Discover millions of ebooks, audiobooks, and so much more with a free trial

Only $11.99/month after trial. Cancel anytime.

Control System Design Guide: Using Your Computer to Understand and Diagnose Feedback Controllers
Control System Design Guide: Using Your Computer to Understand and Diagnose Feedback Controllers
Control System Design Guide: Using Your Computer to Understand and Diagnose Feedback Controllers
Ebook866 pages8 hours

Control System Design Guide: Using Your Computer to Understand and Diagnose Feedback Controllers

Rating: 3.5 out of 5 stars

3.5/5

()

Read preview

About this ebook

Control Systems Design Guide has helped thousands of engineers to improve machine performance. This fourth edition of the practical guide has been updated with cutting-edge control design scenarios, models and simulations enabling apps from battlebots to solar collectors. This useful reference enhances coverage of practical applications via the inclusion of new control system models, troubleshooting tips, and expanded coverage of complex systems requirements, such as increased speed, precision and remote capabilities, bridging the gap between the complex, math-heavy control theory taught in formal courses, and the efficient implementation required in real industry settings. George Ellis is Director of Technology Planning and Chief Engineer of Servo Systems at Kollmorgen Corporation, a leading provider of motion systems and components for original equipment manufacturers (OEMs) around the globe. He has designed an applied motion control systems professionally for over 30 years He has written two well-respected books with Academic Press, Observers in Control Systems and Control System Design Guide, now in its fourth edition. He has contributed articles on the application of controls to numerous magazines, including Machine Design, Control Engineering, Motion Systems Design, Power Control and Intelligent Motion, and Electronic Design News.
  • Explains how to model machines and processes, including how to measure working equipment, with an intuitive approach that avoids complex math
  • Includes coverage on the interface between control systems and digital processors, reflecting the reality that most motion systems are now designed with PC software
  • Of particular interest to the practicing engineer is the addition of new material on real-time, remote and networked control systems
  • Teaches how control systems work at an intuitive level, including how to measure, model, and diagnose problems, all without the unnecessary math so common in this field
  • Principles are taught in plain language and then demonstrated with dozens of software models so the reader fully comprehend the material (The models and software to replicate all material in the book is provided without charge by the author at www.QxDesign.com)
  • New material includes practical uses of Rapid Control Prototypes (RCP) including extensive examples using National Instruments LabVIEW
LanguageEnglish
Release dateJun 27, 2012
ISBN9780123859211
Control System Design Guide: Using Your Computer to Understand and Diagnose Feedback Controllers
Author

George Ellis

George Ellis traveled extensively in Asia in the 1970s and 1980s, writing and translating texts on Yoga, Ayurveda, and Naturopathy. He is the author of The Breath of Life: Mastering the Techniques of Pranayama and Qi Gong. George Ellis and Zhuo Zhao met in Beijing in 1985 and married in 1988.

Read more from George Ellis

Related authors

Related to Control System Design Guide

Related ebooks

Mechanical Engineering For You

View More

Related articles

Reviews for Control System Design Guide

Rating: 3.6666666666666665 out of 5 stars
3.5/5

3 ratings0 reviews

What did you think?

Tap to rate

Review must be at least 10 words

    Book preview

    Control System Design Guide - George Ellis

    safely!

    Chapter 1

    Introduction to Controls

    Chapter Outline

    1.1 Visual ModelQ Simulation Environment

    1.1.1 Installation of Visual ModelQ

    1.1.2 Errata

    1.2 The Control System

    1.2.1 The Controller

    1.2.2 The Machine

    1.3 The Controls Engineer

    Control theory is used for analysis and design of feedback systems, such as those that regulate temperature, fluid flow, motion, force, voltage, pressure, tension, and current. Skillfully used, control theory can guide engineers in every phase of the product and process design cycle. It can help engineers predict performance, anticipate problems, and provide solutions.

    Colleges teach controls with little emphasis on day-to-day problems. The academic community focuses on mathematical derivations and on the development of advanced control schemes; it often neglects the methods that are commonly applied in industry. Students can complete engineering programs that include courses on controls and still remain untutored on how to design, model, build, tune, and troubleshoot a basic control system. The unfortunate result is that many working engineers lay aside analysis when they practice their profession, relying instead on company history and trial-and-error methods.

    This book avoids the material and organization of most control theory textbooks. For example, design guidelines are presented throughout; these guidelines are a combination of industry-accepted practices and warnings against common pitfalls. Nontraditional subjects, such as filters and modeling, are presented here because they are essential to understanding and implementing control systems in the workplace. The focus of each chapter is to teach how to use controls to improve a working machine or process.

    The wide availability of personal computers and workstations is an important advance for control system designers. Many of the classical control methods, such as the root locus method, are graphical rather than analytical. Their creators sought to avoid what was then the overwhelming number of computations required for analytical methods. Fortunately, these calculations no longer present a barrier. Virtually every personal computer can execute the calculations required by analytical methods. With this in mind, the principles and methods presented herein are essentially analytical, and the arithmetic is meant to be carried out by a computer.

    1.1 Visual ModelQ Simulation Environment

    Most engineers understand the foundations of control theory. Concepts such as transfer functions, block diagrams, the s-domain, and Bode plots are familiar to most of us. But how should working engineers apply these concepts? As in most disciplines, they must develop intuition, and this requires fluency in the basics. In order to be fluent, you must practice.

    When studying control system techniques, finding equipment to practice on is often difficult. As a result, designers often rely on computer simulations. To this end, the author developed, as a companion to this book, Visual ModelQ, a stand-alone, graphical, PC-based simulation environment. The environment provides time-domain and frequency-domain analysis of analog and digital control systems. Dozens of Visual ModelQ models were developed for this book. These models are used extensively in the chapters that follow. Readers can run these experiments to verify results and then modify parameters and other conditions to experiment with the concepts of control systems.

    Visual ModelQ is written to teach control theory. It makes convenient those activities that are necessary for studying controls. Control law gains are easy to change. Plots of frequency-domain response (Bode plots) are run with the press of a button. The models in Visual ModelQ run continuously, just as real-time controllers do. The measurement equipment runs independently, so you can change parameters and see the effects immediately.

    1.1.1 Installation of Visual ModelQ

    Visual ModelQ is available at www.qxdesign.com. The unregistered version is available free of charge. The unregistered version can execute all the models used in this book. Readers may elect to register their copies of Visual ModelQ at any time; see www.qxdesign.com for details.

    Visual ModelQ runs on PCs using Windows XP and Windows 7. Download the installation package for Visual ModelQ V7.0 or later; the files are typically installed to C:\Program Files\QxDesign, Inc\VisualModelQ7.0. Visual ModelQ installs with both a User’s Manual and a Reference Manual. After installation, read the User’s Manual. Note that you can access the Reference Manual by pressing the F1 key. Finally, check www.qxdesign.com from time to time for updated software.

    1.1.2 Errata

    Check www.qxdesign.com for errata. It is the author’s intention to regularly update the Web page as corrections become known.

    1.2 The Control System

    The general control system, as shown in Figure 1.1, can be divided into the controller and the machine. The controller can be divided into the control laws and the power converter. The machine may be a temperature bath, a motor, or, as in the case of a power supply, an inductor/capacitor circuit. The machine can also be divided into two parts: the plant and the feedback device (s). The plant receives two types of signals: a controller output from the power converter and one or more disturbances. Simply put, the goal of the control system is to drive the plant in response to the command while overcoming disturbances.

    Figure 1.1 The general control system.

    1.2.1 The Controller

    The controller incorporates both control laws and power conversion. Control laws, such as proportional-integral-differential (PID) control, are familiar to control engineers. The process of tuning — setting gains to attain desired performance — amounts to adjusting the parameters of the control laws. Most controllers let designers adjust gains; the most flexible controllers allow the designer to modify the control laws themselves. When tuning, most control engineers focus on attaining a quick, stable command response. However, in some applications, rejecting disturbances is more important than responding to commands. All control systems should demonstrate robust performance because even nearly identical machines and processes vary somewhat from one to the other, and they change over time. Robust operation means control laws must be designed with enough margin to accommodate reasonable changes in the plant and power converter.

    Virtually all controllers have power converters. The control laws produce information, but power must be applied to control the plant. The power converter can be driven by any available power source, including electric, pneumatic, hydraulic, or chemical power.

    1.2.2 The Machine

    The machine is made of two parts: the plant and the feedback. The plant is the element or elements that produce the system response. Plants are generally passive, and they usually dissipate power. Examples of plants include a heating element and a motor coupled to its load.

    Control systems need feedback because the plant is rarely predictable enough to be controlled open loop — that is, without feedback. This is because most plants integrate the power converter output to produce the system response. Voltage is applied to inductors to produce current; torque is applied to inertia to produce velocity; pressure is applied to produce fluid flow. In all these cases, the control system cannot control the output variable directly but must provide power to the machine as physics allows and then monitor the feedback to ensure that the plant is on track.

    1.3 The Controls Engineer

    The focal task of many controls engineers is system integration and commissioning. The most familiar part of this process is tuning the control loops. This process can be intimidating. Often dozens of parameters must be fine-tuned to ensure that the system meets the specification. Sometimes that specification is entirely formal, but more often it is a combination of formal requirements and know-how gained with years of experience. Usually, experienced engineers are required to judge when a system is performing well enough to meet the needs of the application.

    For some control systems, each installation may require days or weeks to be correctly commissioned. In a complex machine such as a rolling mill, that process can take months. Each section of the machine must be carefully tuned at the site. So even after the design of the machine is complete, the expertise of a controls engineer is required each time a unit is installed.

    Although most controls engineers focus on installation, their job should begin when the machine is designed. Many companies fail to take advantage of their controls experts early in a project; this is short-sighted. A controls engineer may suggest an improved feedback device or enhancements to a machine that will help overcome a stubborn problem. Ideally, the project manager will solicit this input early, because changes of this nature are often difficult to make later.

    The controls engineer should also contribute to the selection of the controller. There are many controls-oriented factors that should be taken into account. Does the controller implement familiar control laws? For digital controllers, is the processor fast enough for the needs of the application? Is the complexity/ease-of-use appropriate for the support team and for the customer base? The selection and specification of a controller involves input from many perspectives, but some questions can be answered best by a skilled controls engineer.

    What is the role for control theory in the daily tasks of controls engineers? At its root, control theory provides understanding and, with that, intuition. Should the company purchase the controller that runs four times faster even though it costs 25% more? Should they invest in machine changes, and what will be the expected improvement from those efforts? How much will the better feedback device help a noise problem? Understanding controls doesn’t guarantee that the engineer will have the correct answer. But a firm grasp on the practical side of controls will provide correct answers more often and thus position the controls engineer to provide leadership in process and product development and support.

    Chapter 2

    The Frequency Domain

    Chapter Outline

    2.1 The Laplace Transform

    2.2 Transfer Functions

    2.2.1 What is s?

    2.2.1.1 DC Gain

    2.2.2 Linearity, Time Invariance, and Transfer Functions

    2.3 Examples of Transfer Functions

    2.3.1 Transfer Functions of Controller Elements

    2.3.1.1 Integration and Differentiation

    2.3.1.2 Filters

    2.3.1.3 Control Laws

    2.3.1.4 Compensators

    2.3.1.5 Delays

    2.3.2 Transfer Functions of Power Conversion

    2.3.3 Transfer Functions of Physical Elements

    2.3.4 Transfer Functions of Feedback

    2.4 Block Diagrams

    2.4.1 Combining Blocks

    2.4.1.1 Simplifying a Feedback Loop

    2.4.2 Mason’s Signal Flow Graphs

    2.4.2.1 Step-by-Step Procedure

    2.5 Phase and Gain

    2.5.1 Phase and Gain from Transfer Functions

    2.5.2 Bode Plots: Phase and Gain versus Frequency

    2.6 Measuring Performance

    2.6.1 Command Response

    2.6.2 Stability

    2.6.3 Time Domain versus Frequency Domain

    2.7 Questions

    The frequency domain provides intuition on practical subjects traversing the field of control theory. How can responsiveness be quantified? How stable is a system and how stable should it be? How do tuning gains work? How well does a system reject disturbances? The frequency domain is the beginning of control theory.

    The competitor to the frequency domain is the time domain. The time domain is, above all, convenient. It is easy to comprehend and easy to measure. The measurements of an oscilloscope are immediately understood. The time domain is often the best way to communicate with customers and colleagues. Thus, the controls engineer should be fluent in both time and frequency. They are two corroborating witnesses, furnishing together a clearer understanding than either can alone.

    This chapter will present the frequency domain, beginning with its foundation, the Laplace transform. Transfer functions are presented with examples of common control elements and plants. Bode plots, the favored graphic display in this text, are presented next. The chapter will discuss two important measures of control system performance — stability and response — using a Visual ModelQ experiment.

    2.1 The Laplace Transform

    The Laplace transform underpins classic control theory.³²,³³,⁸⁵ It is almost universally used. An engineer who describes a two-pole filter relies on the Laplace transform; the two poles are functions of s, the Laplace operator. The Laplace transform is defined in Equation 2.1.

    (2.1)

    The function f(t) is a function of time, s is the Laplace operator, and F(s) is the transformed function. The terms F(s) and f(t), commonly known as a transform pair, represent the same function in the two domains. For example, if f(t)  =  sin (ωt), then F(s)  =  ω/(ω² + s²). You can use the Laplace transform to move between the time and frequency domains.

    The Laplace transform can be intimidating. The execution of Equation 2.1 is complex in all but the simplest cases. Fortunately, the controls engineer need invest little time in such exercises. The most important benefit of the Laplace transform is that it provides s, the Laplace operator, and through that the frequency-domain transfer function.

    2.2 Transfer Functions

    Frequency-domain transfer functions describe the relationship between two signals as a function of s. For example, consider an integrator as a function of time (Figure 2.1). The integrator has an s-domain transfer function of 1/s (see Table 2.1). So it can be said that

    (2.2)

    Figure 2.1 Integrator.

    Table 2.1. Transfer Functions of Controller Elements

    Similarly, a derivative has the transfer function s; differentiating a time-domain signal is the same as multiplying a frequency-domain signal by s. Herein lies the usefulness of the Laplace transform. Complex time-domain operations such as differentiation and integration can be handled with algebra. Dealing with transfer functions in the time domain (that is, without the benefit of the Laplace transform) requires convolution, a mathematical process that is so complicated that it can be used only on the simplest systems.

    2.2.1 What is s?

    The Laplace operator is a complex (as opposed to real or imaginary) variable. It is defined as

    (2.3)

    The constant j The ω term translates to a sinusoid in the time domain; σ translates to an exponential (eσt ) term. Our primary concern will be with steady-state sinusoidal signals, in which case σ = 0. The frequency in hertz, f, is defined as f ≡ ω/2π. So for most of this book, Equation 2.3 will simplify to Equation 2.4:

    (2.4)

    The practical side of Equation 2.4 is that the steady-state response of an s-domain transfer function to a sine wave can be evaluated by setting s = j2πf.

    2.2.1.1 DC Gain

    Often it is important to evaluate the DC response of a transfer function, in other words, to determine the output of the transfer function subjected to a DC input. To find the DC response, set s to zero. For example, we discussed before that the transfer function for differentiation is Vo(s)  = Vi(s) × s. What happens when a DC signal is differentiated? Intuitively, it produces zero, and that is confirmed by setting s to zero in this equation.

    2.2.2 Linearity, Time Invariance, and Transfer Functions

    A frequency-domain transfer function is limited to describing elements that are linear and time invariant. These are severe restrictions and, in fact, virtually no real-world system fully meets them. The criteria that follow define these attributes, the first two being for linearity and the third for time invariance.

    1. Homogeneity. Assume that an input to a system r(t) generates an output c(t). For an element to be homogeneous, an input k × r(t) would have to generate an output k × c(t), for any value of k. An example of nonhomogeneous behavior is saturation, where twice as much input delivers less than twice as much output.

    2. Superposition. Assume that an element subjected to an input r1(t) will generate the output c1(t). Further, assume that the same element subjected to input r2(t) will generate an output c2(t). Superposition requires that if the element is subjected to the input r1(t) + r2(t), it will produce the output c1(t) + c2(t).³²,⁸⁰

    3. Time invariance. Assume that an element has an input r(t) that generates an output c(t). Time invariance requires that r(t − τ) will generate c(t − τ) for all τ > 0.

    So we face a dilemma: Transfer functions, the basis of classic control theory, require linear, time-invariant (LTI) systems, but no real-world system is completely LTI. This is a complex problem that is dealt with in many ways, some of which are detailed in Chapter 12. However, for most control systems, the primary solution is to design components close enough to being LTI that the non-LTI behavior can be ignored or avoided.

    In practice, most control systems are designed to minimize non-LTI behavior. This is one reason why components used in control systems are often more expensive than their noncontrol counterparts. For most of this text, the assumption will be that the system is LTI or close enough to it to use transfer functions. Readers who are troubled by this approximation should consider that this technique is commonly applied by engineers in all disciplines. For example, Ohm’s law, ν = iR, is an approximation that ignores many effects, including electrical radiation, capacitive coupling, and lead inductance. Of course, all those effects are important from time to time, but few would argue the utility of Ohm’s law.

    2.3 Examples of Transfer Functions

    In this section, we will discuss the transfer functions of common elements in control systems. The discussion is divided along the lines of Figure 2.2.

    Figure 2.2 Elements in the control system.

    2.3.1 Transfer Functions of Controller Elements

    The controller elements divide into the control laws and power conversion. Examples of operations used in control laws are shown in Table 2.1. Note that Appendix A shows the implementation of many of these laws using operational amplifier (op-amp) circuits.

    2.3.1.1 Integration and Differentiation

    Integration and differentiation are the simplest operations. The s-domain operation of integration is 1/s and of differentiation is s.

    2.3.1.2 Filters

    Filters are commonly used by control systems designers, such as when low-pass filters are added to reduce noise. The use of filters in control systems is detailed in Chapter 9. Table 2.1 lists the s-domain representation for a few common examples. Notice that second-order polynomials in Table 2.1 are shown in the damping (ζ) form: s² + 2ζωs + ω². Alternatively, they can be listed in the Q form: s² + ωs/Q + ω², where Q, the Quality factor, is defined as Q ≡ 1/(2ζ).

    2.3.1.3 Control Laws

    Control laws produce the control output primarily from the error signal that the control law is intended to minimize; the control output is the command to the power converter in Figure 2.2. Table 2.1 lists the most common control laws: P, PI, and PID, all of which are discussed in Chapter 6. Feed-forward terms are often added to control laws (see Chapter 8).

    2.3.1.4 Compensators

    Compensators are specialized filters. A compensator is a filter that is designed to provide a specific gain and phase shift, usually at one frequency. The effects on gain and phase either above or below that frequency are secondary. Table 2.1 shows a lag and lead compensators. The principles of compensators are described in Chapters 3, 4, and 6.

    2.3.1.5 Delays

    Delays add time lag without changing amplitude. For example, a conveyor belt with a scale can cause a delay. Material is loaded at one point and weighed a short distance later; that causes a time delay which is proportional to that distance and inversely proportional to the belt speed. A system that controls the amount of material to be loaded onto the conveyor must take the time delay into account. Since microprocessors have inherent delays for calculation time, the delay function is especially important to understanding digital controls, as will be discussed in Chapters 4 and 5.

    A delay of T seconds is shown in the time domain as

    (2.5)

    The corresponding function in the frequency domain is

    (2.6)

    2.3.2 Transfer Functions of Power Conversion

    Power conversion is the interface between the control laws and the plant. Power can come from many sources including electric, pneumatic, hydraulic, and chemical. The transfer function of the conversion depends on the device delivering the power. In electronic systems, the power is often delivered through pulse modulation — a family of near-linear methods that switch transistors at high frequency. Seen from the frequency domain, modulation inserts a short delay in the control loop. Modulation also injects nonlinear effects, such as ripple, which are usually ignored in the frequency domain.

    2.3.3 Transfer Functions of Physical Elements

    Physical elements are made up of the plant— the mechanism or device being controlled— and the feedback sensor(s). Examples of plants include electrical elements, such as inductors, and mechanical elements, such as springs and inertias. Table 2.2 provides a list of ideal elements in five categories.³² Consider the first category, electrical. A voltage force is applied to an impedance to produce a current flow. There are three types of impedance: inductance, capacitance, and resistance. Resistance is proportional to the current flow, inductive impedance is in proportion to the derivative of current flow, and capacitive impedance is in proportion to the integral of flow.

    Table 2.2. Transfer Functions of Plant Elements

    The pattern of force, impedance, and flow is repeated for many physical elements. In Table 2.2, the close parallels between the categories of linear and rotational force, fluid mechanics, and heat flow are evident. In each case, a forcing function (voltage, force, torque, pressure or temperature difference) applied to an impedance produces a flow (current, velocity, or fluid/thermal flow). The impedance takes three forms: resistance to the integral of flow (capacitance or mass), resistance to the derivative of flow (spring or inductance), and resistance to the flow rate (resistance or damping).

    Table 2.2 reveals a central concept of controls. Controllers for these elements apply a force to control a flow. For example, a heating system applies heat to control a room’s temperature. When the flow must be controlled with accuracy, a feedback sensor can be added to measure the flow; control laws are required to combine the feedback and command signals to generate the force. This results in the structure shown in Figure 2.2; it is this structure that sets control systems apart from other disciplines of engineering.

    The plant usually contains the most nonlinear elements in the control system. The plant is often too large and expensive to allow engineers the luxury of designing it for LTI operation. For example, an iron-core inductor in a power supply usually saturates; that is, its inductance declines as the current approaches its peak. The inductor could be designed so that saturation during normal operation is eliminated, but this requires more iron, which makes the power supply larger and more expensive.

    Most of the impedances in Table 2.2 can be expected to vary over normal operating conditions. The question naturally arises as to how much they can vary. The answer depends on the application, but most control systems are designed with enough margin that variation of 20–30% in the plant should have little impact on system performance. However, plant parameters sometimes vary by much more than 30%. For example, consider the inertia in rotational motion. An application that winds plastic onto a roll may see the inertia vary by orders of magnitude, an amount too large for ordinary control laws to accommodate simply by adding margin. We will deal with variations in gain of this magnitude in Chapter

    Enjoying the preview?
    Page 1 of 1