Difference between revisions of "Frequency Domain Analysis"

From FBSwiki
Jump to: navigation, search
(created new chapter page)
 
(MATLAB code)
 
(32 intermediate revisions by the same user not shown)
Line 1: Line 1:
{{chheader|Chapter 7 - Transfer Functions|Chapter 8 - Loop Analysis|Chapter 8 - PID Control}}
+
{{chheader|Transfer Functions|Frequency Domain Analysis|PID Control}}
 
In this chapter we study how how stability and robustness of closed loop systems can be determined by investigating how signals propagate around the feedback loop. The Nyquist stability theorem is a key result that provides a way to analyze stability and introduce measures of degrees of stability.
 
In this chapter we study how how stability and robustness of closed loop systems can be determined by investigating how signals propagate around the feedback loop. The Nyquist stability theorem is a key result that provides a way to analyze stability and introduce measures of degrees of stability.
  
<table border=1 width=100%>
+
{{chaptertable begin}}
<tr valign=top>
+
{{chaptertable left}}
<td bgcolor=lightgreen width=50%>
+
 
== Textbook Contents ==
 
== Textbook Contents ==
{{am05pdf|am05-analysis|6nov05|Chapter 8 - Loop Analysis (6 Nov 05)|}}
+
{{am05pdf|am08-analysis|28Sep12|Frequency Domain Analysis|}}
 
* 1. Introduction
 
* 1. Introduction
 
* 2. The Nyquist Criterion
 
* 2. The Nyquist Criterion
 
* 3. Stability Margins
 
* 3. Stability Margins
 
* 4. Bode's Relations
 
* 4. Bode's Relations
* 5. The Small Gain Theorem
+
* 5. The Notion of Gain
 
* 6. Further Reading
 
* 6. Further Reading
 
* 7. Exercises
 
* 7. Exercises
<td bgcolor=lightblue width=50%>
+
 
 +
{{chaptertable right}}
  
 
== Lecture Materials ==
 
== Lecture Materials ==
 
* [[Lecture: Loop Analysis]]
 
* [[Lecture: Loop Analysis]]
* [[#Homework|Additional Homework Problems]]
+
* [[#Exercises|Additional Exercises]]
  
 
== Supplemental Information ==
 
== Supplemental Information ==
 
* [[#Frequently Asked Questions|Frequently Asked Questions]]
 
* [[#Frequently Asked Questions|Frequently Asked Questions]]
* Wikipedia entries:  
+
* [[#Errata|Errata]]
 +
* Wikipedia entries: [http://en.wikipedia.org/wiki/Nyquist_stability_criterion Nyquist stability criterion]
 
* [[#Additional Information|Additional Information]]
 
* [[#Additional Information|Additional Information]]
</table>
+
{{chaptertable end}}
  
== Homework ==
+
== Chapter Summary ==
<!-- Inline:Category:Chapter 8 HW -->
+
 
 +
This chapter describes the use of the Nyquist criterion for determining the stability of a system:
 +
<ol>
 +
<li><p>The ''loop transfer function'' of a feedback system represents the transfer function obtained by breaking the feedback loop and computing the resulting transfer function of the open loop system.  For a simple feedback system
 +
<center>[[Image:loopanal_fbksys.png]]</center>
 +
the loop transfer function is given by <math>L = P C</math>
 +
</p></li>
 +
 
 +
<li><p>The ''Nyquist criterion'' provides a way to check the stability of a closed loop system by looking at the properties of the loop transfer function.  For a stable open loop system, the Nyquist criterion states that the system is stable if the contour of the loop transfer function plotted from <math>s = -j\infty</math> to <math>s = j \infty</math> has no net encirclements of the point <math>s=-1</math> when it is plotted on the complex plane.</p></li>
 +
 
 +
<li><p>The general Nyquist criterion uses the image of the loop transfer function applied to the ''Nyquist countour''
 +
<center>[[Image:loopanal_nyqcontour.png]]</center>
 +
The number of unstable poles of the closed loop system is given by the number of open loop unstable poles plus the number of clockwise encirclements of the point <math>s = -1</math>.</p></li>
 +
 
 +
<li><p>Stability margins describe the robustness of a system to perturbations in the dynamics.  We define the ''phase crossover frequency'', <amsmath>\omega_180</amsmath> as the smallest frequency where the phase of the loop transfer function is <math>-180^\circ</math> and the ''gain crossover frequency'', <math>\omega_{gc}</math> as the small frequency where the loop transfer function has unit magnitude.  The ''gain margin'' and ''phase margin'' are given by
 +
<center><amsmath>
 +
  g_m=\frac{1}{|L(j\omega_{180})|} \qquad  \varphi_m=\pi+\arg{L(j\omega_{gc})}
 +
</amsmath></center>
 +
These margins describe the the maximum variation in gain and phase in the loop transfer function under which the system remains stable.  Two other margins are the ''stability margin'', which is the shortest distance frmo the Nyquist curve to the critical point <math>s=-1</math>, and the ''delay margin'', which is the smallest time delay required to make the system unstable.
 +
</p></li>
 +
 
 +
<li><p>''Bode's relations'' relate the gain and phase of a transfer function with no poles or zeros in the right half plane.  They show that
 +
<center><amsmath>
 +
  \arg{G(j\omega_0)}  \approx
 +
    \frac{\pi}{2} \frac{d\log{|G(j\omega)|}} {d\log{\omega}}.
 +
</amsmath></center>
 +
A ''non-minimum phase'' sytem is one for which there is more phase lag than the amount given by Bode's relations.  Systems with right have plane poles or zeros are non-minimum phase.
 +
</p></li>
 +
 
 +
<li><p>The ''gain'' of an input/output system is defined as
 +
<center><amsmath>
 +
  \gamma=\sup_{u\in\mathcal U}\frac{\|y\|}{\|u\|},
 +
</amsmath></center>
 +
where sup is the supremum.  The ''small gain theorem'' states that if two systems with gains <math>\gamma_1</math> and <math>\gamma_2</math> are connected in a feedback loop, then the closed loop system is stable if <math>\gamma_1\gamma_2</math>.
 +
</p></li>
 +
 
 +
</ol>
 +
 
 +
{{chaptertable begin}}
 +
{{chaptertable left}}
 +
 
 +
<!--
 +
== Additional Exercises ==
 +
The following exercises cover some of the topics introduced in this chapter.  Exercises marked with a * appear in the printed text.
 +
<ncl>Frequency Domain Analysis Exercises</ncl>
 +
-->
 
== Frequently Asked Questions ==
 
== Frequently Asked Questions ==
<!-- Inline:Category:Chapter 8 FAQ -->
+
<ncl>Frequency Domain Analysis FAQ</ncl>
 +
== Errata ==
 +
<ncl>Frequency Domain Analysis errata v2.11b</ncl>
 +
* [[:Category:Frequency Domain Analysis errata|Full list of errata starting from first printing]]
 +
* {{submitbug}}
 +
{{chaptertable right}}
 +
== MATLAB code ==
 +
The following MATLAB scripts are available for producing figures that appear in this chapter.
 +
* Figure 9.4: {{matlabfile|loopanal|nyquist_threepole.m}}
 +
* Figure 9.5: {{matlabfile|loopanal|sketch.m}}
 +
* Figure 9.6: {{matlabfile|loopanal|congctrl_nyquist.m}}, {{matlabfile|.|congctrl_params.m}}, {{matlabfile|.|congctrl_equil.m}}
 +
* Figure 9.7: {{matlabfile|loopanal|nyqcondstab.m}}
 +
* Figure 9.8: {{matlabfile|loopanal|invpend_nyquist.m}}
 +
* Figure 9.10: {{matlabfile|loopanal|simple.m}}
 +
* Figure 9.11: {{matlabfile|loopanal|niklas.m}}
 +
* Figure 9.12: {{matlabfile|loopanal|osc_integral.m}}
 +
* Figure 9.14: {{matlabfile|loopanal|steering_fwdrev.m}}
 +
* Figure 9.17: {{matlabfile|loopanal|desfcn_relay.m}}, {{matlabfile|loopanal|desfcn_hystsim.m}}, {{matlabfile|loopanal|relaysine.mdl}}
 +
See the [[software|software page]] for more information on how to run these scripts.
 +
 
 
== Additional Information ==
 
== Additional Information ==
 +
* [http://www.engin.umich.edu/group/ctm/freq/nyq.html Control Tutorials for Matlab, Nyquist Criterion]
 +
{{chaptertable end}}

Latest revision as of 23:03, 11 November 2012

Prev: Transfer Functions Chapter 9 - Frequency Domain Analysis Next: PID Control

In this chapter we study how how stability and robustness of closed loop systems can be determined by investigating how signals propagate around the feedback loop. The Nyquist stability theorem is a key result that provides a way to analyze stability and introduce measures of degrees of stability.

Textbook Contents

Frequency Domain Analysis (pdf, 28Sep12)

  • 1. Introduction
  • 2. The Nyquist Criterion
  • 3. Stability Margins
  • 4. Bode's Relations
  • 5. The Notion of Gain
  • 6. Further Reading
  • 7. Exercises

Lecture Materials

Supplemental Information

Chapter Summary

This chapter describes the use of the Nyquist criterion for determining the stability of a system:

  1. The loop transfer function of a feedback system represents the transfer function obtained by breaking the feedback loop and computing the resulting transfer function of the open loop system. For a simple feedback system

    Loopanal fbksys.png

    the loop transfer function is given by

  2. The Nyquist criterion provides a way to check the stability of a closed loop system by looking at the properties of the loop transfer function. For a stable open loop system, the Nyquist criterion states that the system is stable if the contour of the loop transfer function plotted from to has no net encirclements of the point when it is plotted on the complex plane.

  3. The general Nyquist criterion uses the image of the loop transfer function applied to the Nyquist countour

    Loopanal nyqcontour.png
    The number of unstable poles of the closed loop system is given by the number of open loop unstable poles plus the number of clockwise encirclements of the point .

  4. Stability margins describe the robustness of a system to perturbations in the dynamics. We define the phase crossover frequency, math as the smallest frequency where the phase of the loop transfer function is and the gain crossover frequency, as the small frequency where the loop transfer function has unit magnitude. The gain margin and phase margin are given by

    math

    These margins describe the the maximum variation in gain and phase in the loop transfer function under which the system remains stable. Two other margins are the stability margin, which is the shortest distance frmo the Nyquist curve to the critical point , and the delay margin, which is the smallest time delay required to make the system unstable.

  5. Bode's relations relate the gain and phase of a transfer function with no poles or zeros in the right half plane. They show that

    math

    A non-minimum phase sytem is one for which there is more phase lag than the amount given by Bode's relations. Systems with right have plane poles or zeros are non-minimum phase.

  6. The gain of an input/output system is defined as

    math

    where sup is the supremum. The small gain theorem states that if two systems with gains and are connected in a feedback loop, then the closed loop system is stable if .

Frequently Asked Questions

Errata

MATLAB code

The following MATLAB scripts are available for producing figures that appear in this chapter.

See the software page for more information on how to run these scripts.

Additional Information