Mechatronics

The Practice of Multidisciplinary Systems Engineering

Dr. Kevin Craig Professor of Mechanical Engineering Rensselaer Polytechnic Institute

Office: JEC 2042 Phone: 276-6671 E-mail: craigk@rpi.edu FAX: 518-276-6025

Course Description:
Mechatronics, as an engineering discipline, is the synergistic combination of mechanical engineering, electronics, control engineering, and computers, all integrated through the design process. It involves the application of complex decision making to the operation of physical systems. Mechatronic systems depend on computer software for their unique functionality. This course studies mechatronics at a theoretical and practical level; balance between theory/analysis and hardware implementation is emphasized; emphasis is placed on physical understanding rather than on mathematical formalities. A case-study, problem-solving approach, with hardware demonstrations, either on video or in class, and hardware lab exercises, is used throughout the course. This course covers the fundamental areas of technology on which successful mechatronic system designs are based: physical modeling, from design model to truth model, and mathematical modeling of dynamic multidisciplinary physical systems; analysis of mathematical models through analysis and computer simulation; measurement systems (analog and digital sensor modeling, analysis, and implementation) for model validation and control; control actuator (electromechanical and fluid) modeling, analysis, and implementation; continuous controller design and real-time analog and digital implementation; analog and digital control and power electronics. Throughout the coverage the focus is kept on the role of each of these areas in the overall design process and how these key areas are integrated, from the very beginning of the design process, into a successful mechatronic system design. Starting at design and continuing through manufacture, mechatronic designs optimize the available mix of technologies to produce quality precision products and systems in a timely manner with features the customer wants. The real benefits to industry of a mechatronic approach to design are shorter development cycles, lower costs, and increased quality, reliability, and performance.

Course Prerequisites: Basic Knowledge of:

Modeling and Analysis of Dynamic Systems Electronics and Instrumentation Feedback Control of Dynamic Systems

Course Learning Objectives:

Understand the importance of the integration of modeling and controls in the design of mechatronic systems. Understand the dynamic system investigation process and be able to apply it to a variety of dynamic physical systems. Understand the importance of physical and mathematical modeling (both from first principles and using system identification experimental techniques) in mechatronic system design and be able to model and analyze mechanical, electrical, magnetic, fluid, thermal, and multidisciplinary systems and identify the analogies among the various physical systems. Understand and be able to model various nonlinear and parasitic effects in real dynamic systems: backlash, time delay, saturation, Coulomb friction, unmodeled resonances. Be able to develop a hierarchy of physical models for a dynamic system, from a truth model to a design model, and understand the appropriate use of this hierarchy of models. Become proficient in the use of MatLab/Simulink and LabVIEW to model and analyze nonlinear and linear mechatronic systems.

Understand the key elements of a measurement system and the basic performance specifications and models of a variety of analog and digital mechatronic sensors. Understand the characteristics and models of various electromechanical actuators (brushed dc motor, brushless dc motor, and stepper motor) and hydraulic and pneumatic actuators. Understand analog and digital circuits and components and semiconductor electronics as they apply to mechatronic systems. Understand the fundamentals of power electronics as it applies to mechatronic system actuators. Understand stability (absolute and relative) and performance (command following, disturbance rejection, robustness) as it applies to feedback control systems. Understand and be able to apply, with the use of the MatLab and LabVIEW, various control system design techniques: open-loop feedforward control, classical feedback control (rootlocus and frequency response), and state-space control. Understand industrial motion control: types of controllers (PID-type control modes and variations), tuning of controllers, and position/velocity control loops with encoders/ resolvers. Understand the digital implementation of control and basic digital control design techniques. Become proficient in the use of MatLab and LabVIEW to design and analyze analog controllers and verify their digital implementation. Have an awareness of more advanced control design techniques, e.g., model predictive control, adaptive control, fuzzy logic control, and multivariable control. Be able to implement a real-time controller through the use of National Instruments control hardware and LabVIEW programming. Be able to apply all these skills to the design of a mechatronic system.

The true mechatronics engineer is that rare individual who has a genuine interest and ability across a wide range of technologies, and who takes delight in working across disciplinary boundaries to identify and use the particular blend of technologies which will provide the most economic, elegant, and appropriate solution to the problem in hand. Furthermore, he/she is a high communicator who has the knack of being able to enthuse others about technologies outside their own, and hence to break down built-in resistance to the use of alternative approaches. To evaluate concepts generated during the design process, without building and testing each one, the mechatronics engineer must be skilled in the modeling, analysis, and control of dynamic systems and understand the key issues in hardware implementation. This course strives to develop in each student a balance between these. This course studies in depth the key areas of technology on which successful mechatronic designs are based and thus lays the foundation for the students to become true mechatronic engineers.

Course Topics and Schedule:

Session 1 Introduction to Mechatronics and the Integration of Modeling & Control Dynamic System Investigation Process Physical Modeling: General Mathematical Modeling: General Modeling: General Concepts Physical & Mathematical Modeling: Mechanical Physical & Mathematical Modeling: Electrical Time Domain vs. Frequency Domain Analogies Among Dynamic Physical Systems Dynamic System Analytical Time Response and Frequency Response Zero-Order Dynamic System Response 1st-Order Dynamic System Response 2nd-Order Dynamic System Response Higher-Order Dynamic System Response Dynamic System Response by Numerical Simulation Physical & Mathematical Modeling: Magnetic and Electromechanical Analogies Among Dynamic Physical Systems Electromechanical Mechatronic Actuators Solenoid, Vibration Exciter Motors: Brushed DC, Stepper, and Brushless DC Introduction to Control Systems Analog vs. Digital Control Open-Loop vs. Feedback Control Stability of Feedback Control Systems: Absolute and Relative Performance of Feedback Control Systems: Time Domain & Frequency Domain; Command Following, Disturbance Rejection, Robustness Classical Control Analysis and Design: Root-Locus and Frequency-Response Techniques Industrial Control Types Measurement Systems Types of Applications of Measurement Instrumentation Generalized Configurations and Functional Descriptions of Measuring Instruments Sensor Fusion Loading Effects Performance Specifications: Static and Dynamic Analog & Digital Sensors for Motion Control

Session 2

Session 3

Session 4

Session 5

 Electronics for Measurement and Control Introduction to Power Electronics Mechatronic System Case Studies Magnetic Levitation System Balancing Robot / Segway Human Transporter Inverted Pendulum Systems: Translational, Rotary, Planetary Gear Introduction to Nonlinear and Parasitic Effects Session 7 Coulomb Friction Gear Backlash Unmodeled Resonances Time Delay Saturation Session 8 Physical & Mathematical Modeling: Thermal Physical & Mathematical Modeling: Fluid (Hydraulic & Pneumatic) Analogies Among Dynamic Physical Systems Physical & Mathematical Modeling: Multidisciplinary Session 9 Fluid Mechatronic Actuators Hydraulic Actuator Modeling, Analysis, and Control Pneumatic Actuator Modeling, Analysis, and Control Hydraulic / Pneumatic Hybrid Systems Session 10 Control System Design State-Space Control Design Digital Control Design Industrial Motion Control PID-type Controllers Controller Tuning Position and Velocity Control Loops Advanced Control Techniques: Overview Session 11 Mechatronic System Case Studies Continued Magnetic Levitation System Balancing Robot / Segway Human Transporter Automobile Traction Control Testbed Inverted Pendulum Systems: Translational, Rotary, Planetary Gear Hydraulically Balanced Beam System Pneumatic Positioning System Hybrid Pneumatic / Hydraulic Positioning System Temperature Control System Session 12 Mechatronic Industrial Case Studies Session 6

General References:
1. 2. 3. 4. 5. 6. 7. 8. 9. Modeling and Analysis of Dynamic Systems Dynamics of Physical Systems, R.H. Cannon, McGraw-Hill, 1967. System Dynamics, E. O. Doebelin, Marcel Dekker, 1998. System Modeling and Response: Theoretical and Experimental Approaches, E.O. Doebelin, Wiley, 1980. Modeling, Analysis, and Control of Dynamic Systems, W.J. Palm, 2nd Edition, Wiley, 1999. Modeling and Simulation of Dynamic Systems, R. Woods and K. Lawrence, Prentice Hall, 1997. Analysis and Design of Dynamic Systems, 3rd Edition, I. Cochin and W. Cadwallender, Addison-Wesley, 1997. System Dynamics, 3rd Edition, K. Ogata, Prentice-Hall, 1998. System Dynamics: An Introduction, D. Rowell and D. Wormley, Prentice-Hall, 1997. Modeling and Analysis of Dynamic Systems, 2nd Edition, C. Close and D. Frederick, Houghton Mifflin, 1993.

Control Systems 1. Control System Principles and Design, E.O. Doebelin, Wiley, 1985 2. Feedback Control of Dynamic Systems, Franklin, G., Powell, J., and Emami-Naeini, A., 4th Edition, Prentice Hall, 2002. 3. Control System Design Guide, 2nd Edition, G. Ellis, Academic Press, 2000. 4. Modern Control Engineering, 4th Edition, K. Ogata, Prentice Hall, 2002. 5. Observers in Control Systems, G. Ellis, Academic Press, 2002. 6. Advanced Control System Design, B. Friedland, Prentice Hall, 1996. 7. Control of Fluid Power: Analysis and Design, 2nd Edition, D. McCloy and H.R. Martin, Ellis Horwood Limited, 1980. 8. Digital Control of Dynamic Systems, Franklin, G., Powell, J., and Workman, M., 3rd Edition, Addison-Wesley, 1998. 9. Discrete-Time Control Systems, 2nd Edition, K. Ogata, Prentice-Hall, 1995. 10. Computer Control of Machines and Processes, J. Bollinger & N. Duffie, Addison-Wesley, 1989. 11. The Control Handbook, W. Levine, Editor, CRC press, 1996. 1. 2. 3. 4. 5. 6. 7. 8. 9. Sensors and Actuators Measurement Systems, E.O. Doebelin, 4th Edition, McGraw-Hill, 1990. Control Sensors and Actuators, C.W. deSilva, Prentice-Hall, 1989. Electromechanical Motion Devices, P. Krause and O. Wasynczuk, McGraw Hill, 1989. Electric Machinery Fundamentals, 3rd Edition, S. Chapman, McGraw Hill, 1999. Driving Force, The Natural Magic of Magnets, J. Livingston, Harvard University Press, 1996. Applied Electromagnetics, M. Plonus, McGraw Hill, 1978. Electromechanical Dynamics, H. Woodson and J. Melcher, Wiley, 1968. Electromechanics and Electric Machines, S. Nasar and L. Unnewehr, Wiley, 1979. Stepping Motors: A Guide to Modern Theory and Practice, 3rd Edition, P. Acarnley, IEE, 1992.

10. Step Motor System Design Handbook, 2nd Edition, A. Leenhouts, Litchfield Engineering Company, 1997. 11. Stepping Motors and Their Microprocessor Controls, 2nd Edition, T. Kenjo, Oxford, 1994. 12. DC Motors, Speed Controls, Servo Systems, The Electro-Craft Engineering Handbook, Reliance Motion Control, Inc. 13. Electric Motors and Their Controls, T. Kenjo, Oxford 1991. 14. Basics of Design Engineering, Machine Design, 1994. 15. Hydraulic Control Systems, H. Merritt, Wiley, 1967. 16. The Analysis and Design of Pneumatic Systems, B. Andersen, Wiley, 1967. Electronics and Microcontrollers 1. Design with Microprocessors for Mechanical Engineers, A.K. Stiffler, McGraw-Hill, 1992. 2. Art of Electronics, Horowitz, P. and Hill, W., 2nd Edition, Cambridge University Press, 1989. 3. Principles and Applications of Electrical Engineering, G. Rizzoni, 4th Edition, McGraw Hill, 2003. 4. The Art of Designing Embedded Systems, J. Ganssle, Newnes, 2000.

Homework Assignments:
11 assignments will be given during the 12-week course. Each assignment will be a combination of problems and concept questions. They are to be individually and professionally done and handed in at the beginning of the class when due. Collaborative discussion (Not Copying!) on homework assignments is permitted and encouraged. MatLab (including Simulink and the Control System Toolbox) and LabVIEW will be used extensively in the solution of these assignments. The assignments will be graded, with the lowest grade dropped, and will count 100% (10 @ 10% each) of the final grade. No late assignments will be accepted without consultation with the instructor prior to the due date for the assignment.

Class Attendance and Participation:

Attendance at all classes is mandatory and participation in class is strongly encouraged. Each session will be conducted with an interactive, applied, mentoring approach. There will be minimal presentation of the lecture material; in each session there will be application of the lecture material to real-world situations and problems. Questions will be answered and difficult concepts explained.

Course Conduct and Academic Integrity:

Students are expected to conduct themselves in a professional manner at all times. Studentteacher relationships are built on trust. Acts which violate this trust undermine the educational process. The Rensselaer Handbook of Student Rights and Responsibilities defines various forms of academic dishonesty and you should make yourself familiar with these; all forms are violations of the trust between students and teachers. Collaborative discussion on homework assignments is encouraged, but copying is not permitted; all homework assignments must represent the students own work. Submission of any homework assignment in violation of this policy will result in a grade of zero for that homework assignment.

Engineering System Investigation Process

The engineering system investigation process is a procedure an engineer follows to thoroughly investigate, i.e., understand, predict, and experimentally verify, how a dynamic engineering system or device performs, no matter how simple or complex the system may be. It is an iterative process, as understanding how the system performs requires simplifying assumptions initially. These initial simplifying assumptions may later be relaxed or changed as understanding develops through comparison of analytical predictions with experimental observations. Comparing the predicted dynamic behavior with the actual measured dynamic behavior is the key step in the investigation process. It is important to note that the steps in this process should be applied not only when an actual physical system exists and one desires to understand and predict its behavior, but also when the physical system is a concept in the design process that needs to be analyzed and evaluated. After recognizing a need for a new product or service, one uses past experience (personal and vicarious), awareness of existing hardware, understanding of physical laws, and creativity to generate design concepts. The importance of modeling and analysis in the design process has never been more important. These design concepts can no longer be evaluated by the build-and-test approach because it is too costly and time consuming. Validating the predicted dynamic behavior in this case, when no actual physical system exists, then becomes even more dependent on one's past hardware and experimental experience. The process, shown in the diagram, consists of the following steps:

Physical System The process starts with an actual physical system or product. It could be an actual engineered product or device, e.g., a bar code scanner or artificial organ, or it could be a basic dynamic system used for instructional purposes, e.g., a spring-mass mechanical system or a resistor-capacitor electrical low-pass filter. The physical system must be completely understood. How does it work? What materials does it use? What problem was it designed to solve? What need was it meant to satisfy? Who was the customer? Why was it designed the way it was? Why is it innovative? What alternative designs were considered? Also, when concepts are developed as part of the engineering design process, the physical system could be one of those concepts which needs to be understood and evaluated, not by building and testing it, but through modeling, analysis, and prediction with some experimental verification. Lets use as an example of a physical system the spring-mass dynamic system, the simplest mechanical dynamic system one could create. A picture of the system is shown. One finds springs and moving masses in vehicle suspension systems and in automatic machinery of all kinds. But what is most important to understand is that the essential characteristics (springiness, mass, and energy loss) of the spring-mass system are present in almost every mechanical system. We will develop this concept more fully as we proceed. The spring-mass system shown consists of a mass hanging at the tip of a tension spring (shown at the right) that is attached to a stationary support. A tension spring can only be stretched. In its rest state, the spring coils are pulled together against each other so no compression of the spring is possible. The motion of the spring-mass system is constrained by a linear ballbearing on the side of the support so that the mass oscillates only in one direction, the vertical direction. For measurement purposes, a non-contact optical (infrared) sensor attached to the base is used to measure the position of the mass; the output of the sensor is an electrical voltage proportional to the mass position. The mass is free to oscillate up and down with energy being dissipated mainly by the friction in the bearing and also by the cyclical motion (stretching and relaxing) of the spring. Air resistance has a minimal effect on the motion of the mass. The mass is set in motion by simply displacing it from its static equilibrium position (i.e., its rest position hanging motionless from the spring) and then releasing it.

10

 Physical Model This step is the key step in the entire process. Unfortunately, it is the least understood. By the use of simplifying assumptions and engineering judgment, learned through much repetition, we develop a physical model, a slice of reality, which is not an actual piece of hardware, but an approximation of the actual system capturing the essential elements of the actual system in as much detail as the need for the model requires. There is a hierarchy of models possible from the less-complex, less-realistic, more-easily-solved design model to the more-complex, more-realistic, less-easily-solved truth model depending on the particular need for the model, e.g., design iteration, control system design, final verification before hardware implementation. Always ask the question Why am I modeling? remembering that a model only has to satisfy the defined need for the range of operation being considered.

Here are simplifying assumptions that one might make to get a fundamental understanding of how the spring-mass systems behaves. The support to which the spring is attached is rigid, i.e., the support does not move as the mass attached to the spring moves. This assumption in effect says that the environment (everything outside of the system boundary) is independent of system motions. The spring is pure, i.e., it only has the characteristic (elasticity or springiness) for which it is named. A pure spring has negligible mass and energy dissipation (damping). This, of course, is an idealization as all springs have mass and dissipate energy upon cycling. If the spring mass is less than 10% of the mass attached to it, neglecting its mass is a reasonable assumption (except in high-speed applications). The energy dissipation in the spring is very small compared to other energy dissipation mechanisms in the system, so neglecting it is also reasonable. The spring is ideal, i.e., there is a linear relationship between spring force and spring displacement in the range of mass motion considered. This can be experimentally verified. The actual spring is a tension spring with some pretension (a force that pulls the coils of the spring together) and so there is a threshold force needed before the spring actually begins to stretch. Also the motion of the mass must be restricted to the range during which the spring is in tension, i.e., large amplitudes of motion of the mass are excluded from consideration.

11

 A schematic representation of a pure and ideal spring is shown, along with a graph of its force vs. displacement behavior.

The attached mass can be treated as a rigid body, i.e., the attached mass does not deform in any way. The mass moves with one degree of freedom in pure translation in the vertical plane. There is no out-of-plane motion and there is no rotational motion of the mass. The mass then can be treated as a point mass. The friction in the system is not intentional; it is what we call parasitic, i.e., no energy dissipation mechanism has been intentionally designed into the system. Compared to the linear bearing friction, air damping due to the motion of the mass in the air is negligible, as is the energy dissipated by the spring. The friction in the linear bearing is the main source of energy dissipation and, based on engineering experience, is a combination of viscous fluid damping (proportional to the velocity of the mass and directed opposite to the mass motion) and dry-friction, or Coulomb damping (essentially constant in magnitude, independent of mass velocity, and directed opposite to the mass motion). Coulomb friction leads to a nonlinear mathematical model, while viscous fluid friction leads to a linear mathematical model. Linear equations are not only easier to solve, but are also easier to get insight into the physical behavior of the system from. However, the desire to have a linear mathematical model does not justify the assumption of viscous fluid damping and the omission of Coulomb damping. If this assumption is not based on sound engineering judgment, then the resulting mathematical model will not predict the actual behavior

12

of the dynamic system. Lets assume that we can capture the major portion of the energy dissipation with the viscous fluid damping device called a damper. This device, like a spring, will be assumed to be pure and ideal, i.e., it has no mass and no springiness, only energy dissipation, and it behaves in a linear fashion, i.e., the force exerted at the ends of the damper, F, is proportional to the difference in velocity of the ends of the damper, v2 v1, which is referred to as the relative velocity of the ends of the damper. The constant of proportionality is called the viscous damping coefficient, B. A shock absorber in your car is an example of a mechanical damper that is intentionally designed into every car. A schematic of a typical actual device along with its engineering symbol and linear relationship are shown.

The system is vertical with the acceleration due to gravity, g, pointing downward and constant in value. All parameters (mass, spring constant, viscous fluid damping coefficient) are constant, i.e., do not change with time or temperature, for example. With these assumptions, a physical model of this physical system looks like the schematic shown below.

13

Physical Modeling Example: As another example consider the simple cantilever beam made of steel. A cantilever beam is a beam fixed to ground at one end while the other end is free. It has strain gages attached to it at its base which are sensors used to determine the amount of beam deflection at the end of the beam when either the beam is loaded with a mass placed on the end of the beam or is simply plucked and allowed to vibrate. The beam bends when loaded, as it is flexible, not rigid it has compliance or springiness and acts like a spring. The beam itself also has mass and when it is plucked and allowed to vibrate, the vibrations will eventually cease as there is dissipation or loss of energy during each cycle of its motion as the steel beam elastically deforms. So this system has the distributed characteristics of mass, springiness, and energy dissipation the characteristics are distributed throughout the beam you cant isolate any one of those characteristics. What would a physical model of this physical system look like? If we are concerned with motion of the tip of the beam, then a physical model of this physical system would look just like the spring-mass-damper system model shown below. While the structure of the model is correct, the challenge of quantifying the parameters in the physical model often makes physical modeling of real devices and systems difficult. This leads us to the next step in the process.

Model Parameter Identification The physical model has elements, not necessarily corresponding to actual physical components in the physical system, with characteristic parameters (spring constant, mass,

14

resistance, inductance, thermal conductivity, fluid viscosity, thermal capacitance, etc.) whose numerical values must be identified. This is done either by numerical calculation, referencing standard handbooks, using vendor information, or through experiment. For the spring-mass system the parameters that were identified are: mass M (5.23 kg) of the attached block obtained by weighing the block; spring constant K (491 N/m) of the spring obtained by attaching two different masses to the spring and measuring the spring stretch for each mass. These two data points (weight and spring displacement) are then plotted so that the assumed linear (straight-line) behavior of the spring can be determined; and the viscous damping coefficient B (1.1 N-s/m) obtained by adjusting the value of B and curve fitting the predicted response with the measured response. For the cantilever beam physical model, while the spring constant K in the model can be calculated from the dimensions of the beam (length L, width b, and thickness, h) and from the beam material property called the modulus of elasticity, E, the values of the mass M and the viscous damping coefficient B need to be determined. Often experiments are used for this determination. Ebh 3 and M = 0.23M beam . Analytically, for the cantilever beam, K = 4L3

Mathematical Model The laws of nature (physics, chemistry, biology) are applied to the physical model (not the physical system) and the mathematical equations describing the system are derived. Here is where the fundamental body of knowledge in science is applied in the process. The laws of physics are primarily used in our investigations, e.g., Newtons Laws, Conservation of Mass, Conservation of Energy, Kirchhoffs Voltage Law, Kirchhoffs Current Law, Faradays Law, Amperes Law, and Lenzs Law. We will use all these laws and apply them to physical models to generate mathematical models. In the case of the spring-mass system, we apply Newtons 2nd Law of Motion to a freebody diagram of the mass in our physical system model to obtain the equation of motion for the system, the mathematical model. Mathematical Analysis: Predicted Behavior The mathematical equations are solved either numerically by computer simulation or analytically (using appropriate theory to obtain a mathematical expression for the solution) to predict the behavior of the engineering system. The purpose of modeling is to gain insight into the behavior of the engineering system. Using a simpler model that allows for an analytical solution often leads to greater insight into system behavior than numerical solutions of a more complicated model. What should be the input into our system to make something happen, i.e., excite a response, so that we can compare a measured response with a predicted response? Engineers use specific inputs to evaluate performance of dynamic systems and compare alternative designs. The inputs used, both in the actual physical system and also in the mathematical model, lead to two complementary points of view: the time domain and the frequency domain. Understanding both points of view is essential for an engineer. Together, time domain and frequency domain give a complete picture of the behavior of a dynamic system. They are essential and complementary.

15

This step is only half the story, for computer simulation or mathematical analysis without experimental verification is at best questionable, and at worst, useless.

System Measurement Experiments are performed on the engineering system to validate the predicted system response. Both time-response and frequency-response measurements are made on the actual system or device. If the physical system is a concept in the design process, then experimental verification is accomplished by selective experimentation focusing on questionable modeling elements. Measurement Analysis The experimental data, both time-domain and frequency-domain, must be reviewed and analyzed for accuracy. Comparison: Mathematical Predications vs. Experimental Observations If the model predictions compare favorably to the experimental observations, then the model is adequate. If not, the experimental measurements, the analytical predictions, and the physical-model simplifying assumptions must be reviewed for accuracy. The physical model might need to be modified to capture the system characteristics which are important and were not initially included. Parasitic, or secondary, effects (e.g., saturation, nonlinear effects, time delays, hysteresis, Coulomb friction, and gear backlash) are added to the physical model to determine if each effect is significant, or if cumulatively they have adverse effects. Eventually, a truth model, which is as realistic a model one could develop, is used to validate system performance prior to hardware implementation. This often eliminates the need for hardware prototyping. The advantages over the build-and-test approach are staggering. Lets view below the comparison of the predicted response with the experimental response for the spring-mass system and the draw some conclusions.

16

The agreement between the predicted response and the actual measured response is quite good. The frequency of the oscillations, obtained by counting the number of cycles of the sine wave in a specified amount of time (here there about 15 cycles in 10 seconds or 1.5 cycles/sec) is the same and the oscillations occur about the zero position which corresponds to the static equilibrium position. The difference in amplitude between the two responses becomes more noticeable as the oscillations diminish. This is as expected because, from engineering experience, Coulomb friction will begin to dominate the response over viscous fluid friction when the system begins to slow down, and the physical model does not contain a Coulomb-friction term.

Design Changes If the model is adequate, but performance is inadequate, then design changes are in order, and the whole engineering system investigation process then starts over again for the revised engineering system.

17