Scribd
Upload
5 views

PID Using SF Controller

Uploaded by

lahlouh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
5 views

PID Using SF Controller

Uploaded by

lahlouh
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Proceedings of the IEEE SoutheastCon 2015, April 9 - 12, 2015 - Fort Lauderdale, Florida

Experiments with PID Controllers using State


Feedback Design Techniques
Robert J. Barsanti
Dept. of Electrical and Computer Engineering
The Citadel
Charleston, SC, USA
[email protected]

Abstract—This paper investigates the application of state


feedback (SF) design methods to implement a Proportional,
Integral, Derivative (PID) controller. A comparison is made with
the Ziegler-Nichols tuning method. Two design examples are
conducted to validate the procedure.

Keywords— Automation, Controllers, PID, State Space.

I. INTRODUCTION Fig.1. Closed loop PID controller and associated pre-filter.


The classical PID controller has gained popularity in part The three PID gains (k1, k2, k3) can be made identical to the
because its implementation does not require detailed three feedback gains of the third order SF system by proper
knowledge of the system to be controlled (the plant). Instead, selection of the SF model. In particular, the PID controller has
armed with observations of the plant response to known inputs the form
the three controller gains (proportional, integral, and
derivative) can be adjusted to achieve a desirable response. 
     . 1
Various rules for trimming the controller can be found in the 
literature. In practice the tuning rules attempt to achieve a While the pre-filter can be found to be [1]
specific response characteristic. For example, the Ziegler-
Nichols tuning method will provide a rapid but slightly 
. 2
oscillatory response to disturbances using quarter amplitude        
damping [4]. Conversely, the Lambda tuning rules provide a
robust but slower over damped response to disturbances [5].
Given the transfer function model for the plant G(s) the SF
Alternatively, if the plant transfer function parameters are design proceeds as follows:
known a-priori, then a PID controller can be employed by
computing the gains analytically. Then observation of the 1) Determine the desired closed loop system poles based on
system response to test inputs would not be required. the required process response characteristics. For example, the
Additionally, if state feedback design methods are used, the poles may be selected to reduce steady- state error, or lessen
overall system closed loop poles can be placed arbitrarily the response time, or to dampen disturbances. A variety of
allowing the designer to tailor the response to the needs of the approaches may be employed, e.g., root locus design [2], [3].
process. 2) Convert the known plant G(s) transfer function to a state
It is the goal of this short paper to compare SF design space model of the form
methods of [1] to the classical Ziegler- Nichols tuning method      . (3)
[4].
Where A and B are matrices of real coefficients, x is a vector of
the system states, and u is the system input. The A and B must
have the particular form described in [1].
II. PID DESIGN IN STATE SPACE
3) Compute the SF loop gains (k1, k2, k3) using Ackerman’s
State feedback (SF) controllers are based on a state space
formula [2] or similar approaches that can be found in [1, 2, 3].
model of the plant, and operate by creating a control input to
the plant formed from a linear combination of the plant states. 4) Form the PID controller using (1).
The SF controller gains can be computed using a variety of
techniques such as can be found in [1], [2], or [3]. 5) Form the pre-filter using (2).

The equivalence of SF and PID controllers is derived for a III. EXPERIMENT AND RESULTS
second order process in [1]. The resulting SF controller can be A comparison of the PID controller tuned using Ziegler-
represented by a PID controller with pre-filter as shown in Nichols [4], and with gains computed using the SF technique
Fig.1. was conducted via computer simulations using Matlab®

7300-5/15/$31.00 ©2015 IEEE


978-1-4673-7300-5/15/$31.00
978-1-4673-7300-5/15/$31.00 ©2015 IEEE
©2015 IEEE
Proceedings of the IEEE SoutheastCon 2015, April 9 - 12, 2015 - Fort Lauderdale, Florida

software. In case 1, a second order DC motor transfer function IV. FUTURE WORK
G(s) was used as a plant. First, the simulated step response of The SF pre-filter adds complexity beyond the PID
the original plant was obtained, and used to compute the PID controller. The next step is to investigate improving controller
controller gains via the Ziegler-Nichols (Z-N) method [4]. performance without pre-filtering.
Next, the PID controller gains were determined using the
design procedure of [1], as explained in the previous section.
The closed loop poles were chosen to produce a two second
settling time with minimal overshoot. The necessary SF gains
were computed using Ackerman’s formula. Plots of the step
response of the original system and two PID controlled systems
are displayed in Fig. 2.
The resulting response curves indicate similar performance
with both controllers closely achieving the theoretical
responses. Note that the Z-N method has a more oscillatory
response as expected. The benefit of using state space
techniques to compute PID controller gains is thus validated in
those cases where the plant transfer function is known.

Fig.3. Design results for case 2, unstable plant. Top: the step response
of the original plant. Middle: PID controller using SF design. Bottom:
PID controller using Ziegler-Nichols design

ACKNOWLEDGMENT
The author would like to acknowledge The Citadel
Foundation for the grant that supported this research.
REFERENCES
[1] R.J. Vaccaro, “Digital Control, A State Space Approach”, McGraw-
Hill, 1995.
[2] B.C. Kuo, Digital Control Sytems, Oxford University Press, 1992.
[3] K. Ogata, Discrete-Time Control Systems, Prentice Hall, 1995.
Fig.2. Design results for case 1, stable plant. Top: the step [4] J.G. Ziegler and N.B. Nichols, Optimum settings for automatic
response of the original plant. Middle: PID controller using SF design. controllers. Transactions of the ASME, 64, pp. 759–768, 1942.
Bottom: PID controller using Ziegler-Nichols design. [5] Rivera, D.E., M. Morari, and S. Skogestad, Internal Model Control 4.
In case 2, a second order unstable transfer function G(s) PID Controller Design, Industrial Engineering and Chemical Process
Design and Development, 25, p. 252, 1986
was used as a plant. Again, the closed loop poles were chosen
to produce a two second settling time with minimal overshoot.
Plots of the step response of the original system and two PID
controlled systems are displayed in Fig. 3. The SF design
procedure again performs well. The Z-N method however is
known to perform poorly for an unstable plant [5].

7300-5/15/$31.00978-1-4673-7300-5/15/$31.00
©2015 IEEE ©2015 IEEE

You might also like