identified is limited time. EGR 448 and 449 is
a two semester series in which research and design should be completed within the first
semester, leaving the second for building and implementing the projects. With only a school
year, about 8 months, the design team must consider time throughout many aspects of the
project. Finally, taking on such a large project requires large funds. The design team has
identified funding as an important constraint limiting many aspects of design including the
components used, specifically the types of actuators and data analysis software.
Tasks
The four main tasks of this project include (1) design and build the vibration
experimental setup, (2) develop Matlab interface to acquire and analyze the data from the
vibration simulator, (3) develop a lab procedure and manual for students and instructors, and
(4) make all relevant documentation available to all of academia via Internet.
As mentioned, one aspiration of the road vibration simulator senior design is to enhance
the learning experience and future engineering curricula. Education cannot be limited to
learning solely from books, and because of this, the project team’s primary intentions for the
final device is that it be used by future engineering students and faculty at Miami University
to supplement the traditional vibrations courses. Following this idea, all four of the general
tasks identified in the problem statement have an end goal of enhancing education. The
6
road vibration simulator must have an experimental focus. Therefore, the first task classifies
the device as an experimental setup and designing and building it is of primary concern for
the design team. Engineering students use Matlab or other data acquisition devices to collect
and analyze data during experimental procedures, thus, it is important for the design team to
incorporate Matlab into the experimental setup. This allows students to use the simulator
and understand it with ease. The third task listed is also of vital importance when it comes
to the education process. The design team must develop a lab procedure for students to use
in supplementation of vibrations courses. In addition, the team would like to construct a
manual for students and instructors to refer to when using the road vibration simulator.
Finally, with a broader reach in mind, the project team would like to make all design
information, data, and documentation available to academia via Internet. The project team
hopes to not only enhance Miami University, but also allow other universities with limited
resources to improve the learning experience of their engineering students by taking ideas
from the team’s research and design to either duplicate or add to the experimental setup.
TEAM OBJECTIVE
With a problem definition focused on building equipment and designing experiments to
enhance the educational experience at Miami, the design team has created a team objective
to further clarify what this means to them. Education is fundamental to the conception of
the project in the team’s mind. They feel it is important to consider the enhancement of
education in every step of the design and later the implementation process. Because of this
an underlying, shared goal has been formed: To build intuitive, hands on experiments that
supplement vibrations courses to enhance the educational experience of the engineering
student. In order to achieve this goal, the team must constantly consider the future students
7
when making decisions about design. This will be referred to throughout the research and
proposal sections. A simple example comes when choosing the layout for the graphical user
interface (GUI). The team had to take the seat of the students when choosing this design.
They had to consider, as students without any previous experience in vibrations or use of the
software, what would make the interface easy to use and to understand. A more specific
discussion of what went into this process can be found in the research and proposal section
of GUI.
The team also shares a common drive for choosing this project and working hard at
providing intuitive educational aids. Throughout their educational experiences, the team
feels that they have experienced many different effective as well as ineffective teaching
techniques. It is well agreed upon that classes involving hands on experiences are the most
effective in terms of material learned. The team feels that the reason for this is that these
types of experiences force a student to apply the knowledge he or she has learned from a
book, which in turn reinforces what has been taught allowing for better absorption of the
material.
A final objective of the team is related more to functionality of equipment and
experiment. The team has made sure to keep in mind the goal of building a versatile device.
Because of this, they have taken steps to insure all pieces are mobile and independent. First,
this will be useful for when the new engineering building is built and the equipment will need
to be moved. And secondly, this is important so the equipment researched and purchased
can easily be used for future projects.
8
THE BIG PLAN
The design team developed the diagram below to show how they view the project as a
whole. The team feels there are three main interrelating tasks within their project: Hardware,
Software, and Classroom Experiments. As the tasks are interrelating, the team feels that
singling one out to start work on would be an inappropriate way to approach the project.
The team feels that moving forward in each task simultaneously is the best approach. This
allows for gaining feedback and dealing with troubleshooting early on. They feel that leaving
a category untouched until it is necessary to begin may cause problems in the other two
categories if a problem is encountered in the third. The design team keeps this diagram in
mind in determining goals as well as in determining how one category affects the rest of the
project. The Big Plan can be seen in Figure 1.
Figure 1
RESEARCH
IMPORTANCE OF VIBRATION IN DESIGN
Vibration is a very important issue to study and analyze when it comes to design because
it has many effects on the overall system and how the system performs. Safety, human
9
factors and ergonomics, and style are three primary issues that account for vibrations
considerations in their designs.
Safety
One of the primary reasons for analyzing a system’s vibrations properties is safety. Since
structural integrity and coupling are affected by changes in vibrations properties, if a system
is not designed to withstand the vibrations it is going to encounter, the structure could fail
and human lives could be in danger. A perfect example of when vibration was not properly
considered and analyzed in design is the Tacoma Narrows Bridge Disaster. According to the
Tacoma Narrows Bridge site, on November 7, 1940, at approximately 11:00 AM, the first
Tacoma Narrows suspension bridge collapsed due to wind-induced vibrations
[http://www.civeng.carleton.ca/Exhibits/Tacoma_Narrows/]. Situated on the Tacoma Narrows
in Puget Sound, near the city of Tacoma, Washington, the bridge had only been open for
traffic a few months.” The bridge had been designed for structural grace with slender
towers and shallow stiffening trusses. The bridge was so light that it behaved much like an
airplane wing, generating lift with significant enough winds. On the day that the bridge
failed, the wind speed was such that the bridge began to oscillate at its resonance frequency,
thus causing the bridge components to fail.
Until the Tacoma Narrows Bridge disaster, designing with resonance frequencies in
mind was virtually non-existent, and thus the study of design against resonance frequencies
is a fairly new field. Now, experiments in the field include placing tiny accelerometers all
over a structure to monitor how it is behaving and to be able to detect changes in the system
behavior. This way, if there is a significant change in the behavior of the structure, the
proper measures can be taken to ensure the safety of those using the structure.
10
Other safety considerations have to do with the modal properties of structures. Certain
physical quantities are characteristic of every structure; these properties are known as Modal
properties. Modal properties (resonance frequencies, mode shapes, etc.) of the structure are
very closely related to how vibrations affect the system and are functions of their physical
quantities; consequently, when damage in the system results in a change in the physical
properties of the structure, the outcome is a corresponding change in modal properties of
the structure. If these changes are too great, the integrity, and therefore the safety, of the
structure will be compromised. For this reason, it is very important to consider the modal
properties of all the materials to be used in a structure.
Human Factors/Ergonomics
Another important factor when it comes to vibrations in design is the consideration of
human factors and ergonomics. According to the Merriam Webster Online Dictionary
ergonomics is “an applied science concerned with designing and arranging things people use
so that the people and things interact most efficiently and safely -- called also human
engineering.” Therefore, vibrations is important to consider when it comes to ergonomic
designs because if the vibration of a structure that a human is using is causing the person to
become uncomfortable in someway, this may cause them to become less focused on using
the structure properly and could lead to safety hazards. A perfect example of the importance
of this issue is when it comes to driving a car. Vibrations in the structure of the car can
cause fatigue and decrease the safety of the driver. According to a study done by Australian
Transport Safety Bureau (ASTB), “Some studies identified vibration as a stimulant and
others as inducing drowsiness…random and stimulating vibration exposure can act to
increase alertness and that low frequency monotonous vibration has the opposite effect by
increasing drowsiness.” If there are many high frequency vibrations, this could keep the
11
driver awake as “vibration and noise exert a stimulating effect on the central nervous system,
however result in muscle fatigue in the back and arm muscles due to the strain of controlling
the car. Body chemistry changes due to the vibrations experienced when driving. According
to the ASTB study, “the increase in lactic acid is due to muscle activity in the spine directed
at maintaining posture under exposure to vibration.” This muscle fatigue could be due to
changes in the body’s metabolism. Despite the reason for increased fatigue due to
vibrations, this could be avoided if the car had been designed with vibrations and how they
affect ergonomics in mind. On the other hand, if there are too many low frequency
vibrations, more similar to a rocking motion, this can increase the driver fatigue and
drowsiness and also decrease the safety of the driver. Again this could have been avoided
had vibrations been considered in design. In general it was found that drivers became more
fatigued from a large amount of low frequency vibrations rather than a large amount of high
frequency vibrations, however, regardless of the type of vibrations induced on the driver by
the car, high and low frequency vibrations are both dangerous and therefore need to be
considered in the design of cars and any other structure where the user can be negatively
effected by the vibrations the structure causes just due to normal use.
Style
Another aspect when it comes to the importance of vibrations in design is the style of
the structure. It is possible to tune the engine such that it vibrates at a certain frequency to
give a desired sound. This is usually done for cosmetic reasons, such as the Ford Mustang,
which has what one would call a signature sound. Harley Davison has actually patented its
sound and now each new Harley Davison engine must be tuned to that desired, patented
frequency or it is not of acceptable quality for the Harley Davison Company.
12
REASEARCH INTO CURRENT VIBRATION SIMULATION EXPERIMENTS
In order to justify the cost and time expenses of designing a vibrations experiment from
scratch, the team obtained some catalogues and brochures from Dr. Ettouney for various
educational experiments. These companies included Armfield and Hampden, which are
leading companies in the field of educational experiments. Researching this information
showed that the companies do not currently make any type of vibrations simulation
educational experimental setups. The types of experiment packages available from Armfield
and Hampden companies are in subject areas such as fluid mechanics, heat transfer,
thermodynamics, unit operations, controls, and the like. As well as looking at the brochures
from Hampden and Armfield, the team contacted the companies to see if they knew of
other businesses that make vibrations experiments, as well as doing extensive research online
about whether such experiments are available. From this research, the team discovered that
there are currently no educational vibrations experiments available. Because of this, the team
needed to create an original vibration experiment to be used for education. Looking at the
booklets also gave the team an idea of what was needed in a laboratory set up description
that would be available to other universities interested in modeling the experimental set up.
For example, the Hampden brochures outline how well written experiments have a purpose,
description, and specifications in addition to the actual hardware and software associated
with the experiment.
TRIP TO UNIVERSITY OF CINCINATI VIBRATION LAB
One of the first steps in the team’s research was to visit the University of Cincinnati’s
vibrations laboratory in order to get an idea about what a vibrations simulator looks like and
how such a system works. The team was unable to go at a time where they could see the
13
simulator in operation; however, the team was able to get an idea about what types of
components go into such a the design of such a structure.
A picture of the simulator used at the University is shown in Figure 2.
Figure 2
http://www.eng.uc.edu/dept_min/research/
This simulator is very large scale. An entire car can be driven onto the simulator for
analysis and a separate controls room is used to monitor the forces going into the car, as well
as the vibrations effects which result on the system. Hydraulics are used as the force of
power for the simulator and each wheel is placed on one of the four hydraulically operated
shakers. A specially designed floor built on top of springs is used to isolate the experiment
from vibrating the rest of the building during the experimental testing.
The simulator is part of the SDRL (Structural Dynamics Research Laboratory) at
University of Cincinnati, which has a large list of corporate sponsors. The Boeing Company,
Ford Motor Company, General Motors Corporation, Hewlett Packard, Leuven
14
Measurements and Systems (LMS), The MathWorks, The Modal Shop, MTS. Noise and
Vibrations Division, PCB Piezotronics, Pemtech, Structural Dynamics Research Corporation
(SDRC) and Vold Solutions all cooperate with UC’s SDRL laboratory in some form.
DISCUSSION OF PRIMARY DESIGN COMPONENTS
Building a vibrations simulator like the one used at the University of Cincinnati would
require a much more space, funding, and time than the team has available for their project;
consequently, the goal is to be able to scale down UC’s design and create a table top
simulator that can be used to conduct an experiment with the same principles in mind.
Although the simulators will have obvious differences, the four major components needed
in the design remains the same: the physical structure, sensors, actuators, and data
acquisition and analysis software. The following section discusses some of the team’s
research and reflection on each of these components:
Structure
The first consideration the team debated was what physical structure to test using their
simulator. An entire car, an entire scaled model of a car, or individual car components were
all options for what object should be tested. Testing an entire car was eliminated
immediately on account of a lack of space and funding available. The next consideration
was the level of intuitiveness involved in the design. Since the simulator was designed with
the goal of being used for a beginner’s vibrations experiment, the team wanted to make the
design very intuitive in order for students to be able to readily see the real world applications
involved. The team believed that using an entire scaled down model of a car supported this
goal. Also, if individual car components were to be the subject of testing, the experiment
15
would want to eventually relate the effects on the individual components back to see the
entire ride of the car later in the experiment. A system of relating the two would have to be
developed, and consequently, by testing an entire scaled down model, this step can be
eliminated from the process.
As research showed, the world of scale model vehicles is vast including anything from
monster trucks to motor cycles at scales ranging from 1/36th to 1/4th. Dr. Fazeel Khan
spoke with the team specifically about 1/10th model cars and the realistic components found
on these cars. He brought in a car and the team was able to see the car’s size as well as its
many components. He also spoke with the team about what he felt would be a good model
size and type for the project at hand. He recommended a scale remote control car larger
than 1/10th. The team was able to agree with this recommendation prior to research due to
the shared desire for intuitive learning. The 1/10th scale was too small, allowing for little
variation in the placement of accelerometers or the actuator. Also, the addition of
accelerometers to a car weighing less than 5 pounds would add a substantial percentage of
mass to the system. The size also affected the acquisition of accurate data. It was found that
using a small scale car would produce frequency data consistently at high ranges, while a full
scale car would produce frequency data at low ranges. The problem here exists in that
analyzing data at either high or low ranges is not only difficult to do, but difficult to do
accurately. A car between 1/10th and full scale would produce data over a middle range
giving more consistent data that is easier to analyze. Again with the desire to make this
experiment applicable to the real world, the team chose a car as the remote vehicle as most
students drive cars in comparison with motorcycles or trucks.
16
Since a size between 1/10th and full scale was needed, the team started researching the
1/5th scales and 1/8th scales. The design team contacted a company, Technokit, who made
1/5th and 1/8th cars to determine whether one scale was more realistic than the other. Their
response suggested that their 1/5th scale cars were more similar to a real car in terms of
weight, engine, tires and response than their 1/8th cars. Upon discussing these results with
Dr. Shukla and Dr. Khan, the design team decided to focus their future research on
obtaining information and dealers of 1/5th scale cars.
Support Structure
An important part of the required hardware is a support structure for the car. There are
two main reasons for the structure. First, the car weighs approximately 23 lbs, and the shaker
can only support a payload of up to 10 lbs. Therefore, the structure will need to support the
weight of the car. Also, the support structure needs to prevent the car from being in contact
with the ground. Since the team does not have an isolated laboratory, outside vibrations may
be transmitted from the ground to the car. Also, the ground would dampen vibrations from
the shaker to the car. In order to fulfill these requirements, the team considered two possible
support structures. The first idea was to have the car suspended from the ceiling by 4 bungee
cords. This would be a simple and cost effective way to accomplish both requirements. The
other option was to build a separate support structure, and suspend the car from bungee
cords attached to the structure.
Sensors
Two types of sensors will need to be used in this experimental setup. These are force
sensors and accelerometers. Force sensors will be used to measure the input force applied to
17
the system by the actuator. The team will need to select their actuator before they can select
a force sensor that can be used in combination to accurately measure the impact. Miami
University already has three accelerometers available to be used in the experiment. The team
hopes to use these readily available accelerometers in order to cut back on some of the costs
involved with the project.
Load Cell
There are many options available to measure the force transmitted from the shaker to
the car. A good but potentially costly choice would be to use a load cell to measure the force
output from the shaker. One option would be to run an open loop experiment that would
not require a load cell. However, this would limit the data recorded, and reduce the accuracy
of the data. There are still other alternatives to using the load cell. If the team wanted to
achieve similar results at a cheaper price, they could use an accelerometer to measure the
input.
The team is using a load cell to incorporate the idea of force measurement (hence not
use an accelerometer) in the experiment. Force measurement is important in the experiment
because the team is attempting to create an intuitively obvious experiment. An important
aspect of this is to see the forces applied to the car by the road conditions. Theoretically, a
shaker would send the same signal as you generate in Matlab but using a load cell is more
precise. It also helps in detecting any loose connections between shaker and the car. The
voltage is input through Matlab, then amplified through the amplifier, so there will be many
chances for the data to differ from the predicted values. The load cell will record the actual
data, thus increasing the accuracy of the results. The team has not chosen a specific load cell
to purchase yet.
18
Actuators
The three primary types of actuators the team debated using in the design of the simulator
were hydraulic, electronic, and impact systems. The design team primarily investigated these
three types since they are the most commonly used techniques, and information on them is
readily available and accurate.
Hydraulic
Hydraulic actuators operate using Pascal’s Law, which states, “when a pressure is applied
to a liquid in a sealed container, the pressure is distributed across the liquid, and the pressure
is the same everywhere throughout the liquid.” In hydraulic systems, a power source must
supply pressure to the fluid. The pressure is the same throughout the fluid, which is
consequently used to push and pull a piston.
Hydraulic systems are lightweight and compact. Low compressibility of the medium and
low sensitivity to temperature are also advantages. Hydraulic actuators are best for heavy
loads and high speeds allowing for large forces and accuracy. The use of fluids allows for
inherent damping which results in no added vibration. The lightweight components in a
hydraulic system allow for quick acceleration and low heat generation. The vibrations
simulator used at the University of Cincinnati, as well as the Model 3.53.10 Multi-Axial
Simulation Table, both have impact forces powered by hydraulic systems.
Some disadvantages exist as well. These systems are typically complex and require
maintenance to numerous filters, fitting, valves and tubes. Maintenance must be performed
by someone with expertise in the hydraulic field and frequent testing of the hydraulic oil
must be done. Noise as well as leaking of oil are concerns, especially since the oil can be a
fire hazard. Costs can add up quickly as a separate power source is needed as well as
component replacements and energy to generate power.
19
Electric
Electrodynamic and electromagnetic are two different options for electric actuators.
These electric actuators use an electrical power supply as a drive. The electrodynamic uses
an electric motor, while the electromagnetic uses magnets and varying voltage to drive the
actuator.
Both offer low noise advantages as well as low power supplies which mean lower energy
costs. Electric actuators are excellent with light loads and low speeds. Electrodynamic
actuators are lightweight and do not use pipes or pumps. Without the use of flammable or
compressed fluids, the electric systems are environmentally safe. Neither unique nor
extensive maintenance is needed in electric systems. Electromagnetic systems have a quick
response time and can generate high forces.
The electromagnetic actuator is heavy and this can cause slow acceleration. Elements,
such as screws or gears, within the electrodynamic actuator may transmit vibrations to the
moving parts. As the parts of the system wear, the system may lose accuracy.
Impact/Manual
Impact testing is the third option the team has considered as the actuator for the applied
vibration force. In impulse testing, the input force to the system is a single impulsive force
applied by a specially designed impact device, such as an impact hammer.
There are several benefits to impact testing. It can be said that “the usefulness of the
impulse technique lies in the fact that the energy in an impulse is distributed continuously in
the frequency domain rather than occurring at discrete spectral lines as in the case of period
signals. Thus, an impulse force will excite all resonance frequencies within its useful
frequency range.” (Halvorsen) This basically means that impact testing can capture all of the
20
same results as the other two methods for system excitation, obtaining relatively accurate
representations of the system’s vibrations properties. Impact testing is also simple to setup
and take measurements for, as well as highly portable. Finally, one of the largest benefits to
impact testing is the relatively inexpensive
One of the largest downsides to impact testing is that it is not a very intuitive method of
testing, one of the primary goals for the team’s experimental design setup. It is doubtful that
vibrations beginners would believe that the force applied by an impact hammer can
represent a road vibration as accurately as an electrical motor. Also, impact testing is very
subject to human experimental error. In order to obtain the most accurate results, it is
desirable to average multiple hits on the system. In doing so, testers need to make sure that
they are striking the exact same place each time they repeat the experiment. They also must
strike the object being tested directly along a single axis of impact in order to obtain accurate
results. Finally, the tester must be sure to only strike the system twice, for if vibrations cause
a second impact to take place, the results on the system will not be accurate.
Electrodynamic Shaker Options
The group considered purchasing electromagnetic shakers from 3 different companies:
DataPhysics, Ling Dynamics, and MB Dynamics. After doing preliminary research, the
group discovered that MB Dynamics was the leading shaker supplier in the industry. Still, the
group wanted to consider all possible options before making a decision. Datasheets for the
shakers from these companies can be found in the Appendix.
Ling Dynamics, which distributes their products through the LDS (Ling Dynamic
Systems) Group, seemed to have exactly what was needed for the project. The shaker
considered by the group was the V406/8 – PA 500L. This was their 44lb sine force shaker.
21
Ling specified that their products could be used for laboratory experiments. They also
offered cooling fans and amplifiers in a package with their shakers. However, upon
contacting the LDS group, the team found that no academic discounts were available, and
the sales representative also made reference to MB dynamics as a reputable company that
may be able to suit our needs.
DataPhysics offered a 45 lb sine force shaker that produced a 31 lb random force, the
DP-V031. They also offered packages that included amplifiers and cooling systems.
However, DataPhysics was hard to contact, and did not seem to be willing to go out of their
way to fulfill the team’s needs.
MB Dynamics was very highly regarded by every shaker authority the team contacted,
and Dr. Shukla also felt that MB Dynamics was a reputable company that would be easy to
deal with. MB Dynamics offered a Modal 50A exciter that produced a 50 lb sine force. They
also offered amplifiers, cooling systems, amplifiers, and accessories kits. MB Dynamics also
responded quickly to any requests the team made, and offered an academic discount for any
items the team would purchase. MB had a great reputation for service, which the group
found to be very important due to the unique needs of our project.
Amplifier
The shaker requires an input voltage of up to approximately 20 volts. However, it is
dangerous to produce a voltage of much more than 5 volts through a software program,
such as Matlab. Therefore, an external amplifier is required to produce the necessary voltage.
Amplifier selection is a relatively simple process, since many shakers have recommended
amplifiers. The team also determined that they would need an accessories kit, or a similar
product, that would supply all of the necessary cables and attachments.
22
Cooling System
Cooling systems usually are necessary when the shaker is used at more than 50% of the
full performance force. To determine the necessity of a cooling system, the team asked many
representatives of the shaker suppliers their opinion, based on the projected uses for the
shaker. They also considered the requirements the experiments would have for the shaker,
and the possibility of future experiments that may have a greater force requirement from the
shaker and would therefore need a cooling system. The team decided that they will not be
using the shaker near its peak force, and the shaker experiments will be no longer than a few
seconds each. The alternatives they considered were as follows:
• No cooling system.
• Shop air cooling system. This would connect directly to the shop air supply. The
current vibrations laboratory has access to shop air, so this would be a feasible
option for us.
• Portable cooling system. This uses a vacuum cleaner type idea that can be used when
shop air is not easily accessible. This is a little more expensive than the shop air
cooling system, but because of its portability and air flow generation, it is a very good
option.
DATA COLLECTION & ANALYSIS
Matlab
Matlab is a software program designed to aid in the mathematical analysis of system
models and experimental analysis. Some of the capabilities of Matlab include technical
computing, control design, test and measurement, and image processing. In this experiment,
the goal will be to develop the capability to integrate Matlab into the experiment in order to
23
collect and analyze the experimental data. Matlab will also likely be used to obtain transfer
functions (graphs of the system output/ input with respect to time) and other relevant
vibrations graphs for the laboratory experiment.
LabVIEW
An alternative to MatLab is LabVIEW. LabVIEW is another software program that can
provide a graphical way for the group to acquire and measure the signal, analyze the
measurements, and present the data in an easy-to-use fashion. It is advertised to be easier to
use than other traditional development tools.
Selection
Primarily due to the fact that three of the students have had at least some previous
experience with Matlab, and two of the students have used Matlab in a vibrations-based
experiment for data acquisition as well as analysis, the team decided to use Matlab for the
software portion of the project. Information manuals and faculty expertise towards Matlab
is also more readily available, which would be helpful in troubleshooting problems that could
occur in learning this complicated and relatively new software program. Finally, research did
not concluded that LabVIEW’s data acquisition and analysis capabilities were superior to
Matlab’s in a way that would counteract the benefits of using Matlab.
DAQ
In order to measure and analyze physical phenomena, a data acquisition system is
necessary. This system typically provides the tools and resources necessary to perform the
24
aforementioned task. A data acquisition system is basically a connection between software
and hardware. Typically, a data acquisition system consists of:
• Data acquisition hardware
• Sensors and actuators
• Signal conditioning hardware
• Computer
• Software
The computer needs to read and analyze the data gathered by the accelerometers on the
car’s vibrations. The accelerometers read data in analog form. This data is then sent to a
signal conditioning box which also reads in analog. The signal conditioning box improves
accuracy with amplification, filtering, and allows for simultaneous sampling. From this signal
conditioning box, the data is sent to the computer. In the computer, there is a card that
converts analog data to digital data so it can be processed. Now, the data acquisition
software become important. The data acquisition software talks to this computer card and
brings in the digital data to Matlab. This software allows the user to read information from
the accelerometers. Data analysis begins once the GUI (Graphical User Interface) has been
initialized. Below is a diagram of the general flow of data:
There are not a wide variety of data acquisition software programs available to the design
team that are suitable. The Data Acquisition Toolbox in Matlab is the most suitable
NI Signal Conditioning Box: Analog
Analog Digital Converting Card inComputer
DAQ: Digital MATLAB Code To Analyze Data (GUI is used to initialize): Digital
Accelerometers: Analog
25
software because it is compatible with the already-owned signal conditioning hardware. It is
also built on the same platform as the GUI and analyzation code.
GUIDE To enable a user to initialize data collection and analysis through Matlab, a graphical user
interface has to be created. Creating one in Matlab is the most efficient option because this
will allow all parts of the design project program software to be built on the same platform.
In order to make use of a graphical user interface through Matlab, GUIDE has to be used.
GUIDE is an acronym for Graphical User Interface Development Environment. A set of
tools for creating GUIs is provided in GUIDE. When GUIDE is opened, the Layout
Editor appears. This Layout Editor is what contains the GUI as well as the control panel for
all GUIDE tools. Push buttons, pop-up menus, axes, and other components are available
from the component palette. They can be easily dragged into the layout area. Figure 3
shows an example of a GUIDE Layout Editor and its components.
Figure 3 Courtesy of www.mathworks.com
26
GUI
In order to make the road vibration simulation laboratory more user friendly, the design
team created a graphical user interface (GUI) in Matlab. A GUI is a program interface that
takes advantage of the computer’s graphics capabilities to make the program easier to use. If
designed appropriately, taking ergonomics into account, a GUI can free the user from
learning complex command languages. GUI’s can also make it easier to move data from one
application to another.
Some of the basic components of a GUI can be seen in Table 1. The design team used
some of these components, but not all of them, as the team wanted to keep their GUI as
simple and user friendly as possible.
GUI COMPONENTS
COMPONENT ACTION Pointer A symbol that appears on the display screen that
the users moves to select objects and commands.Pointing Device A device, such as a mouse, that enables the user
to select objects on the display screen. Icons Small pictures that represent commands, files, or
windows. Desktop The area on the display screen where icons are
grouped. Windows Areas into which the desktop is divided. Menus A way for the user to execute a command by
selecting a choice from a menu. Table 1
After reading manuals and documents about how to create a GUI using Graphical User
Interface Development Environment (GUIDE) command in Matlab, the team began
developing a GUI that they wanted to have be their final user interface that people using
their vibration simulator would use. After a short period, the team realized that the code
associated with the size of GUI that they wanted to use was quite large just for the different
components of the GUI such as pull down menus and graph windows. Once the team
27
starts adding code to make the GUI work with the Data Acquisition (DAQ) Toolbox and
other components of the simulator, the code will be quite complex, however, in order to add
this code in the right part of the GUI code, it is necessary for the team to understand the
code. When one large GUI is made at once, the code is too complicated for the team to
understand because of their relatively limited experience with Matlab. When developing a
GUI using Matlab, several lines of code are written for each element that is added to the
GUI. Because of this, the code behind the GUI can become quite complex very quickly. In
order to manage the amount of code they would have to work with, the team first developed
a simple example of a pull down menu. As can be seen in Figure 7, the team’s final GUI
contains several pull down menus, so it is import for the team to understand the code
associated with a pull down menu. Once they learn the basic code they will be able to add
more code and make the GUI do more complex operations. Similarly, because of the
amount of graphs on the team’s final GUI, the team first created a simple GUI that plots
just two functions of X based on a user input value. Understanding the code behind this
simplistic graph model will allow the team to make more complicated graph functions as
desired for the final GUI. The simple pull down menu and graph GUI’s can be seen in
Figure 4 and Figure 5 respectively. The code for these can be found in the Appendix for this
report.
Figure 4 Figure 5
28
Through research, the team found that when designing a GUI, it is very important to
consider ergonomics. Ergonomics is the study of designing for human safety and comfort.
The more ergonomically designed something is, the more likely the users are to use it
correctly without being injured. This is especially important when it comes to computers.
Some of the things that ergonomics focuses on in design are simplicity, consistency,
immediate feedback, and intuition.
Simplicity involves ease of use of the GUI. The simpler the GUI is to use, the more
easily the user will be able to figure out how to work the program, and thus gain more
benefits from it. According to congitivent.com, “the [GUI] designers attempted to
communicate with users more effectively by making the computer communicate in ways the
brain uses more readily; using icons for instance, because the visual part of the brain can
track their presence and state much better than words. They developed ways of organizing
information in patterns which the eye can track through more easily, drawing attention to
the work in progress. They developed the model of WYSIWYG (what you see is what you
get) to improve print proofing performance, and found through testing that the digital
representation of black text on a sheet of white paper increased information legibility and
retention.” Thus, including simplicity in the design of the GUI allows for users to use it
more effectively and learn more, which is exactly what the design team wants.
According to omnica.com, “there are a number of basic ergonomic guidelines which can
be incorporated into a product design that will eliminate or reduce learning curve times.”
Consistency, which is one of these guidelines, involves making sure everything on the GUI is
displayed in a consistent, almost repetitive manner. “Maintain consistency by making sure
similar or identical interfaces operate in a similar manner. Controls should also respond the
same way every time. While multifunctional buttons may conserve space and reduce cost,
29
they can also be a source of user aggravation.” The way the team achieved this in their GUI
is discussed further in the proposal section below.
Immediate feedback involves the ability of the user to perform an action with the GUI,
and them being allowed to see the results of their action immediately. According to
omnica.com, “if possible, the device should confirm to the user that input has been received
and is appropriate. Such feedback should be immediate, to prevent the user from repeating
the function. Some ways to accomplish this are to provide tactile, visual, or audio responses.
In addition, the designer can help prevent user confusion by giving different types of
feedback for dissimilar functions.” A perfect example of an item in a GUI is a pull down
menu. The user selects an item from the list in the pull down menu and they are
immediately able to see what they have selected. This makes the design much simpler and
the user is able to see whether they have performed a particular action correctly right away.
This improves productivity and learning. Again, according to omnica.com, immediate
feedback is an ergonomic guideline that, when incorporated into a design, will eliminate and
reduce learning curve times. The methods incorporated by the team in order to have
immediate feedback in their GUI are discussed in the following proposal selection.
Finally, intuition involves the idea that human tendencies can be incorporated into the
design of the GUI such that the actions that need to be performed by the user are very
intuitive in nature. According to mtnmath.com, “intuition is defined as knowing or sensing
without the use of rational processes.” It is important to incorporate this idea when
designing the GUI because if the GUI is designed such that students can just know or sense
what to do next, then it will be easier for them to perform the tasks required of them from
the GUI, perform the experiment and thus learn something from the whole simulation.
30
SIMULINK
Although the team would be unable to acquire the shaker as early in the design process
as was wanted, they did have the equipment to perform an impact hammer testing already
available in the vibrations laboratory. Because the team wanted to move forward on the
project in as many directions as possible early in the design process, it was decided that a
numerical experiment was needed that could analyze the differences in data resulting from
impact hammer excitation and shaker excitation. This would allow the team a larger portion
of time to develop analysis code and troubleshoot the problems that arose throughout the
process.
The team decided to use Matlab’s Simulink tool to develop the beginning model for the
numerical experiment. Simulink is a Matlab tool that can be used for modeling, simulating
and analyzing complex systems. Simulink’s ability to generate a wide range of signal types
made it fitting to the team’s objective for the numerical experiment.
Within the Simulink model, signal generators were used to create two different signals,
one to model the shaker excitation (a random signal), and one to model the impulse hammer
signal (two step signals added together, each with a different start time, such that the
addition results in an impulse force). These signals were then fed to a transfer function that
represents a general tendency for a system’s behavior, based on general trends for how the
output of the system results to input forces. The output response could be viewed through
the scopes placed at the end of the Simulink model. The Simulink model is shown in Figure
6.
31
Figure 6
The data from before and after the signals encounter the transfer function are outputted
to two separate files that can be loaded into a Matlab file. Within the Matlab file, the
function ‘fftmspec’ is used to take the Fourier transform of the input and output files and
convert them from time domain to frequency domain. From here, the signal output can be
divided by the signal input, and the transfer functions resulting from the two different forms
of excitation can be examined. The code developed so far for the corresponding Matlab file
is given in the appendix to this report.
For accurate results in this experiment, the team needed to ensure that the transfer
function used in the experiment was representative of a stable system. By making sure that
the poles (roots of the denominator of the transfer function equation) were negative, the
32
team made sure that the system model developed was indeed that of a stable system. Ideally,
the transfer function would be representative of the car; however, it was not necessary to
have that high of a degree of accuracy in the model to achieve the general purpose of
identifying differences in behavior.
The team is still working on the numerical experiment and hopes to use the data to
examine the affects which changing such experimental variables as windowing and averaging
have on the outcome of the transfer function.
CLASSROOM EXPERIMENTS
It is important to consider the importance of intuition in experiment and design, and as
previously mentioned, according to mtnmath.com, “intuition is defined as knowing or
sensing without the use of rational processes.” It is important to incorporate this idea when
designing an experiment because if the experiment is designed such that students can just
know or sense what to do next, then it will be easier for them to perform the experiment and
thus learn something from it. Continuing with this idea, the team wanted to use a car that
would respond much like a real size car since that is what students have had experience with.
Through research, the team found that 1/5th scale cars were more similar to a real car in
terms of weight, engine, tires and response than their 1/8th cars; which since students are
familiar with how a real car responds, using a 1/5th scale model increases the intuitiveness of
the experiment.
One important idea to consider when dealing with any type of experiment is the
accuracy of the data obtained from the experiment. This is important because there is no
point to doing the experiment and finding data from it if the data found is inaccurate and
therefore unable to be used. Because of this, it is important to look at ways of checking the
33
accuracy of data obtained from an experiment. There are several ways in which to
accomplish this with number of averages, cycle time, and windowing being some examples.
One technique for checking the accuracy of data taken from an experiment is changing
the number of averages taken when analyzing the data. The number of averages is going to
affect the accuracy of the data collection. Depending on the number of averages taken, the
outliers, high, or low points in the data, could be accentuated or diminished, thus affecting
accurate data collection. If your data remains consistent with many different numbers of
averages, then it can be considered that your data collection methods were accurate, but if
your data varies a lot depending on the number of averages taken, the data collection may
not have been too accurate. Once the data is sampled, then the Fast Fourier Transform is
computed to form graphs of the input excitation and output response. These functions are
averaged and used to compute the coherence function and the frequency response function;
both of which are important for modal data acquisition. The coherence function is used as a
data quality assessment tool which identifies how much of the output signal is related to the
measured input signal. The frequency response function contains information regarding the
system frequency and damping and a collection of frequency response functions contain
information regarding the mode shape of the system at the measured location.
Another technique for checking the accuracy of the data taken from an experiment is
sample time, which is going to affect the accuracy of the data because, as with any
experiment, a reasonable amount of time is needed in order to gather accurate data. In
general, it has been found that the more time the data is taken over, the more accurate the
results of the experiment. There does, however, reach a point, where increasing the cycle
time does not improve the accuracy of the taken data; which is something that is unique to
34
each individual experiment, and would therefore need to be analyzed and dealt with based
on the experiment being performed.
A third technique for checking the accuracy of your experimental data is a technique
known as windowing. Windowing is the process of adding a weighting function to the
measured data. This weighing function, or window, is used to force the data to better satisfy
the periodicity requirements of the Fourier Transform Process. This process minimizes the
distortion effects of leakage, which is a distortion in the data during the transformation of
time data to the frequency domain using the Fast Fourier Transform.
PROPOSALS
STRUCTURE
It was found that in America, the popular scale for remote control cars is like Dr. Khan
suggested, the 1/10th scale. The 1/5th scale cars were found primarily in Europe, so finding
retailers and information in the United States was quite difficult. After e-mailing companies
for information and getting little response, the design team determined that calling dealers
was the best way to get what they wanted. The team found that the best approach was to
explain the project at hand and then ask the companies for their opinion on what was
needed.
Upon discovering that a car with engine and electronics was nearly $1000, the team
decided they could make do with a realistic chassis and suspension system if it were less
expensive. After contacting several dealers, importers and race track companies explaining
the project and then asking if they could either sell the team parts or build the team
something, the team was given two options of put together machines, no electronics and no
35
engines. These two options were slightly over $500, which was the team’s maximum car
budget. While doing research focused on systems instead of cars, the team also came across
an offer from a company for a 1/5th scale car with engine and body, but no electronics, for a
discounted list price of $500 not including shipping. This was the only option that came
with a body, which added to the team’s intuitive initiative, making the test structure to look
more like a car than what a chassis and suspension system would look like. However, the
fact that the engine was included was the decision maker in determining which option the
team wanted. The team liked the idea of not having to add weight, which they would have
had to do with a chassis system since it would be missing key items like body, engine and gas
tank. The team also liked the fact that, for an extra $200 and a trip to the hobby shop,
future students or faculty could purchase the electronics to make the car run if research or
experiments warranted such a purchase. In order to fully justify their purchase decision, the
team constructed a selection matrix which can be seen in Table 2.
Selection Matrix (1-5)
Intuitiveness FunctionalityMetal Construction Suspension Cost Total:
Alro Racing- whole car 4 4 4 4 3 19
GB Direct-parts only 3 3 3 3 3 15
Victory Speedway-parts only
3 3 4 4 3 17
Build own-H frame & Suspension
2 2 4 3 4 15
Table 2
With the selection matrix, the design team weighed each option in several categories. A
discussion of the team’s definition of each category is necessary. The team considers
36
intuitiveness to be a measure of how much the car affects the students understanding of the
experiment. For example, with the whole car option, the team feels that this ranks highest
because the true to life appearance of the car will aid the student in understanding the
experiments application to a real car. With the built H frame option, the team felt that the
put together appearance of the structure would hinder the student in understanding that the
experiment was supposed to simulate a real car. Functionality is defined by the team as
opportunities for future use. The whole car option gets a four in this category because for
$200 someone could purchase the electronics to get the car to run if a running car was
needed for future senior design projects or other engineering classes. The suspension
systems built by GB Direct and Victory Speedway are rated with a three because there would
be a possibility that the companies supplying them could provide faculty with other parts to
enhance possible use. Therefore, the H frame option is rated with a two as there isn’t much
use outside of this project. Metal Construction is obvious in definition, but not so in
importance. The reason for using this category in the selection matrix is that metal
construction of the car affects the stiffness of the chassis and thus the realistic nature of the
model. All options were made with metal parts except there was some mention of older
plastic parts with the GB Direct offer. The team therefore rated this option slightly lower
than the rest. Suspension is used for obvious reasons as it is most important to the study of
vibrations in cars. The team rated the GB direct offer with a three because, although the
suspension system was made from realistic parts, the plastic used may affect the operation of
the system as a test structure. Also, the team rated the H frame option with a three because
the amateur construction of the system may affect the quality of suspension. And finally, cost
is another important aspect in determining selection. The only obvious savings came with
the H frame option as the team would be building most of it on their own. The other three
37
options were similarly priced at a little over $500, therefore all these options were rated the
same.
It was seen through the selection matrix, as well as the general team opinion, that the
entire car offer from Alro Racing, an importer of Technokit cars, was what the team wanted
for their experimental test structure. A picture of this car as well as its specifications can be
found in the Appendix. This car is made with metal parts and has four independent
suspensions. Each suspension system consists of 1/5th scale springs as well as hydraulic
shock absorbers. The car is 45 inches long, 14.5 inches wide and 10.5 inches high. Its
weight is 21 pounds. The car has been purchased and delivered to the team and they are
anxiously awaiting the first test with software and impact hammer.
SUPPORT STRUCTURE
The team’s two options for the support structure were:
• Suspend car from ceiling of vibrations lab
• Suspend car from independent support structure
As the team started looking into the feasibility of suspending the car from the ceiling, they
found that there was not a very good place to do this in the current vibrations lab. The team
also didn’t know what resources the vibrations laboratory would have in the new engineering
building. Finally, they agreed that they wanted to have the entire experiment be self-
contained and mobile. Therefore, the team chose to build an independent support structure
for the car. They will actually build the structure next semester in 449. A dimensioned Solid
Edge drawing can be found in the Appendix.
38
SHAKER
In order to achieve the desired results of the road vibration simulation, the design team
selected an electromagnetic shaker as the actuator for the experimental set up. This shaker
generates either a random or a sine force, and transmits the vibration force to the car
through the shaker’s stinger. For the 1/5 scale model car that is being used for the
experiment, many shaker companies that were contacted recommended a 50 lb sine force
shaker. A shaker that can produce a 50 lb sine force cannot generate as high of a force when
a random signal is generated, and the team determined that for the most accurate vibration
simulation, a random input force should be used. By looking at different shakers from
various companies, the team found that a 50 lb sine force shaker could generate
approximately 30 – 35 lb random force. However, a random signal force of 35 lbs will still
be enough to achieve a large enough amplitude for the experiments. Research also found
that a 50 lb shaker can usually only support a payload of 7 to 10 lb. By building a support
structure for the car this problem can be accounted for, even though the car weighs
approximately 21 lbs.
One of the first decisions that had to be made was the necessary size of the shaker. The
required size of the shaker had to correspond to the scale of the car. With a 1/5 scale model
car, a 50 lb sine force shaker could be used. If the car was any smaller, the team could
purchase a smaller shaker, but the results of the experiment would be less accurate. If the car
was any bigger, a bigger shaker would be required, which would be much more expensive.
The team contacted multiple shaker dealers, such as DataPhysics, Ling Dynamics, and MB
Dynamics. Although each company offered help in recommending equipment for the
project, MB Dynamics was the only company that submitted a price quote to the team. They
also offered an academic discount for the team. One of the benefits of this particular device
39
is its lightweight, compact design, which make it portable and easy to setup. Another
advantage is that the device is designed to excite small structures, making it applicable to the
team’s scaled-car testing setup. MB Dynamics has been extremely helpful in communicating
with us, and they are located in Cleveland, OH. This is a major advantage over some of the
other companies which were located overseas and thus would take a longer time to order
and would be harder to communicate with if problems were encountered with its
implementation into the team’s simulator design. None of the other companies felt they
could beat MB Dynamics’ price, and MB is a very reputable company that the team felt
comfortable dealing with. Therefore, the team chose to purchase the modal 50A exciter
from MB Dynamics. Specifications and a picture of this shaker can be found in the
Appendix.
Amplifier
MB Dynamics has a specific amplifier that is recommended for the modal 50A shaker.
Many of the other shaker companies such as DataPhysics and Ling Dynamics both
recommended purchasing an amplifier that was intended for the shaker that would be
purchased. Since no other alternative options are feasible for the project, the team chose to
purchase the amplifier from MB Dynamics.
Cooling System
The cooling systems are relatively cheap compared to the cost of a shaker and
amplifier, and by having a cooling system, the team will be able to use the shaker at its peak
force if necessary. To be safe, the team decided a cooling system would be a good
40
investment. The cooling system would also allow future experiments to be performed using
the shaker at whatever force was necessary. It was agreed that since the lab experiment was
supposed to be portable and self-contained, it would be best to purchase the portable
cooling system from MB Dynamics. Economically, this would also be the safest system to
buy, since it has not been determined if shop air will be available in the vibrations laboratory
in the new engineering building.
GUI
The design team’s final GUI for their vibration simulation laboratory can be seen in
Figure 7. As can be seen in Figure 7, the team’s GUI consists of three main areas; an area
for shaker input information, an area for the accelerometer output information, and an area
for the transfer function graphs.
Figure 7
41
The shaker input area has three pull down menus. One menu pull down menu is a
where the user can enter the number of averages. This is helpful information because the
number of averages taken during data collection can affect the accuracy of the data taken.
With this pull down menu option, the user can compare the data with different numbers of
averages taken, and make conclusions about how this affects the data. The second pull
down menu is for the sample time, which is the length of time during which the data is
taken. This can be input by the user, and will affect the outcome of the experiment, and can
therefore be changed to see how the output changes. The third and final pull down menu in
the shaker input area is for the force window. This will only have two options: on or off.
Windowing is one technique for checking the accuracy of experimental data. Windowing is
the process of adding a weighting function to the measured data. This weighting function,
or window, is used to force the data to better satisfy the periodicity requirements of the
Fourier Transform Process. This process minimizes the distortion effects of leakage, which
is a distortion in the data during the transformation of time data to the frequency domain
using the Fast Fourier Transform. The shaker input area on the GUI also includes a graph
window in which the shaker sine wave can be viewed by the user.
The accelerometer output area consists of a color key code to let the user know which
accelerometer is associated with which color on the graph and the actual graph output
window. The graph will show four different outputs, each one in a different color, that will
correspond to the data obtained from each respective accelerometer. Channel 0
corresponds to accelerometer 1 and is blue; channel 1 corresponds to accelerometer 2 and is
magenta; channel 2 corresponds to accelerometer 3 and is green; and channel 3 corresponds
to accelerometer 4 and is black. The accelerometer output area also includes two pull down
42
menus. The first menu is where the user can select the accelerometer number, then in the
second pull down menu the user selects the accelerometer location corresponding with the
chosen accelerometer number. The user then clicks the save button that is also located in
the accelerometer output area of the GUI. The save button saves the chosen accelerometer
number and corresponding location and the user can then choose more accelerometer
numbers and locations. This way, the user can view all the outputs from the accelerometer
in one graph.
The final area on the GUI is the transfer function area. This area contains a graph which
will output the transfer function of the data received from the Matlab Data Acquisition. The
user will not be able to input any information in this area, as it is only an area for the user to
view the results of the data collected.
As mentioned in the research section above, ergonomics in the design of many things
including GUIs. As the design team created their GUI, they kept the ergonomic ideas of
simplicity, consistency, immediate feedback, and intuition in mind. The team incorporated
simplicity into their design by making the number of input variables the user could change to
a minimum. This minimizes confusion on the part of the user and allows them to focus on
obtaining data from the program rather than entering it. Also, because the GUI is laid out
such that all the components line up at right angles, only the necessary information is
present, and much white space is used, the user isn’t confused by a lot of clutter or
randomness, which in turn enhances simplicity. Finally not many different types of
components are used, i.e. just pull down menus, push buttons, and graph windows, which
also helps add to the simplicity of the design. The simpler the design, the easier it is for the
user to use, which by definition is ergonomic design.
43
Consistency is shown in the team’s GUI because, as previously mentioned, all
components are in line, and appear in a logical order on the screen. First on the upper left
hand corner is the input information, followed by the output information in the upper right
hand corner, and finally there’s the transfer function graph area below these two and in the
middle. This makes sense because the transfer function is a combination of inputs and
outputs. The team’s GUI also shows consistency because all similar items have similar font
sizes for the headings; same sizes for each pull down menu, and the same size for the input
and output graphs. This makes the GUI easier to use as well as easier to look at, which
shows the team’s desire to design with ergonomics in mind.
Immediate feedback is achieved with this GUI design because it is created in Matlab and
the DAQ code as well as all other code associated with the vibration simulation is written in
Matlab, the GUI can immediately show the information obtained by the DAQ toolbox.
Also, pull down menus show immediate feedback because the user selects the desired option
and can immediately see what has been selected. The same is true for when the user selects
the accelerometer number and location and hits the save button; they can immediately what
combination they have chosen and what color the line on the graph will be. This is
important because if there wasn’t immediate feedback, then the user would have no way of
knowing what had been selected and whether or not they had selected the right option.
Immediate feedback is ergonomic design because the user doesn’t waste time looking for the
results, as they are created on the same screen as the inputs parameters are entered.
Finally, in order to enhance the ergonomic design of their GUI, the team incorporated
the concept of intuition when creating it. Because of the design of a pull down menu, as
well as the familiarity of pull down menus to most people, a pull down menu itself is very
44
intuitive. Most people know that the desired input needs to be selected from the pull down
list. As mentioned, the team incorporated several pull down menus in their GUI layout,
which makes the GUI intuitive. Also, the GUI area is laid out in a very logical order. With
the input area on the top left, the output area on the top right and the transfer function area
in the middle on the bottom, the order of flow of information follows the order in which
humans read. This is very ergonomic as well as the fact that the inputs are on the left, the
outputs on the right, and the transfer function, which is a combination the inputs and
outputs is in the middle.
Addressing all four of the aforementioned ergonomic issues makes the GUI much more
user friendly as well as limits the amount of confusion for the user. Both of these in turn
enhance the ability of the user to learn; which is the main focus of the team’s entire road
vibration simulation project. The code for these can be found in the Appendix for this
report.
DAQ
The design team chose to utilize the Data Acquisition Toolbox provided in Matlab because
no suitable systems for the team’s needs are available. These needs are based upon the signal
conditioning hardware already obtained (manufactured by National Instruments) as well as
the platform being used to design the experiments with. The Data Acquisition Toolbox in
Matlab is a collection of M-file functions as well as MEX-file dynamic like libraries built on
the Matlab technical computing environment. It allows for bringing in live, measured data
into Matlab. The Data Acquisition Toolbox within Matlab will be used because it is built on
the same platform as Matlab. This will allow for smooth transitions and less programming
conflicts. In order to test the Data Acquisition Toolbox, code was written in Matlab. This
45
code was tested by seeing if the computer could communicate with the accelerometers. The
code written can be found in the Appendix. The Data Acquisition Toolbox code that has
been written has been able to open up the lines of communication between the computer
and the accelerometers. Therefore, the Data Acquisition Toolbox in Matlab will be used to
talk to the computer card that converts analog to digital data and bring in the digital data
into Matlab.
FLOW OF INFORMATION
Figure 8
46
Figure 8 shown above demonstrates the flow of information between the software and
hardware of the team's experimental setup. The red lines show information that is being
sent from the computer to the car and the blue lines show information that is traveling back
as an input to the computer.
The information flow begins with a random signal generated by Matlab being sent from
the computer to the signal conditioning board. Because Matlab is only capable of generating
a signal on the order of 0-10 Volts and the necessary input voltage to the shaker is
sufficiently larger (on the order of 20 Volts) an amplifier is needed between the signal
conditioning board and the shaker. Once the signal reaches the shaker, the signal output
travels in two different directions. First, the shaker output is sent to the load cell in order to
measure the input force that is actually being sent to the car. This information is then
forwarded to Matlab by the load cell to be used in the data analysis. Second, the shaker
output is sent to the actual car, where the resulting output force of the car is measured by the
accelerometers. This information is then sent back to the signal conditioning board, where it
can also be used in Matlab for data analysis.
As a result of this information flow, both the input and output forces on the car are
being sent to Matlab. This information can be read from the Matlab GUI and analyzed by
the students according to the procedures developed in the corresponding labs.
CLASSROOM EXPERIMENTS
Because the goal of this entire project is to create a lab experiment that students taking
the vibrations class can use in the Miami University vibrations laboratory, the design team
had to keep in mind the experiments that they would write to go with their project. The
team wanted to focus on intuition when designing the experiments. The key objective of the
47
classroom experiments that the team is going to develop is intuitive learning. The team
wants to make sure that the students can perform the experiments such that a large amount
of intuition is involved on the student’s part. For example, as previously mentioned, the
GUI is designed with intuitiveness in mind. Also, the team chose the real 1/5th scale model
car because it looks and reacts the most like a real car. Because all the students would have
some idea as to how a real car performs, as they have certainly had much experience with
them throughout their lives, having a classroom experiment involving a car that very much
resembles a real car makes the experiment very intuitive, and by definition, the students will
learn more because the intuitiveness will allow them to perform the experiment more easily
and probably accurately.
The first classroom experiment the design team wants to develop and have the
vibrations students do in the laboratory is a basic experiment that introduces students to the
equipment as well as the basics of transfer functions. This experiment will involve a basic
introduction to the shaker and how to set it up with the amplifier, accessories kit, and
cooling system such that it performs in the correct manner. The students will also learn
about accelerometers and how to correctly connect them to the car and the signal
conditioning box. They will also learn how to connect the car properly to the support
structure. Not only will the students learn the basic hardware associated with this vibration
experiment, but they will also learn the basics of a transfer function. Because, at this point in
their schooling, students haven’t really had much experience with transfer functions, this
experiment will give them a basic idea of what a transfer function is and how they can use it
to analyze the data. This experiment will also involve learning how to use the GUI to input
data as well as run the experiment and get data. The overall point of this experiment will
48
only be to familiarize the student with all the components and steps involved in using this
road vibration simulator to gather data.
The second experiment the team will develop for the students to use perform in the
classroom is an experiment that involves changing the location of the shaker on the car and
comparing the different data acquired from the different shaker locations. The design team
would like to have a total of five preset locations at which the shaker can measure vibrations.
The team has currently selected the driver seat, front driver side wheel, rear driver side
wheel, front bumper, and rear bumper as locations to put the shaker. In order to make sure
the data taken by the students is consistent, the team will mark, using a permanent marker or
perhaps a paint pen, points on the aforementioned parts of the car so that the shaker is
placed in the same spot on that particular part of the car every time an experiment is
performed. Also, specifics in the Matlab code will pertain to these points, and thus it is
critical to place the shaker at the same point each time data is collected on that point.
With this experiment, the team hopes to allow the students to see how this model car
responds to vibrations in a similar manner as a real size car. The team wants to show the
students how the relationship between the vibrations felt in the driver seat as compared to
the driver side front wheel in the model car is similar to the relationship between vibrations
felt in those two spots on a real car. This will show students how vibrations are an actual
real world phenomenon that must be carefully analyzed and dealt with in the real world.
The team wants the students to be able to be able to see how vibrations they feel when
riding in a real car can be modeled by experiments and thus dealt with. The team hopes that
students will see that they are performing this experiment for a real reason, as it has real
world applications, and that this is not just a pointless experiment with no applications to
their lives. The team is certain that, since most students have undoubtedly had large
49
amounts of experience with cars in their lives that they will appreciate the work that goes
into making sure they have a comfortable ride when in a car.
Finally, a possible third experiment that the design team would like to develop for the
classroom is an experiment that teaches students about ways to analyze the accuracy of the
data collection technique they have used during the simulation. As can be seen in the GUI
discussion above, the shaker input area has three pull down menus; one for number of
averages, one for sample time, and one for windowing. These three items are all different
ways of analyzing and checking the accuracy of the data collection method.
The design team wants the students to be able to select the number of averages taken
when taking their data, and then see how changing the number of averages affects the data.
From this, they will be able to see whether they have collected the data accurately. The
team will have the students vary the sample time of the simulation to help them understand
the concept that increasing sample time will increase the accuracy of the data taken. Another
possibility with this experiment is to have the students analyze at what point it becomes
unnecessary to increase sample time in order to obtain more accurate data.
The final way the team will have the students check for the accuracy of their data is
through the technique of windowing. As discussed in the preceding research section,
windowing is adding a function to the data to make it easier to analyze and read. This
experiment will just be very basic when it comes to widowing, as windowing is a difficult
concept and can be very complicated. The team just wants the students to be exposed to the
idea of windowing and how it can help analyze the accuracy of the data collected.
The design team feels that if they have the students perform the above experiments that
they will gain a good insight into what goes into simulating road vibrations while learning
50
about how to interface hardware equipment with software, setting inputs and taking data,
then seeing how it applies to the real world.
CONCLUSIONS AND FUTURE WORK
CONCLUSIONS
The design and development of a road vibration simulator for the vibration laboratory
has many components to it. The initial component to be decided upon is the physical
structure to be used as the specimen being tested. The three preliminary ideas for this were
an entire car, an entire car scale model, or an individual car component. The design team
chose to design the simulator for use with a scaled model of a car, mostly based on
laboratory constraints and the team’s goal of intuitive learning. The next component to be
decided upon is the type of sensor to be used. The two main preliminary ideas for this were
force sensors or accelerometers. After performing some initial research, it was decided that
a force sensor would be used to measure the input while an accelerometer would be used to
measure output. The third component to be decided upon is the actuator. There were three
preliminary ideas for this: an electrodynamic shaker, a hydraulic shaker, or an impact tester.
The design team decided, after performing much research, that an electrodynamic shaker
would be the most suitable option for this simulator. The final component that needed to
be chosen was the type of data acquisition/analysis package to use. The preliminary ideas
were Matlab and Labview. Matlab was chosen as the type of data acquisition and analysis
package, primarily due to the fact a majority of the design team members have prior
experience with Matlab.
In terms of Hardware, once the team had decided to use a scaled model of a car, the
exact size and the particular model needed to be selected. After weighing the alternatives of
51
different scaled models and obtaining price quotes for several options, the team purchased
the TKT99 Junior 1/10 scaled model car from Alro Racing. The team also needed to select
a particular shaker to be purchased for the setup. The Modal 50A Exciter from MB
Dynamics was selected. This shaker was purchased along with the necessary corresponding
components (amplifier, cooling system, accessories kit) for this specific device. Finally, a
support structure was designed that could support the car from bungee chords during the
experimental testing.
In the software category, the team developed a mathematical experiment in Simulink that
would help them to understand the differences between shaker and impact hammer testing.
Code for data acquisition was also developed to allow the computer to communicate with
the data acquisition hardware. An overall goal for a GUI was designed, as well as individual
working GUI components that would allow for better understanding of development code.
Finally, three preliminary ideas were developed for laboratory experiments. These ideas
include an introduction to experimental testing and transfer functions lab, an analysis of
shaker location and relation to real world automobile lab, and a data collection and analysis
issues lab. The teams design goal of intuitive learning was a primary consideration in the
development of these designs.
FUTURE WORK
As can be seen in the Timeline, in Gantt Chart form, attached in the appendix, nearly
two-thirds of the project tasks have been either started or completed. All of the designs for
the system have been finalized during 448, leaving the only the remainder of the
implementation process to be completed during 449. For each of the major sections of the
52
project, (hardware, software, and experiments), specific tasks are described below that will
need to be completed during EGR 449. In addition, the team must constantly keep in mind
that it is the integrating of these three sections (hardware, software, and experiments) that is
key to the success of the road vibrations simulator.
Task Name Duration [Day] Start Date Finish Date
Test Components 45 days Mon 1/26 Fri 3/26
Assembly of Apparatus 15 days Mon 3/29 Fri 4/16
Develop Data Collection/Analysis Software 120 days Mon 11/3 Fri 4/16
Writing of Final Report 130 days Mon 11/3 Fri 4/30
Write Lab Manual 110 days Mon 11/17 Fri 4/16
Run User Test on EGR 315 Students 5 days Fri 4/16 Thu 4/22
Post Information to the Internet 3 days Mon 4/19 Wed 4/21
Table 3
53
REFERENCES
ALRO Racing: http://www.alroracingusa.com/ Technokit: http://www.technokit.it GB Direct: http://www.gbdirectusa.com/ Victory Speedway: http://www.victoryspeedway.com/ FG Modelsport: http://www.fg-america.com/site/lv1/productlines.htm RDP Corporation: http://www.rdp-corp.com/ MB Dynamics: www.mbdynamics.com Data Physics: http://www.dataphysic.co.uk/html/shakers.htm Random Vibration: http://www.algor.com/products/analysis_types/linear_dynamics/random_vibration.asp Modal Exciters: http://www.bksv.com/2329.htm Shaker Selection: http:/autotestreport.com/2002_articles/09_shaker.htm Ling Dynamics: http://www.lds-group.com/ Labworks, Inc: http://www.labworks-inc.com Derritron: http://www.derritron.com/ Quick Tech Hobby: http://www.quicktechhobby.com/prices/cars/5%20SCALE.htm# Ready-to-Run Modal Analysis: http://www.mecha.uni-stuttgart.de/Forschung/forschung_exp_meth/forschung_expmodal_en.htm Modal Analysis: http://www.gmi.edu/~drussell/GMI-Acoustics/Modal.html Modal Analysis: http://www.mts.com/nvd/PDF/1992_AMATLST.pdf Harm Racing: http://www.harm-racing.de/Cars_Street.asp?Lang=EN MCD Racing: www.mcdracing.com GUI Ergonomics: http://cognitivevent.com/gui.html Intuition Information: http://www.mtnmath.com/whatrh/node105.html GUI Ergonomics: http://www.omnica.com/omniview_industrial_design2.htm Math Works: www.mathworks.com National Instruments: www.nationalinstruments.com Mathworks. Data Acquisition Toolbox: For Use with Matlab User’s Guide. Version 2. 2001.
The Mathworks, Inc. Natick, MA.
54
ABOUT THE TEAM
There was a wide variety of engineering background within the group, with Manufacturing,
Management, Mechanical, and Environmental majors represented. This variety of
backgrounds has led to a wide array of skills, and many different approaches to problem
solving. Each member of the group shares the same feeling learning style. The group was
able to take advantage of the shared learning styles by taking a hands-on approach to the
project, for instance scheduling a trip to see a full scale road vibration simulator at the
University of Cincinnati. Two of the group members, Sarah and Mandy, have taken a
Vibrations class at Miami University, and Mandy spent last summer at Miami assisting in
vibrations research. Sarah and Jessica also had professional experience in the Engineering
Industry. Their prior knowledge and ability to share this knowledge with the rest of the team
has greatly benefited the group.
Each member of the group was interested in the subject matter behind this project, and the
team wanted to be able to help improve the Engineering Department at Miami University.
At first, small tasks were divided up between group members, and the main tasks were
accomplished as a group. As the semester progressed, the team decided that the main
sections of the project should be divided between subgroups within the team, so the team
divided into 2 sections. Due to their experience with Matlab, software development, and
vibrations software, Jessica, Mandy, and Sarah focused on software development for the
group. Due to their experiences as Engineering Management majors, Carolyn and Matt
worked primarily on the selection and acquisition of hardware for the experiment. Still, the
group met multiple times each week in order to ensure that all members of the group were
aware of what was being accomplished. By doing this, the group was able to maximize the