ISA-TR75.25.02-2000 (R2010)

ANSI Technical Report prepared by ISA

ANSI/ISA-TR75.25.02-2000 (R2010)

Control Valve Response Measurement from Step Inputs

Approved 18 July 2010

Copyright International Society of Automation

Provided by IHS under license with ISA

No reproduction or networking permitted without license from IHS

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---

Licensee=Sabic Engineering and Project Mgmt/5951674001, User=Soliman, Mohamed

Not for Resale, 01/03/2017 21:38:37 MST

ANSI/ISA-TR75.25.02-2000 (R2010)

Control Valves Response Measurement from Step Inputs

ISBN: 978-1-936007-50-9

Copyright

©

2010 by ISA. All rights reserved. Not for resale. Printed in the United States of America.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise), without the prior written permission of the Publisher.

ISA

67 Alexander Drive

P. O. Box 12277

Research Triangle Park, North Carolina 27709






- 3 - ANSI/ISA-TR75.25.02-2000 (R2010)

Preface

This preface, as well as all footnotes and annexes, is included for information purposes and is not part of

ANSI/ISA-TR75.25.02–2000 (R2010).

The standards referenced within this document may contain provisions which, through reference in this text, constitute requirements of this document. At the time of publication, the editions indicated were valid. All standards are subject to revision, and parties to agreements based on this document are encouraged to investigate the possibility of applying the most recent editions of the standards indicated within this document. Members of IEC and ISO maintain registers of currently valid International

Standards. ANSI maintains registers of currently valid U.S. National Standards.

This document has been prepared as part of the service of ISA toward a goal of uniformity in the field of instrumentation. To be of real value, this document should not be static but should be subject to periodic review. Toward this end, the Society welcomes all comments and criticisms and asks that they be addressed to the Secretary, Standards and Practices Board; ISA; 67 Alexander Drive; P. O. Box 12277;

Research Triangle Park, NC 27709; Telephone (919) 549-8411; Fax (919) 549-8288; E-mail: standards@isa.org.

The ISA Standards and Practices Department is aware of the growing need for attention to the metric system of units in general, and the International System of Units (SI) in particular, in the preparation of instrumentation standards. The Department is further aware of the benefits to USA users of ISA standards of incorporating suitable references to the SI (and the metric system) in their business and professional dealings with other countries. Toward this end, this Department will endeavor to introduce

SI-acceptable metric units in all new and revised standards, recommended practices, and technical reports to the greatest extent possible. Standard for Use of the International System of Units (SI): The

Modern Metric System , published by the American Society for Testing & Materials as IEEE/ASTM SI 10-

97, and future revisions, will be the reference guide for definitions, symbols, abbreviations, and conversion factors.

It is the policy of ISA to encourage and welcome the participation of all concerned individuals and interests in the development of ISA standards, recommended practices, and technical reports.

Participation in the ISA standards-making process by an individual in no way constitutes endorsement by the employer of that individual, of ISA, or of any of the standards, recommended practices, and technical reports that ISA develops.

CAUTION — ISA DOES NOT TAKE ANY POSITION WITH RESPECT TO THE EXISTENCE OR

VALIDITY OF ANY PATENT RIGHTS ASSERTED IN CONNECTION WITH THIS DOCUMENT, AND

ISA DISCLAIMS LIABILITY FOR THE INFRINGEMENT OF ANY PATENT RESULTING FROM THE

USE OF THIS DOCUMENT. USERS ARE ADVISED THAT DETERMINATION OF THE VALIDITY OF

ANY PATENT RIGHTS, AND THE RISK OF INFRINGEMENT OF SUCH RIGHTS, IS ENTIRELY THEIR

OWN RESPONSIBILITY.

PURSUANT TO ISA’S PATENT POLICY, ONE OR MORE PATENT HOLDERS OR PATENT

APPLICANTS MAY HAVE DISCLOSED PATENTS THAT COULD BE INFRINGED BY USE OF THIS

DOCUMENT AND EXECUTED A LETTER OF ASSURANCE COMMITTING TO THE GRANTING OF A

LICENSE ON A WORLDWIDE, NON-DISCRIMINATORY BASIS, WITH A FAIR AND REASONABLE

ROYALTY RATE AND FAIR AND REASONABLE TERMS AND CONDITIONS. FOR MORE

INFORMATION ON SUCH DISCLOSURES AND LETTERS OF ASSURANCE, CONTACT ISA OR

VISIT WWW.ISA.ORG/STANDARDSPATENTS.

OTHER PATENTS OR PATENT CLAIMS MAY EXIST FOR WHICH A DISCLOSURE OR LETTER OF

ASSURANCE HAS NOT BEEN RECEIVED. ISA IS NOT RESPONSIBLE FOR IDENTIFYING

PATENTS OR PATENT APPLICATIONS FOR WHICH A LICENSE MAY BE REQUIRED, FOR






ANSI/ISA-TR75.25.02-2000 (R2010) - 4 -

CONDUCTING INQUIRIES INTO THE LEGAL VALIDITY OR SCOPE OF PATENTS, OR

DETERMINING WHETHER ANY LICENSING TERMS OR CONDITIONS PROVIDED IN CONNECTION

WITH SUBMISSION OF A LETTER OF ASSURANCE, IF ANY, OR IN ANY LICENSING

AGREEMENTS ARE REASONABLE OR NON-DISCRIMINATORY.

ISA REQUESTS THAT ANYONE REVIEWING THIS DOCUMENT WHO IS AWARE OF ANY PATENTS

THAT MAY IMPACT IMPLEMENTATION OF THE DOCUMENT NOTIFY THE ISA STANDARDS AND

PRACTICES DEPARTMENT OF THE PATENT AND ITS OWNER.

ADDITIONALLY, THE USE OF THIS STANDARD MAY INVOLVE HAZARDOUS MATERIALS,

OPERATIONS OR EQUIPMENT. THE STANDARD CANNOT ANTICIPATE ALL POSSIBLE

APPLICATIONS OR ADDRESS ALL POSSIBLE SAFETY ISSUES ASSOCIATED WITH USE IN

HAZARDOUS CONDITIONS. THE USER OF THIS STANDARD MUST EXERCISE SOUND

PROFESSIONAL JUDGMENT CONCERNING ITS USE AND APPLICABILITY UNDER THE USER’S

PARTICULAR CIRCUMSTANCES. THE USER MUST ALSO CONSIDER THE APPLICABILITY OF

ANY GOVERNMENTAL REGULATORY LIMITATIONS AND ESTABLISHED SAFETY AND HEALTH

PRACTICES BEFORE IMPLEMENTING THIS STANDARD.

THE USER OF THIS DOCUMENT SHOULD BE AWARE THAT THIS DOCUMENT MAY BE

IMPACTED BY ELECTRONIC SECURITY ISSUES. THE COMMITTEE HAS NOT YET ADDRESSED

THE POTENTIAL ISSUES IN THIS VERSION.

The following people served as members of ISA Subcommittee ISA75.25 and approved

ANSI/ISA-TR75.25.02-2000:

NAME COMPANY

C. Langford, Chairman

W. Weidman, Managing Director

J. Beall

D. Bennett

W. Bialkowski

W. Black

M. Boudreaux

S. Boyle

D. Buchanan

F. Cain

N. Cammy

M. Coughran

Cullen G. Langford, Inc.

Parsons Energy & Chemicals Group

Eastman Chemical Co.

Samson Controls, Inc.

Entech Control Engineering, Inc.

Cashco, Inc.

Exxon Mobil Chemical

Neles Automation

Union Carbide Corporation

Flowserve Corporation

UOP LLC

Fisher Controls

J. Jamison

S. Kempf

P. Maurath

R. McEver

N. McLeod

G. McMillan

J. Reed

K. Senior

Bantrel Inc.

Harold Beck & Sons, Inc.

Procter & Gamble Company

Bettis Corporation

Elf Atochem

Solutia, Inc.

Norriseal

Dupont Dow Elastomers

The following people served as members of ISA Committee ISA75 and approved

ANSI/ISA-TR75.25.02-2000:

NAME COMPANY

D. Buchanan, Chairman


Union Carbide Corporation







- 5 - ANSI/ISA-TR75.25.02-2000 (R2010)

A. Abromaitis

H. Backinger

G. Barb

H. Baumann

H. Boger

G. Borden

S. Boyle

R. Brodin

F. Cain

C. Corson

A. Engels

H. Fuller

J. George

A. Glenn

L. Griffith

B. Guinon

F. Harthun

B. Hatton

J. Jamison

R. Jeanes

J. Kersh

C. Langford

A. Libke

R. Louviere

O. Lovett

A. McCauley

R. McEver

H. Miller

T. Molloy

L. Ormanoski

J. Ozol

W. Rahmeyer

J. Reed

K. Schoonover

A. Shea

E. Skovgaard

H. Sonderegger

R. Terhune

Red Valve Company, Inc.

J. F. Kraus & Company

Retired

H B Services Partners LLC

Masoneilan Dresser

Consultant

Neles Automation

Fisher Controls International, Inc.


Fluor Daniel Inc.

Praxair, Inc.

Valvcon Corporation

Richards Industries

Flowserve Corp.

Consultant/Retired

Shell Chemical

Retired

DeZurik Division Unit

Bantrel, Inc.

TXU Electric

M. W. Kellogg Company

Cullen G. Langford, Inc.

DeZurik Valve Company

Creole Engineering Sales Company

Consultant/Retired

Chagrin Valley Controls, Inc.

Bettis Corporation

Control Components, Inc.

CMES Inc.

Frick Company

Commonwealth Edison

Utah State University

Norriseal

Con-Tek Valves, Inc.

Copes-Vulcan, Inc.

Leslie Controls, Inc.

Tyco Flow Control

Retired

ANSI/ISA-TR75.25.02-2000 was approved for publication by the ISA Standards and Practices Board on

30 December 2000:

NAME COMPANY

M. Zielinski, Vice President

D. Bishop

P. Brett

M. Cohen

M. Coppler

B. Dumortier

W. Holland

A. Iverson

R. Jones

V. Maggioli

T. McAvinew

Fisher-Rosemount Systems, Inc.

Consultant

Honeywell, Inc.

Senior Flexonics, Inc.

Ametek, Inc.

Schneider Electric

Southern Company

Ivy Optiks

Dow Chemical Co.

Feltronics Corp.

Bateman Engineering, Inc.






ANSI/ISA-TR75.25.02-2000 (R2010) - 6 -

A. McCauley, Jr.

G. McFarland

D. Rapley

R. Reimer

J. Rennie

H. Sasajima

R. Webb

W. Weidman

J. Weiss

J. Whetstone

M. Widmeyer

R. Wiegle

Chagrin Valley Controls, Inc.

Westinghouse Process Control Inc.

Rapley Consulting Inc.

Rockwell Automation

Factory Mutual Research Corp.

Yamatake Corp.

Altran Corp.


EPRI

National Institute of Standards & Technology

EG&G Defense Materials

CANUS Corp.

C. Williams

G. Wood

Eastman Kodak Co.

Graeme Wood Consulting

The following people served as members of ISA Subcommittee ISA75.25 and reaffirmed

ANSI/ISA-TR75.25.02-2000 (R2010):

NAME COMPANY

J. Beall, Chairman


D. Bennett

W. Bialkowski

M. Boudreaux

S. Boyle

N. Cammy

J. Faramarzi

J. Jamison

S. Kempf

J. Kiesbauer

Emerson Process Management

Worley Parsons

Samson Controls Inc.

Entech Control Engineering Inc.

ExxonMobil Chemical

Metso Automation USA Inc.

UOP LLC

Control Components Inc.

EnCana Corporation Ltd.

Harold Beck & Sons Inc.

Samson Aktiengesellschaft

C. Langford

P. Maurath

R. McEver

N. McLeod

G. McMillan

J. Reed

J. Young

Consultant

Procter & Gamble Company

Consultant

Arkema

CDI – Process & Industrial

Consultant

The Dow Chemical Company

The following people served as members of ISA Committee ISA75 and reaffirmed

ANSI/ISA-TR75.25.02-2000 (R2010):

NAME COMPANY

J. Young, Chairman


H. Baumann

J. Beall

M. Bober

H. Boger

G. Borden

S. Boyle

J. Broyles

F. Cain

W. Cohen

R. Duimstra

J. Faramarzi


Worley Parsons

Consultant


Copes-Vulcan

Masoneilan Dresser

Consultant

Metso Automation USA Inc.

Enbridge Pipelines Inc.


KBR

Fisher Controls International Inc.

Control Components Inc.






- 7 - ANSI/ISA-TR75.25.02-2000 (R2010)

J. George

J. Jamison

J. Kiesbauer

A. Libke

G. Liu

H. Maxwell

J. McCaskill

A. McCauley

R. McEver

V. Mezzano

H. Miller

T. Molloy

Richards Industries


Samson Aktiengesellschaft

DeZurik

Consultant

Bechtel Power Corp.

Expro Group

Chagrin Valley Controls Inc.

Consultant

Fluor Corporation

Consultant

CMES Inc.

L. Ormanoski

J. Reed

E. Skovgaard

Consultant

Consultant

Control Valve Solutions

This standard was approved for reaffirmation by the ISA Standards and Practices Board on

3 June 2010:

J. Tatera

P. Brett

M. Coppler

E. Cosman

B. Dumortier

D. Dunn

R. Dunn

J. Gilsinn

E. Icayan

J. Jamison

D. Kaufman

K. P. Lindner

V. Maggioli

T. McAvinew

A. McCauley

G. McFarland

R. Reimer

N. Sands

H. Sasajima

T. Schnaare

I. Verhappen

R. Webb

W. Weidman

J. Weiss

M. Widmeyer

M. Wilkins

M. Zielinski

NAME COMPANY

Tatera & Associates Inc.

Honeywell Inc.

Ametek Inc.


Schneider Electric

Aramco Services Co.

DuPont Engineering

NIST/MEL

ACES Inc.


Honeywell International Inc.

Endress + Hauser Process Solutions AG

Feltronics Corp.

Jacobs Engineering

Chagrin Valley Controls Inc.

Emerson Process Mgmt. Power & Water Sol.

Rockwell Automation

DuPont

Yamatake Corp.

Rosemount Inc.

Industrial Automation Networks Inc.

ICS Secure LLC

Consultant

Applied Control Solutions LLC

Kahler Engineering Inc.

Yogogawa IA Global Marketing (USMK)







This page intentionally left blank.






- 9 - ANSI/ISA-TR75.25.02-2000 (R2010)

Contents

Foreword ..................................................................................................................................................... 11

Abstract ....................................................................................................................................................... 11

Key Words................................................................................................................................................... 11

1 Purpose................................................................................................................................................ 13

2 Scope ................................................................................................................................................... 13

3 Definitions ............................................................................................................................................ 13

4 Control valve response ........................................................................................................................ 18

4.1

Measurement of control valve response ...................................................................................... 18

4.2

System response.......................................................................................................................... 18

4.3

Test environments used to determine control valve response .................................................... 19

4.4

Size of input signal change - regions ........................................................................................... 20

5 Maintenance and design issues affecting process control .................................................................. 21

5.1

Stem seal...................................................................................................................................... 21

5.2

Valve seat shutoff ......................................................................................................................... 22

5.3

Valve seat type ............................................................................................................................. 22

5.4

Process fluid effects ..................................................................................................................... 22

5.5

Mechanical tolerances.................................................................................................................. 23

5.6

Structural stiffness ........................................................................................................................ 23

5.7

Pneumatic positioner.................................................................................................................... 23

5.8

Actuator size/type ......................................................................................................................... 23

5.9

Electric and hydraulic actuators ................................................................................................... 24

5.10

Flow effects .................................................................................................................................. 24

5.11

Valve sizing and selection ............................................................................................................ 24

6 Process and control design issues ...................................................................................................... 24

6.1

Control loop process gain – range and variability ........................................................................ 24

6.2

Over-sizing ................................................................................................................................... 26






ANSI/ISA-TR75.25.02-2000 (R2010) - 10 -

6.3

Control valve inherent characteristic ............................................................................................ 26

6.4

Closed loop performance – control valve dynamic specification ................................................. 26

6.5

Nonlinear regions ......................................................................................................................... 28

7 Static behavior tests............................................................................................................................. 28

7.1

Important measures of static behavior ......................................................................................... 28

7.2

Applications affected and classes of performance....................................................................... 29

7.3

Testing considerations ................................................................................................................. 30

7.4

Data presentation ......................................................................................................................... 32

7.5

Design and maintenance factors important to static behavior ..................................................... 33

8 Small amplitude and medium amplitude dynamic response tests (regions 2 and 3) .......................... 33

8.1

Important measures for regions 2 and 3 ...................................................................................... 34

8.2


8.3

Testing considerations for small amplitude and medium amplitude dynamic response (regions 2 and 3) ..................................................................................................................................................... 36

8.4

Data presentation for regions 2 and 3.......................................................................................... 37

8.5

Design and maintenance factors at small amplitude and medium amplitude (regions 2 and 3) . 37

8.6

Oscillatory response..................................................................................................................... 38

8.7

Performance near the closed position.......................................................................................... 39

9 Large amplitude dynamic response tests (region 4) ........................................................................... 39

9.1

Important measures in region 4 ................................................................................................... 39

9.2


9.3

Testing considerations for large amplitude dynamic response.................................................... 40

9.4

Design and maintenance factors important at large amplitude.................................................... 41

10 References ....................................................................................................................................... 41






- 11 - ANSI/ISA-TR75.25.02-2000 (R2010)

Foreword

Publication of this Registered Technical Report has been approved by the Accredited Standards

Developer. This document is registered as a Technical Report series of publications according to the procedures for the Registration of Technical Reports with ANSI. This document is not an American

National Standard and the material contained herein is not normative in nature. Comments on the content of this document should be sent to the Secretary, Standards and Practices Board; ISA; 67

Alexander Drive; P. O. Box 12277; Research Triangle Park, NC 27709.

This technical report applies to throttling control valves in closed loop control applications. The concept has some application to open loop control applications but does not address control valves used in on-off control service. In the context of this document, the control valve includes these components

⎯ a valve, an actuator, a motion conversion mechanism, and accessories such as a positioner, transducer, signal booster relay, air set, snubber, etc.

Abstract

This technical report describes the characteristic response of a control valve to step input signal changes, considering the factors that affect this response, the impact of the response on the quality of process control, and the appropriate control valve specifications. In this document, a control valve is the complete control valve body, with actuator and any accessories required for normal operation assembled and ready for use. This document identifies and defines four regions of control valve response to step input changes of varying sizes and provides guidance that can be used to relate the control valve performance to process control.

Key Words

Accessories, actuator, amplitude, backlash, bench test, closed loop, closed loop time constant, control valve, dead band, dead time, dynamic response, flow coefficient, hunting, in-process test, laboratory test, limit cycle, motion conversion mechanism, nonlinear, open loop, overshoot, process control, process response, resolution, response, static, steady state, stem position, step change, step input signal changes, step response time, step size, step test, throttling control valve, time constant, valve, velocity limiting.

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---






This page intentionally left blank.






- 13 - ANSI/ISA-TR75.25.02-2000 (R2010)

This technical report describes the characteristic response of a control valve to step input signal changes.

It considers the factors that affect this response, the impact of the response on the quality of process control, and the appropriate control valve specifications. In this document, a control valve is the complete control valve body, with actuator and any accessories required for normal operation assembled and ready for use. This document supports standard ANSI/ISA-75.25.01-2000 (R2010), "Test Procedure for Control

Valve Response Measurement from Step Inputs." See the standard for the test procedures.

Users and manufacturers have developed a better understanding of the effects of control valve response characteristics on process control. This document identifies and defines four regions of control valve response to step input changes of varying sizes. Existing standards do not include the definitions and methods to measure certain valve characteristics now understood to be important. This technical report provides guidance that can be used to relate the control valve performance to process control.

2 Scope

This technical report applies to throttling control valves in closed loop control applications. The concept has some application to open loop control applications. It does not address control valves used in on-off control service. The “control valve” in the context of this document includes the following components:

Valve: A valve is a device used for the control of fluid flow. It consists of a fluid containing valve body assembly, one or more ports between connection openings and a moveable closure member, which opens, restricts or closes the port(s) (see ANSI/ISA-75.05.01-2000 (R2005), "Control Valve

Terminology").

Actuator: An actuator is a device that supplies the force and causes the movement of the valve closure member. Commonly these are fluid or electrically powered (see ANSI/ISA-75.05.01-2000 (R2005)).

Actuators often use air but other types use electric, hydraulic and electro-hydraulic power.

Motion conversion mechanism: A mechanism installed between the valve and the power unit of the actuator to convert between linear and rotary motion where required. The conversion may be from linear actuator action to rotary valve operation or from rotary actuator action to linear valve operation.

Accessories: Additional devices used in the operation of the control valve. As described in

ANSI/ISA-75.05.01-2000 (R2005), typical examples include a positioner, transducer, signal booster relay, air set, snubber, etc.

3 Definitions

This document and ANSI/ISA-75.25.01-2000 (R2010) make use of terms as defined in ISA-51.1-1979

(R1993) "Process Instrumentation Terminology," and some of the essential terms are repeated here for convenience. In the specific area of nonlinear dynamics, it was determined that some terms defined in

ISA-51.1-1979 (R1993) lacked the precision desired for these documents. Others were inconsistent with the terminology used in the nonlinear control literature. A common set of definitions is used in

ANSI/ISA-75.25.01-2000 (R2010) and this document. Those used only in this document are marked with an asterisk (*).

3.1 backlash:* in process instrumentation, a relative movement between connected mechanical parts, resulting from looseness when motion is reversed [ISA-51.1-1979 (R1993)]. Sometimes also referred to as slop, lost motion, or free play.

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---






ANSI/ISA-TR75.25.02-2000 (R2010) - 14 -

Output

b a a < resolution

≤

b

Input

c d c

≤

dead band < d

Time

Dynamics are not shown

Figure 1 —Dead band and resolution

3.2 closed loop time constant:* the time constant of the closed loop response of a control loop, used in tuning methods such as Internal

Model Control (IMC) and Lambda Tuning. The closed loop time constant is a measure of the performance of a control loop. the range through which an input signal may be varied, with reversal of direction, without initiating an observable change in output signal [ISA-51.1-1979 (R1993)]. In this technical report and in standard

ANSI/ISA-75.25.01-2000 (R2010) it is defined in percent of input span. Note that in some other literature this definition is used for dead zone.

3.4 dead time ( T d

): the time after the initiation of an input change and before the start of the resulting observable response. a zone of input for which no value of the output exists [ISA-51.1-1979 (R1993)]. the time-dependent output signal change resulting from a defined time-dependent input signal change.

Commonly used input signal changes include impulse, pulse, step, ramp, and sinusoid [McGraw-Hill

"Dictionary of Scientific and Technical Terms", sixth edition, 2002]. Dynamic means that the control valve is moving. Dynamic response can be measured without process loading in bench top tests with simulated or active loading in a flow laboratory or under normal process operating conditions.

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---






- 15 - ANSI/ISA-TR75.25.02-2000 (R2010)

G

Z

/ G

Z02

: the response gain G

Z

divided by the response gain G

Z02

determined from the multi-step test performed with a step size of 2 percent. The ideal gain ratio equals 1.0 for tests about any nominal position.

G

R

=

G

Z

/ G

Z02

3.8 hunting:* an undesirable oscillation of appreciable magnitude, prolonged after external stimuli disappears [ISA-

51.1-1979 (R1993)]. Hunting can have two forms: oscillations occurring near the stability limit of a linear system or the limit cycling tendency of a nonlinear system. an oscillation caused by the nonlinear behavior of a feedback system. These oscillations are of fixed amplitude and frequency, and can be sustained in a feedback loop even if the system input change is zero. In linear systems, an unstable oscillation grows theoretically to infinite amplitude, but nonlinear effects limit this growth [Van De Vegte, J., "Feedback Control Systems", 2nd edition, Prentice Hall, 1990, p. 14]. See also hunting [ISA-51.1-1979 (R1993)].

3.10 memory:* in the context of small signal nonlinear dynamics, is that property of a nonlinearity which makes it sensitive to the current direction, and the history of the input signal. Memory requires the inclusion of direction arrows on those line segments of an X-Y plot that are directionally sensitive. a nonlinear system is one whose response depends on the amplitude and the nature of the input signal, as well as the initial conditions of the system. As an example, a nonlinear system can change from being stable to unstable by changing the size of the input signal.

When a nonlinear system is driven towards a setpoint by feed back control action, it is likely to develop a limit cycle. The amplitude and frequency of such limit cycles are a function of the nature of the nonlinearities which are present, and the effective gain of the feed back control action. As the gain of the feed back is increased, the frequency of the limit cycle is likely to increase. More aggressive gain increases may produce behavior such as bifurcation, frequency doubling and eventually chaotic behavior.

3.12 nonlinearity:* there are many types of nonlinearities, although they can be generally grouped into two main groups: simple nonlinearities without memory and more complex nonlinearities with memory [Van De Vegte, above, Gibson, J. E., "Nonlinear Automatic Control," McGraw-Hill, 1963]. Not the same as in

ISA-51.1-1979 (R1993), linearity: the closeness to which a curve approximates a straight line.

3.13 overshoot: the amount by which a step response exceeds its final steady state value. Refer to Figure 24 of

ISA-51.1-1979 (R1993). Usually expressed as a percentage of the full change in steady state value. the position of the closure member relative to the seated position. In this technical report and in standard

ANSI/ISA-75.25.01-2000 (R2010) expressed as a percent of span.

3.15 resolution: smallest step increment of input signal in one direction for which movement of the output is observed.

Resolution is expressed as percentage of input span. The term in this document means: the tendency of a control valve to move in finite steps in responding to step changes in input signal applied in the same direction. This happens when the control valve sticks in place, having stopped moving after the previous step change.






ANSI/ISA-TR75.25.02-2000 (R2010) - 16 -

3.16 response: the time history of a variable after a step change in the input. In this technical report and in standard

ANSI/ISA-75.25.01-2000 (R2010), the step response can be stem position, flow, or another process variable.

3.17 response flow coefficient C vR

:* apparent flow coefficient as determined by testing in an operating type environment. The data available in the operating environment may differ from the laboratory data required by valve sizing standards.

:

Z the ratio of the steady state magnitude of the process change

Δ

Z divided by the signal step

Δ s that caused the change. One special reference response gain is defined as that calculated from the 2 percent step size response time test. This is designated as G

Z02

.

G

Z

=

Δ

Z

Δ s

G

Z 02

=

Δ

Z

02 Δ s

02

Δ t s

: the time increment between sampled data points. It is the inverse of the sampling rate, f

0

.

Δ t s

= 1 / f

0

.

As used in this technical report and in standard ANSI/ISA-75.25.01-2000 (R2010), since more than one variable is being sampled, it is the time between the sets of sampled data. Ideally, all variables in one set are sampled at the same time. If data is recorded using analog equipment, the time constant for the recording equipment shall be less than or equal to the maximum allowed

Δ t . rate

0

: the rate at which data samples are taken or the number of samples per unit time. See sampling interval. in rotary valve systems, the tendency of the drive shaft to twist under load while the closure member is stuck at a given position. friction r

or T r

: the force or torque required to maintain motion in either direction at a prescribed input signal ramp rate.

3.23 static: means without motion or change [McGraw-Hill, "Dictionary of Scientific and Technical Terms," sixth edition, 2002]; readings are recorded after the device has come to rest. Static performance can be measured either without process loading (bench top tests), with simulated or active loading, or under process operating conditions. This kind of test is sometimes called a dynamic test [McGraw-Hill above], which may cause confusion. The static behavior characteristics identified as important to the control valve performance are the dead band, the resolution, and the valve travel gain. a condition of a dynamic system when it is at rest at a given value. In testing the responses of a dynamic system, step test methods are often used. The resulting system transitions from an initial steady state value to a new final steady state value.






- 17 - ANSI/ISA-TR75.25.02-2000 (R2010) a nearly instantaneous step change made to an input signal of a dynamic system with the intention of stimulating a step response of the dynamic system. Such a test is used to characterize the step response of the dynamic system.

3.26 step change time

Δ t sc

: the time between the start of a signal input step and when it reaches its maximum value. the application of a step change to an input signal in order to test the step response dynamics.

3.28 step response time ( T

86

): the interval of time between initiation of an input signal step change and the moment that the response of a dynamic reaches 86.5% of its full steady state value. The step response time includes the dead time before the dynamic response.

3.29 step

Δ s : the difference between the beginning and ending signal in a step change expressed as a percent of the signal span.

3.30 stiction (static friction):* resistance to the start of motion, usually measured as the difference between the driving values required to overcome static friction upscale and downscale [ISA-51.1-1979 (R1993)].

3.31 stick/slip:* a term that attempts to explain jerky or “sticky” motion by postulating that static friction differs substantially from sliding friction. However, friction is rarely directly measured, and “sticky” behavior can be caused by other physical effects (e.g., positioner behavior, at small amplitudes). a term that attempts to describe a limit cycle caused when the control valve “sticks” and suddenly “slips” during a change in input signal. It is the result of static friction combined with a positioner and actuator system that does not provide enough force to overcome friction at low positioner error values. constant : for first order dynamic systems, the interval of time between initiation of an input signal step change and the moment that a first order dynamic system reaches 63.2% of the full steady state change. The term is used in this technical report and in standard ANSI/ISA-75.25.01-2000 (R2010) to describe the dynamic characteristics of the analog measuring instruments.

3.34 valve travel gain:* the change in closure member position divided by the change in input signal, both expressed in percentage of full span.

G

X

=

Δ

X

Δ s

3.35 valve system approximate time constant (

τ ′

):* the time constant of a first order response without dead time, which may fit the actual control valve step response reasonably well. The approximate time constant is defined to provide a basis for comparison of the valve with other time constants, such as the closed loop time constant for the control loop. A first order system reaches 86.5% of its final step response value in two time constants; the approximate time constant is considered to be one half of the step response time, T . The use of the approximate time

86 constant in no way implies that the response of the control valve is first order. The step response of the






ANSI/ISA-TR75.25.02-2000 (R2010) - 18 - control valve is typically complex, having dead time initially, followed by potentially complex dynamics before the steady state is achieved. T includes the dead time in the initial part of the response, as well

86 as the possibility of slower settling in the last portion of the response. Some valve positioner designs attempt to achieve a slow-down in the final part of the response in order to limit overshoot.

τ ′

attempts to produce a simple linear time constant approximation of the control valve dynamic response, which can be compared to the closed loop time constant of the control loop on the same basis in time constant units.

Note that as the portion of T that is dead time increases, this approximation becomes less ideal.

86 the maximum rate of change that a system can achieve due to its inherent physical limitations. time t w

:* the time spent after a step input change waiting for the response to come to the new steady state value. a plot of the output excursions plotted against input excursions. Input-output plots are useful for defining the steady state characteristics of nonlinearities.

4 Control valve response

4.1 Measurement of control valve response

The control valve is an integral part of the process control loop. The other parts are the sensing devices, the controller, and the process under control. A certain level of quality of process control performance is required for every application. To achieve this, the control valve must have the appropriate response characteristics.

The control valve modifies a fluid flow in response to a signal from a process controller. This is accomplished by moving the closure member resulting in a change in the flow coefficient. This response has both static and dynamic behavior characteristics. The static behavior characteristics identified as important to the control valve performance are the dead band, the resolution, and the valve travel. The dynamic response characteristics of interest are the dead time, overshoot and the step response time.

Control valve performance may be improved by reducing the dead band, improving the resolution, stabilizing the travel gain, and reducing both the dead time and the step response times and by minimizing overshoot. The effect of the dynamic properties is a function of the process time constants.

The effect of the static properties is a function of the process static gain properties.

Several parameters can be used to describe the response of the control valve. These parameters are the flow coefficient, the value of the measured variable, and the stem position. They are discussed below in detail.

4.2.1 Flow coefficient

The fundamental parameter is the valve flow coefficient. The flow coefficient, or C , is calculated from v the flow through the valve, the fluid density and the differential pressure across the valve [see

ANSI/ISA-75.01.01 (60534-2-1 Mod)-2007, “Flow Equations for Sizing Control Valves” and

ANSI/ISA-75.02.01-2008, “Control Valve Capacity Test Procedure”]. For in-process testing, a response flow coefficient C , is defined in this technical report. The response flow coefficient vR

C is calculated vR from the flow, the differential pressure and the density. It is defined this way to allow the use of the available data and test environment where this differs from the standard definition of C . This coefficient v




--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---



- 19 - ANSI/ISA-TR75.25.02-2000 (R2010) may differ from catalog values. Changes in the closure member position will change the flow coefficient even under choked flow conditions. The response flow coefficient will test for the following uncertainties: a) The control system output signal may not be the exact stem position. b) The stem position may not be the exact closure member position. c) The relation between the closure member position and the flow coefficient is not certain.

4.2.2 Process response

The process response provides information similar to that provided by using the flow coefficient. The data is the change in temperature or in composition or in flow rate. It may be distorted by any errors in the process measurement, changes in differential pressure across the valve, noise, and changes in the fluid density. The process measurement instruments may not provide the desired resolution or accuracy.

Any errors and all the dynamics of the transmitter and the process can distort the data. Consider these uncertainties in reporting the response of the control valve.

If the process installation includes a flowmeter, or if it is a relatively “fast”, self-regulating process, then this task is made easier. If the process is not a relatively “fast”, self-regulating process, it may be difficult to use the process response to determine the performance of the control valve. For instance, for an integrating process (such as a level or pressure), the fluid flow rate can be calculated from the rate of change or from differences over time in the process variable. In some applications, the process information may provide only a limited set of performance parameters such as dead band and resolution.

Stem position measures the response of the physical parts of the control valve. Some control valve accessory systems include the ability to measure and report control valve stem position and other parameters. This method can eliminate the uncertainties of the dynamic response parameters of the process and process measuring instruments. It is necessary to correct for the response of the stem measurement instrument. The stem position measurement is well suited for determining the dynamic response characteristics (dead time, step response time, overshoot) of the control valve. It does not account for the differences between the stem movement and the closure member movement or between the closure member movement and the flow coefficient. This makes it less useful for measuring dead band and resolution. These differences can be caused by such things as backlash in the actuator system, or stem

11

windup on rotary valves. However, if the correlation between the stem position and the flow coefficient is known for a particular valve or valve design, it can be combined with the response of the stem position to predict the response of the flow coefficient.

4.3 Test environments used to determine control valve response

The control valve can be tested in three different environments. Each of these environments has tradeoffs in practicality and measurement uncertainties. The test environment will determine which parameter (flow coefficient, process response or stem position) can be used to measure the response characteristics of the control valve. The test environments are discussed further and are as follows: a) Bench test without process flow (e.g. plant instrument shop, laboratory, manufacturing site) b) Laboratory test with flow, simulating a plant process

_____

1

Stem position refers to rising stem valves. Shaft position refers to rotary valves. This document uses stem to mean either stem or shaft.




--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---



ANSI/ISA-TR75.25.02-2000 (R2010) c) In-process test during plant operation

- 20 -

Table 1 — Control valve output measurements and test environments

Test Environment Parameter Recommended for Measurement

(listed order of preference within each environment)

Static Performance

Measures - Resolution, dead band, and valve travel gain.

Dynamic Performance

Measures - Dead time, response time, and overshoot.

Stem position Bench test

Tested on a bench, in a plant, a lab, a shop, etc. with no flow

(may be under applied pressure), with valve completely assembled, and with recommended packing load.

Stem position (use only if a laboratory or in process test is not available)

Laboratory test

Tested in a laboratory flow loop with flow

Flow coefficient for dead band and resolution

Process response, where possible

Stem position

Flow coefficient

Process response

In-process test application

Tested in the actual process

Flow coefficient (if measurements available)

Process response, where possible

Stem position

Flow coefficient (if measurements available)

Process response

Table 1 summarizes the control valve test environment and the resulting information. Some of these combinations may be eliminated by constraints mentioned later in this document.

4.4 Size of input signal change - regions

The character of the control valve response usually changes with the size of the change in the input signal. For the purpose of this report, four regions are defined.

The four regions are defined only as an aid to understanding control valve response. They are not intended for catalog data or purchase specifications. The size of each region and the boundaries between the regions are determined completely by the specifications used. For specification guidance see 6.5.2.

Region 1 is defined as small input steps which result in no measurable movement of the closure member within the specified wait time.




--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---



- 21 - ANSI/ISA-TR75.25.02-2000 (R2010)

Region 2 is defined as input step changes which are large enough to result in some control valve response with each input signal change, but the response does not satisfy the requirements of the specified time and linearity.

Region 3 is defined as step changes which are large enough to result in flow coefficient changes which satisfy both the specified maximum response time and the specified maximum linearity.

4.4.4 Region 4

Region 4 is defined as input steps larger than in region 3 where the specified magnitude response linearity is satisfied but the specified response time is exceeded.

5 Maintenance and design issues affecting process control

Control valves consist of a valve body, an actuator, and often accessories such as a positioner to form a control valve system. The interrelationship of these parts can be complex. It is advisable for the specifier and the manufacturer to communicate and to understand the requirements and constraints before selecting the valve that will perform correctly in service and can be maintained properly. Design, installation, service, and maintenance will affect valve performance. The user site may also have maintenance and business requirements that must be satisfied. Ongoing operational conditions and requirements may affect ongoing control valve performance. It is not possible to make an intelligent selection of a suitable control valve for a given application without adequate information. Process fluid characteristics must be known: pressures, temperatures, (both normal and maximum), flows, (minimum, normal, and maximum), fluid characteristics (density, viscosity, vapor pressure, etc.), required leakage class, temperatures, and corrosion issues, are typical application requirements. The capabilities of the local maintenance organization, support by the manufacturer, local needs and desires should be considered. Some requirements may be determined by regulations and laws. Good knowledge resources include catalogs, experience, and knowledgeable manufacturer representative. The required speed of response should be estimated based on anticipated process control requirements.

Installation and operation outside the design conditions may damage the valve. The only defense is knowledgeable maintenance and management support. Poorly trained mechanics, poor records, inadequate spare parts, and no time to do it right, can lead to sticky valves, broken stems, actuators that do not move, and positioners that do not function. The discussion below is not intended to dictate technology.

5.1.1 Materials

Most common valve stem seals are based on PTFE (polytetrafluoroethelene) and die-formed graphite.

Until recently, graphite was required for applications involving high temperature, high pressure, fire safety, or because of chemical incompatibility with PTFE. Some implementations of graphite have a coefficient of friction much higher than PTFE and a tendency to stick to the stem, which will degrade the control behavior of the control valve.

Recent developments provide alternatives. These often offer better sealing and reduced friction and can tolerate a wide range of temperatures. New materials and new formulations are available, along with combinations of the old materials and various design details (see references 15 through 18).






ANSI/ISA-TR75.25.02-2000 (R2010) - 22 -

The packing system is now an engineered component of the control valve. Factors to be considered include a) temperature, pressure, and chemical compatibility; b) sealing and fire safe requirements; c) friction with actuator sizing and process control requirements; d) service life, maintenance, and total life costs; and

5.1.2 Design

Limitations on fugitive emissions have encouraged manufacturers to develop innovative material selections, combinations of packing materials, and installation configurations, such as “live loading”. Live loading is the use of a spring, typically a Belleville washer, to maintain a constant force on the packing.

The more nearly constant and optimum compression load on the packing may improve the life of the packing and minimize both leakage and friction. New configurations typically use innovative material combinations and are designed to minimize leakage and friction. Extensive testing on packing leakage and stem friction by the vendors provides the basis for their design and recommendations. Deviations from these recommendations may result in higher leakage and greater friction. Attention to the details is paramount. The manufacturers offer a variety of design packing details, such as the number of rings of packing, style, and materials and filling materials. The packing box design provides a number of choices and details.

There is a balance between stem seal leakage and packing loading. Over tightening can lead to excessive stem seal friction. It will destroy the packing and result in poor dynamic performance and stem leakage.

5.2 Valve seat shutoff

Undesired static and dynamic control valve characteristics which may develop near the shut off position can be avoided when control is normally, and preferably, with the closure member well away from the seat position.

There are designs where the valve seat continues to contact the closure member beyond the initial opening. This creates friction and may degrade resolution and dead band depending on the actuator and positioner.

5.3 Valve seat type

Valves that have seats in contact with the flow closure element while in the control range include some ball and plug valves and some cage guided globe valves. The magnitude of these effects will depend on the seat style, the design details and the materials used. Friction affects the static performance parameters; resolution and dead band. Review the catalog data and drawings for information.

5.4 Process fluid effects

Viscous and sticky fluids such as resins will tend to resist stem movement and will increase the resistance to movement. Rotary valves are less affected by process fluid and packing friction.






- 23 - ANSI/ISA-TR75.25.02-2000 (R2010)

Mechanical backlash in any of the connections between the closure member, the actuator, and the positioner, will increase dead band. That is, two steps in the same direction may have a different result than if the direction were reversed between the steps.

Weak or flexing linkages between the valve, the actuator, and the positioner can increase dead band and resolution.

5.7 Pneumatic positioner

The positioner air flow capacity limits, varying gain, linkage wear, and internal friction can add significant non-linearity to the control valve response time. Internal friction can degrade dead band and resolution.

Interactions between the positioner and the actuator can create dynamic non-linearity.

Restrictions in the flow path into and out of a pneumatic actuator will slow the response. These restrictions may be undersized tubing, damaged tubing, undersized fittings, solenoid valves orifice size, and manual valves. Restrictions will have a lesser effect on small changes in signal but will affect the time of response for large (>10%) signal changes.

The volume of air that the positioner must supply and exhaust limits the speed of response. A greater actuator volume requires a larger change in the mass of air in the actuator and may delay the response and increase the dead time. Three volumes may be described. The stroke volume is the change in volume during the stroke. The dead volume is the total volume minus the stroke volume. The total volume is fixed by the design. The stroke volume will vary with stem position and pressure creating a dynamic non-linearity. In any calculations, consider that it is actually the mass of air in the actuator that creates the pressure.

5.7.3 Supply pressure and capacity

Inadequate air supply capacity and pressure will limit the dynamic performance of the control valve.

Undersized and limiting piping or tubing, air filters and supply regulators limit the air capacity. Dirty filters and partially closed block valves will slow or prevent response. The dead time and the time constant will be increased. The times for opening and for closing the valve will probably differ from each other. See

ISA-7.0.01-1996, "Quality Standard for Instrument Air."

5.7.4 Valve pneumatic accessories

Accessories such as volume boosters, quick-release valves, and solenoid valves will all affect the performance. A volume booster can improve speed of response.

Selection of the actuator size requires accurate information on friction, and process pressures, temperatures, and fluid characteristics. Safety factors in sizing actuators must consider safety, and the quality of the available information. Actuators are sized based on the minimum air supply pressure available but the actuator design must also withstand the maximum air supply pressure. To reduce dead time and to minimize the dead band, available stroking power must be greater than the minimum force






ANSI/ISA-TR75.25.02-2000 (R2010) - 24 - required to move the valve stem. An inadequate or undersized actuator or a positioner with poor performance will result in poor response in both response time and magnitude.

The selection of the type of actuator involves considerations of valve size, and the design details, air supply pressure, and manufacturer offerings.

5.9 Electric and hydraulic actuators

The actuator discussion above applies to all types of actuators, pneumatic, hydraulic and electric, but especially to pneumatic. Hydraulic actuators are expected to provide very good performance and are typically used for larger valves and the more difficult applications. Some types of electric motor actuators will have good resolution and may have a very small dead band, but they may lack the required speed when used with larger valves. These statements are only broad generalities. The user must investigate the data and claims from the manufacturers.

Dynamic imbalance from the effects of flow on the control valve closure member can degrade repeatability and dynamic linearity. Choking will limit flow capacity and vibration can affect the positioner performance. A valve operated in the flow-to-close mode may show stem instability as the plug approaches the seat and the hydrodynamic plug forces increase rapidly. Some butterfly designs have a reversal in flow induced shaft torque depending on position. These forces will vary with flow rate and pressure drop.

5.11 Valve sizing and selection

The installed flow characteristic of the valve may not be the same as the inherent flow characteristic of the valve. See further discussion in 6.3.

6 Process and control design issues

6.1 Control loop process gain – range and variability

A fluid process is much easier to control if the control dynamics remain nearly constant over the full range of operating conditions.

The key dynamic parameters include: process gain, process time constant, dead time, controller dynamics, sensor dynamics, and control valve static and dynamic properties. The values of these parameters and their changes over the operating range will determine how well the process can be controlled to achieve the following: a) Acceptable level of process performance during process operation b) Low process variability c) Final product of acceptable uniformity d) Low manufacturing cost e) Ability to meet manufacturing demand f) High level of plant safety g) High level of environmental compliance




--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---



- 25 - ANSI/ISA-TR75.25.02-2000 (R2010) h) Successful startups, product grade transitions, and shutdowns i) Recovery from process upsets

The control loop is subject to setpoint changes and to upsets. It establishes new operating conditions to recover from load disturbances or to meet the new setpoint. Parameters such as the required time constant and minimum dead time are set by the process design and instrument selection. The process gain is central to the valve selection, process dynamics, and fluid transport system. The process gain is determined by the process dynamics, the control scheme and the fluid transport system. The fluid transport system characteristics are determined by the pump/compressor, piping, equipment, and the control valve. In all cases, the valve capacity and characteristics influence the process dynamic control.

Control design strategy can compensate to a certain extent for physical equipment and piping design limitations. The task of the design process is to select the right sized control valve and suitable characteristics. This will help keep the installed control loop process gain in the acceptable range, and reduce the variation over the normal operating range of the process.

The process gain of a self-regulating process is the ratio of the process variable change (e.g., temperature, composition, pressure, or flow) compared to the change in the controller output that caused the change. The process gain of an integrating process is the ratio of the change in the rate of change of the process variable (e.g., level, gas pressure) compared to the change in the controller output that was made to cause the change.

The process gain may be determined on-line by carrying out a series of step changes in the controller output. It may also be predicted from sizing calculations, or from a dynamic simulation of the plant design. In flow control applications the process gain is influenced by the relative pressure drop taken across the control valve, as compared to the rest of the fluid transport system, and the span of the sensor, transducer, or transmitter.

For non-integrating (self regulating) processes (integrating processes tend to be tank levels and certain types of vessel pressures) such as flows and pressures, the process gain should be within the range of

0.5 to 2.0 (% of span process variable change) / (% output change)

Example:

In a step test, if the controller signal output changed 5%, and the flow signal changed by 7.2% of span, then process gain = 7.2% / 5% = 1.44.

If the process gain is higher than 2.0, then the valve is oversized; has the wrong flow characteristic; the process dynamics are too sensitive (high gain); or the fluid transport system is oversized (i.e., pump may be oversized).

If the process gain is higher than 2.0, the process may be difficult to control. The valve dead band and resolution are multiplied by the process gain, therefore a high process gain increases the effective process dead band and resolution. The minimum flow change (dead band) determines the resolution of the resulting control action. A process gain lower than 0.5, will result in small flow changes and require higher controller gain. This does not create a control problem, but it may indicate a limitation to achieve adequate capacity.

There is a tendency for dead time to be longer with small changes in controller output for pneumatic actuators. This will have a de-stabilizing effect on most control loops. Dead time is a common cause for control loop limit cycles in control loops where the step response time in this region is not significantly shorter than the normal closed loop time constant of the control loop.

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---






ANSI/ISA-TR75.25.02-2000 (R2010) - 26 -

It is desirable to keep the process gain as constant as possible over the operating range of the process.

This will reduce the need to retune the controller with changes in operating conditions. Process control performance is a function of the total loop.

6.2 Over-sizing

Control valves, pumps, and pipelines are very often over-sized during the design of a plant to provide for an easy increase in plant capacity. The result is that the control valve will operate nearly closed. This often results in extremely poor control performance. The whole range of plant operation is reduced to a very narrow controller output and valve travel range. A system with a centrifugal pump operating at a small fraction of the design basis flow will have a higher control valve pressure drop. The pump will operate in the high head, low flow condition. The pipe and equipment friction pressure drops will be lower. Cavitation and damage to valves and pumps may occur. The control valve is an integral part of the fluid system, and the sizing deserves careful consideration.

In a flow control loop, if the dead band and resolution are 1% of travel, and the process gain is 1.0

(percent of flow change per percent input signal change), then the flow signal will exhibit a dead band and resolution equal to 1% of span. If the process gain is 5, the flow resolution will be 5 x 1% = 5%. This may be so coarse that good control is impossible. No controller tuning can hide or eliminate this problem.

The performance of the loop can be improved by either reducing the control valve dead band and resolution, or by lowering the process gain.

The pump impeller and the control valve trim should be selected for the present actual required capacity.

As production increases are desired, the control valve trim and the pump impeller can be replaced. The economic justification is the ability or inability to manufacture a product of adequate uniformity with poor control.

6.3 Control valve inherent characteristic

The inherent characteristic of a control valve is the relation between flow and valve stem position at a fixed differential pressure. The installed characteristic is the relation between flow and stem position in a real installation where the pressures and density may vary with the flow and time. It is likely that the installed flow characteristics will differ from the inherent one. The choice between the standard catalog trim characteristics of "linear", “equal percentage” and "quick opening" is made to reduce the range of process gain over the control range. If the process gain varies excessively it can be at least partially corrected through control design strategy, by compensating for valve position in the control system by nonlinear compensation, or by compensating the controller gain by gain scheduling. Or, even through the adjustment or modification of the characterizing cam in the positioner. It is advisable to minimize the variation in the process gain, or the effective loop gain, over the operating range of the process so that the net variation is no more than +/- 50%.

6.4 Closed loop performance – control valve dynamic specification

The speed of response of a control loop is determined by the objectives of the process control strategy, which in turn is set by the process manufacturing goals. The required speed of response sets the control loop performance, which can be measured in terms of bandwidth, cross-over frequency, phase margin

(see reference 10), closed loop time constant, setpoint overshoot, resonant peak and other measures.

Modern tuning methods, such as those based on design synthesis, Internal Model Control (IMC) and

Lambda Tuning (see references 12 through 14), are all based on the determination of control loop performance by specifying a desired closed loop time constant. The closed loop time constant will differ depending on the tuning method used. The required speed of response will vary widely from loop to loop.




--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---



- 27 - ANSI/ISA-TR75.25.02-2000 (R2010)

70

65

60

(%)

55

50

Flow Rate

Actuator Position

Input Signal

45

0.5% Steps 1% Steps 2% Steps 5% Steps 10% Steps

40

35

0

4-Inch Segmented Ball Valve with Diaphragm Actuator & Positioner

50 100 150 200 250 300 350 400 450

Time (sec)

Figure 2 — Example of step response test with lost motion

One example is a design requirement to control the level of a large capacity tank with a closed loop time constant of 30 minutes in order to use the surge capacity of this large tank to reduce upsets in the rest of the process. This level controller requires a valve with a small dead band to avoid cycling but does not require fast response. Another example is a time critical loop, such as pressure control of an incompressible liquid header supplying a number of critical users. This may require a closed loop time constant as short as one second. This high speed of response reduces the interaction between the various flow control loops that supply each user.

6.4.1 Control valve speed of response – step response time T , approximate time constant

τ ′

86

The control loop contains many dynamic elements: the process, the control valve and the transmitter. It is theoretically possible to tune a control loop to be faster than its internal dynamic elements. This can only be accomplished in a stable and robust manner when the internal dynamics are well behaved, and have parameters that are constant with time. This is seldom true in a plant environment, and even less so when control valve dynamics are involved. As dynamic parameters vary, the control loop could become unstable and process variability will increase. For this reason, it is common practice to tune control loops to be slower than the open loop dynamic of the component elements.

The control valve system will not limit the control loop response if its speed of response is 10 to 20% of the next slowest control loop component. Minimum step sizes for good valves may range from 0.2 to

2.0% for the required T .

86

6.4.2 Impact of dead time on loop stability

The step response time ( T ) is made up of two components, the dead time (

86

T ) and the remainder of d the time. Control loop stability is especially sensitive to dead time; this is the most de-stabilizing of the time dependent dynamics for a control loop (see reference 15). Equally de-stabilizing is the tendency of the dead time to vary. Pneumatic actuators tend to exhibit dead time while the positioner transfers sufficient power air to the actuator to overcome friction and to move the valve closure member. This tendency also is often amplitude dependent and smaller step changes exhibit a much longer dead time than larger step changes. For simplicity, a single response time specification, T , is used to represent

86






ANSI/ISA-TR75.25.02-2000 (R2010) - 28 - the total step response time of the control valve. It remains desirable to minimize the dead time portion of

T .

86

The control valve response is nonlinear. Subclause 4.4 defines valve characteristic responses for four ranges of input step change size. Region 3 is the normal operation step size. T

86 has meaning only within this range. A very good valve may have a dead band of 0.1 to 1.0% and a step resolution of 0.05 to 0.5%.

6.5.1 Control valve minimum position

It is more difficult for most control valves to make accurate small signal step movements when the closure member approaches the closed position (see Figure 8). Friction increases as the seat is contacted. The change in flow coefficient also becomes less predictable at small openings. It is normal to specify a minimum valve position for which the valve dynamic specification applies.

6.5.2 Control valve dynamic specification

Control valve dynamic specifications should be based on the control loop dynamic requirements.

For example:

Desired control loop dynamics – closed loop time constant of 10 seconds.

Control valve step response time T is 40% of the closed loop time constant, or 4 seconds, hence the

86 desired control valve approximate time constant is 2 seconds.

The valve system shall respond to step changes from 1% to 10%. This means that the control valve will respond within the T

86 maximum specification for step changes ranging from a minimum of 1% to a maximum of 10%.

The control valve has to operate within the above specification down to a minimum of 10% open.

The example above serves to illustrate that a range of dynamic specifications is needed for control valve speed of response, which depend on the specific requirements of the control loop. There are applications where a very fast response is needed, with a

τ ′

of 0.2 seconds, and a T

86 of 0.4 seconds.

Other applications exist where the speed of response of the control valve is not an issue, and it may be acceptable to specify a

τ ′

of as long as 10 seconds, and a T

86 of 20 seconds. Such a specification may allow the selection of a less expensive valve actuator. A

τ ′

in the 1 to 3 second range, and a T

86 in the 2 to 6 second range will suit the majority of control valve applications in most process plants. Larger valves and larger actuators tend to have longer time constants.

7 Static behavior tests

7.1 Important measures of static behavior

See ANSI/ISA-75.25.01-2000 (R2010) for the standard testing procedures.

Clause 5 of ISA-51.1-1979 (R1993) gives test procedures for a variety of static performance parameters that measure differences in the input-output relationship. Many of these are derived from the full-scale bench top calibration cycle familiar to most instrument engineers [e.g., Figure 30 of






- 29 - ANSI/ISA-TR75.25.02-2000 (R2010)

ISA-51.1-1979 (R1993)]: dead band, linearity, repeatability, and reproducibility. These bench tests are valuable and provide considerable information. They do not measure the effects of varying friction and hydraulic forces. The hysteresis and dead band cannot be separated from this test, hysteresis and linearity are usually dominated by smoothly varying errors that accumulate significantly only over large strokes, and repeatability and reproducibility are important only in open loop systems

2

.

For most closed loop applications the key static performance discrepancies occur with small amplitude changes when the input reverses direction or when it continues in one direction after coming to rest.

These are identified as dead band and resolution

3 respectively.

The in-process test is conducted with the control valve installed in the process, and usually with the process transmitter used to measure the result (see reference 4). Bench top tests (without process loading) measuring actuator position or stem position provide necessary but not sufficient data to predict in-process performance. Possible development of better instrumentation would allow reliable detection of the closure member position. This could be used during normal process operation; any erratic flow behavior caused by the valve trim might still be missed.

Another possible solution is to use actuator position data for conditions in which the manufacturer can provide data showing, for typical operating conditions (typical temperature, pressure drop, and valve friction), that actuator motion causes distinct changes in flow coefficient.

As a compromise, laboratory tests, with flow simulating a plant process load, allow accurate measurements with reasonably realistic loading. Such tests allow predictions of results under normal operating conditions.

7.2 Applications affected and classes of performance

In many control loops, the ability of the control valve to make small moves (<1%) accurately is more important than the ability to make large moves quickly. The ability of the control valve to make small moves allows the process to be properly regulated. Reducing the controller gain, to eliminate the cycling, degrades the response of the control loop to load disturbances and setpoint changes. Control is further degraded when the dissatisfied operator switches to manual control.

The non-linearities that occur at small signal amplitudes are important because they tend to generate limit cycles. The magnitude of the non-linearities determines the amplitude of the cycle. Controller tuning several minutes, creating upsets that may affect the entire operation. It is very common for flow and pressure control loops to cycle, especially when the valves have high process gains. Control loops in a process area interact, sometimes very strongly. A pressure controller for a liquid header that supplies several users is a good example. When such a loop limit cycles, the flow loops supplying liquid to each user will also cycle, thus destabilizing the whole process area.

Dead band and resolution are also important in control of composition, pH, some level control applications, and temperature. These would be considered relatively slow systems, or “lag-dominant”

(see reference 6), hence the importance of static behavior over dynamic response. Difficulties in

_____

2

These conclusions hold for the majority of troublesome applications experienced by the authors. However, readers interested in these other static performance measures will find through discussion of test methods in ISA-51.1-1979 (R1993) and

ANSI/ISA-75.13.01-1996 (R2007).

3

The meaning of dead band--measured when reversing direction--is explicit in ISA-51.1-1979 (R1993). However, the present meaning of resolution--measured while continuing in one direction--is adopted here consistent with our physical understanding and for lack of a better term in ISA-51.1-1979 (R1993).






ANSI/ISA-TR75.25.02-2000 (R2010) - 30 - establishing the desired process gain also contribute to the importance of static behavior in these applications. Advanced Process affects only the period of the cycle, not the amplitude. The period can range from fractions of a minute to Control systems such as Dynamic Matrix Control make use of complex plant models that assume well-behaved control valves; these generally require dead band less than 1%.

The “small valve in parallel with a large valve” scheme has been used for difficult pH applications (see reference 6). The large valve is periodically adjusted to keep the small valve near 50% open. This is called “position control”. The small valve provides the control resolution, the large valve provides the rangeability. The disadvantage is in the dynamics for large changes in demand. The position control loop is tuned with low gain and moderate reset, response is delayed, and an upset is caused by the change in the relatively less precise large valve.

The following factors should be considered when defining the test procedure: a) Will a pre-programmed input sequence be used or will it be manually operated? b) How abruptly should the input be changed? Step changes are simplest, but not necessarily the most realistic. c) What should be the amplitude of changes? d) What should be the wait time or duration of steady input after each small change is made? At least one wait period of several minutes should be specified to detect sustained oscillation, which would invalidate measurement of static behavior. e) How many cycles of reversal should be conducted? f) What will be the nominal positions for the small-amplitude tests?

There is no single test sequence that can be generally applied to all control valves in all four environments. An example test is shown in Figure 2. Testing requires a wait time long enough to allow the valve to reach the final position. This time may be several minutes. The pressure gauges associated with the positioner will provide an indication of the end of positioner action.






- 31 - ANSI/ISA-TR75.25.02-2000 (R2010)

60.0

59.5

62.0

Valve X / Actuator Y / Positioner Z, packing per instructions

Tested at nominal 60% open, 600 gpm, 38 psid

61.5

59.0

Input

Signal

(%)

58.5

58.0

57.5

0.1%

≤

Dead Band no response upon reversal

<

0.3%

61.0

60.5

60.0

Actuator

Position

59.5

(%)

0.1%

<

Resolution

≤

0.3%

57.0

59.0

56.5

no response while con-

58.5

tinuing in same direction

56.0

58.0

900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300

Time (seconds)

Figure 3 — Example of a test using a series of small steps

7.3.2 Tests without process load in a plant instrument shop or control valve manufacturing site

This environment provides valve data but not process control data because field conditions are not present and the flow is not measured. Some realism can be simulated by tightening the valve packing to a specified value as described in 6.3.4.2 of ANSI/ISA-75.13.01-1996 (R2007), "Method of Evaluating the

Performance of Positioners with Analog Input Signals and Pneumatic Output". Detailed test sequences can be run at multiple nominal positions. However, other factors work to limit realism; some known factors with today’s designs are listed below. These factors may also affect in-process testing. a) Friction in many valves changes during the first hundred cycles of operation. Non-live-loaded PTFE packing may experience a relaxation of stress after a few cycles of operation. b) Ball valves with tight shutoff and seals that self-lap in service may experience much higher friction after only a few hours of operation in the plant, but later friction may decrease. c) Graphite packing for high-temperature service may experience much higher friction at room temperature. d) PTFE packing for moderately high temperatures may experience lower-than-realistic friction at room temperatures. e) Significant vibration of the process piping may change the results for an installed valve. f) It is not certain that the valve will hold the flow coefficient steady when the closure member is stationary (see reference 7). g) Testing at manufacturing sites normally will be limited by operational limitations.






ANSI/ISA-TR75.25.02-2000 (R2010) - 32 -

7.3.3 Laboratory testing with flow, simulating a plant process

This environment provides more information on valve performance by applying process loading and using the process variable to measure the results, as required in some specifications (see reference 4). Prior to testing, stroke the valve sufficient cycles to break-in all sealing and guiding surfaces.

61.5

58.6

55.6

52.6

49.7

Input Signal

Stem Position

0.0 90.2 180.4 270.7 360.9 451.1

Figure 4 — Step response test 0.5% steps 7 - 8% dead band

7.3.4 In-process testing during plant operation.

If the process is self-regulating

4

(or, for integrating processes, if the time derivative can be calculated to infer flow results) and if noise and process disturbances are relatively small, this environment provides useful data but only at the allowable operating conditions. For some valve styles, this is the only way to get a useful answer. The process should be allowed to operate for some time after the valve is installed, to allow break-in of all sealing and guiding surfaces to their asymptotic frictional state.

The simplest method of presenting the results is direct plotting of the time series as in Figures 2, 3 and 4.

The dynamic response for the small steps can be measured on the same graphs. Multiple cycles can be shown, and multiple output variables can be shown if available (e.g., stem position and flow). The test shown in Figure 3 is very time-consuming. Figure 4 shows an extreme case of a small step size test with a large dead band, a valve with high friction, and no positioner. Using a pre-programmed series of steps, bounds on the dead band and resolution can be set, as shown in the figure, providing overshoot does not occur on the output. In contrast, the ISA-51.1-1979 (R1993) method of measurement requires an operator to move the input slowly in the smallest possible increment until an output change is observed.

_____

4

Defined in ISA-51.1-1979 (R1993)






- 33 - ANSI/ISA-TR75.25.02-2000 (R2010)

The labels on Figure 3 show data interpretation according to ISA-51.1-1979 (R1993) as follows.

Dead band is the range through which the input signal may be varied in one direction, after a reversal of direction, without initiating an observable change in output signal. After the second step in the up direction, some motion did occur (albeit small); therefore, the dead band is less than the sum of two steps. Following ISA-51.1-1979 (R1993), largest dead band from all recorded reversals is reported and from all nominal positions if more than one position was tested.

Another method is to plot only the static results (after the control valve stops moving on each step) in the x-y domain. However, if overshoot occurs, which prevents measurement of dead band and resolution, this information will be lost in transforming from the time series to x-y plots.

Either plotting method can be used to determine the actual cause of the measured dead band and resolution if further measurements are taken of at least one internal state of the control valve. For example, with pneumatic actuators a pressure measurement enables determination of friction.

Measuring motion at multiple locations enables determination of backlash.

Labeling of graphs should state clearly what loading was applied to the control valve, and the condition of the valve (new versus worn). For bench top testing, if comparisons are to be made of two control valves, or altered conditions within one control valve, it is helpful to have an independent measure of the total friction.

7.5 Design and maintenance factors important to static behavior

Friction is usually the dominant factor in the static behavior. Actuator sizing, positioner design, drive-train design, and other factors are also important. These factors are discussed in Clause 5.

8 Small amplitude and medium amplitude dynamic response tests (regions 2 and 3)

Dynamic response, in this document, means valve response versus time for changes in control valve input signal. Dynamic response can be measured either without process loading (bench top tests) or with active loading.

This clause covers input amplitudes below which velocity limiting, region 4 (defined in Clause 4), occurs.

Steps smaller than 10% commonly avoid velocity limiting. It is possible for velocity limiting to occur on

5% steps, or even 2% steps, if the positioner has a small air flow capacity relative to the actuator volume.






ANSI/ISA-TR75.25.02-2000 (R2010) - 34 -

65

Stem

Position

(%)

67

158 159

68

A

67

B

160 161

Time (seconds)

162 163 164 165

C

166 167 168

66

I/P Input Signal

2% Steps on 4-Inch Globe Valves with Standard Actuators and Positioners

Nominal Q = 600 gpm,

Δ

P = 38 psid, water T = 120 o

F

Packing maintained per instructions

C

A

66

65

B

64

178 179 180 181 182 183

Time (seconds)

184 185 186 187

Figure 5 — Example graph of time series tests showing step response times

Because control valves are not linear in behavior, the dynamic response is strongly amplitude dependent.

Usually, 1% steps will give response times very different from 5% steps. The response is often much faster for 5% steps than 1% steps

5

. Small-step dynamic response cannot be scaled down from 100% steps, since the two amplitude ranges are governed by completely different dynamics. Finally, results become inconsistent as the amplitude is reduced to approach the dead band and resolution limits; this is the definition of region 2.

8.1 Important measures for regions 2 and 3

Frequency response, using sine wave inputs, is a useful test described in standard ANSI/ISA-75.13.01-

1996 (R2007). The controller output signals of most control loops are simulated better by sine waves than by square waves (even for digital controllers correctly applied and tuned). However, frequency response is generally too difficult for tests within the scope of ISA75.25, especially tests under normal process operating conditions. The simplest and currently most popular test, step response, may be the only method feasible in all environments for which tests are desired.

_____

5

For an example of this trend, see reference 8.

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---






- 35 - ANSI/ISA-TR75.25.02-2000 (R2010)

The closed loop input signal to a control valve is usually of limited bandwidth; open loop step response is misleading for some applications. Particular areas of uncertainty are the measured dead time and overshoot. Dead time, T , is commonly measured after a step input. Tests show that dead times can be d much longer with the slowly changing controller output signals typical of many control systems. Whether the control valve will overshoot depends on controller signals.

1 0 0

1 0

1

0 . 1

0 . 1 1 1 0

S t e p S i z e , P e r c e n t

Figure 6 — Step response for four valves

Measurement of response time close to the final position is difficult because the stem often approaches the final position asymptotically; for example, the three control valves in Figure 5. T , for example, is

98 especially difficult to measure if there is significant noise on the signal or limited signal resolution.

Rather than requiring one set of numbers for all processes, it has recently been proposed that consistency of dynamic response is more important. The user might ask that gain and dead time be consistent within some tolerance

6

for all steps over a range of amplitudes, regardless of direction and history. With pneumatic actuators and positioners, consistent dynamic response, which defines region 3, usually occurs only on steps much larger than the dead band.

Due to the amplitude dependence of control valve dynamic response in region 2, any specification of response time for region 2 must state the amplitude of step (or other input) for the test.

_____

6

For example, a tolerance of +/- 10% on gain and +/- 50% on dead time.






ANSI/ISA-TR75.25.02-2000 (R2010) - 36 -


Process dynamics will vary over a wide range from loop to loop. In a well-designed process control strategy, it is common to specify the required speed of response of each control loop, so that all loops meet the manufacturing requirements in a coordinated fashion. The required speed of response can be specified by selecting a closed loop time constant. To achieve this speed of response means that the process dynamics, the transmitter dynamics and the valve dynamics ideally are 5 times, and possibly 10 times, faster in order to maintain closed loop stability. This is the control valve dynamics in region 3, together with the need to have these dynamics remain consistent.

In liquid pressure and flow control, the control valve dynamics are typically slower than the process dynamics. The control valve dynamics will dominate the open loop response and determine the closed loop time constant.

Typically in hydraulic process applications, such as in pulp and paper applications for instance, there are pressure and flow applications that require very fast dynamics. Requirements exist for some pressure control loops to have a closed loop time constant as fast as one second. This would require the control valve T to be in the range of 0.1 to 0.2 seconds. Control of liquid flows often results in desired closed

86 loop time constants of about 10 seconds. A control valve T of about 1 to 3 seconds would be

86 reasonable to meet this performance. Temperature control is always dominated by the thermal lags of the process, with time constants in the sub-minute to several minute ranges. Closed loop time constants of a minute or more are common. Similarly, level control is usually a slow process. Closed loop time constants of one minute for a small capacity tank might be common. On the other hand, the closed loop time constant for a large capacity tank might well be 30 minutes. For temperature and level control, the requirements for control valve dynamic response can be relaxed considerably, with a T of 2 to 12

86 seconds or longer being justifiable. However, in all of the above cases, there is an equal need for the control valve dynamics to be consistent, otherwise the result will be control loop instability, producing oscillations.

8.3 Testing considerations for small amplitude and medium amplitude dynamic response

(regions 2 and 3)

See discussion in 7.3.






- 37 - ANSI/ISA-TR75.25.02-2000 (R2010)

80 80

70

(%)

65

75

0.5% Steps 1% Steps 2% Steps 5% Steps 10% Steps

Actuator Position

I/P Input Signal

75

70

(%)

65

60 60

55 Flow Rate

3" Butterfly Valve / Diaphragm Actuator / Spool Valve Positioner

50

0 50 100 300 350 150 200

Time (seconds)

250

55

50

400

Figure 7 — Control valve Oscillations

8.4 Data presentation for regions 2 and 3

The time series for individual steps can be plotted, for example as shown in Figure 5. For the control valves shown here, 2% steps were sufficiently large that the actuator response represented the entire control valve response. Note the tendency of control valve C to move very slowly initially, making the definition of dead time ambiguous. Alternately, for region 2 behavior, “zooming in” on data such as in

Figure 3 would be appropriate. Figure 6 shows the relation between step size and T for bench tests on

86 four valves.

Labeling of graphs should state clearly what loading was applied to the control valve. For bench top testing, if comparisons are to be made of two control valves, or altered conditions within one control valve, it is helpful to have an independent measure of the total friction.

8.5 Design and maintenance factors at small amplitude and medium amplitude (regions 2 and 3)

For small steps in region 2 (i.e., of amplitude approaching the dead band), the important factors are the same as in region 1; friction, positioner design, drive train design, etc.

Actuator design is usually less important than positioner design. Positioner frequency response data can be used to predict comparative results on a given valve and actuator. However, standards for frequency response testing usually require it be measured without actuator load and at amplitudes well above the dead band (see ANSI/ISA-75.13.01-1996 (R2007)). The first condition excludes the desired realism, and the second condition excludes regions 1 and 2.

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---






ANSI/ISA-TR75.25.02-2000 (R2010) - 38 -

S t e m

P o s i t i o n

( % )

0

1 0 0

9 0

8 0

7 0

6 0

5 0

8 0

7 0

6 0

5 0

4 0

3 0

2 0

1 0

1 1 0

6 7 8 9 1 0 1 1 1 2 1 3 1 4 1 5 1 6

1 0 0

9 0

I /

S

P i g n

I n p u t a l

T i m e ( s e c o n d s )

B

C

1 0 0

A

S t a n d a r d A c t u a t o r s

F u l l

% p u

S t m e p p s o

A n 4 s p e e d

" G l a o w i t h n b d e

8 4 p

V s a i l d v e a t s w i t h

P o s i t i o n e r s s h u t o f f

4 0

3 0 B

C

2 0

1 0

0

3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1 4 2 4 3 4 4

T i m e ( s e c o n d s )

Figure 8 — Example of large amplitude or “stroking time” testing actuator

8.6 Oscillatory response

Figure 7 shows an example of control valve “hunting”, which ISA-51.1-1979 (R1993) defines as “an undesirable oscillation of appreciable magnitude, prolonged after external stimuli disappear”. The oscillation would be expected to continue with this frequency and amplitude while the controller output is constant. Probably all users would agree that the behavior in Figure 7 is unacceptable, unless the process stream passes into a large holding tank or a large gas volume. However, the importance may be relative; it is conceivable that hunting with amplitude below the dead band specification might be acceptable in any process loop.

It is important to understand that the hunting occurs without stimulus from the control loop. Therefore, this behavior is distinct from limit cycling of the entire control loop induced by finite dead band or resolution combined with integral action. Technically, however, the hunting is still a “limit cycle” within the control valve, not an underdamped linear oscillation, since it is dominated by nonlinear behavior. Various design factors can influence the amplitude and frequency of hunting. Tests with pause times of several minutes may be required to detect hunting.






- 39 - ANSI/ISA-TR75.25.02-2000 (R2010)

8.7 Performance near the closed position

Figure 8 includes an example, with control valve C, of irregular behavior near the seat. Oscillatory behavior is observed in this unbalanced, flow-down globe valve with process flow. At present, no standard procedure exists for testing near the seat; steps from 0 to 5% open have been suggested.

However, some rotary valves, including certain ball valve styles, do not pass any flow until the actuator opens approximately 10%. One could argue too that control valves should not be sized to routinely operate near the seat.

These issues should be considered when specifying the minimum nominal valve position as in 6.5.1.

9 Large amplitude dynamic response tests (region 4)

This clause covers input changes large enough to cause velocity limiting or relay saturation. This often occurs on step changes 10% or larger, although with some actuator and positioner combinations it occurs on 5% or even 2% steps.

9.1 Important measures in region 4

Stroking time is the most common measure; although it is not defined in present known standards, it is commonly understood to mean the time to move 100%. Vendor software programs will estimate stroking time for 100% steps. Measurement of actuator position is usually adequate since lost motion, which would be significant at small amplitudes, is not significant relative to a 100% move. Stroking time includes the dead time that occurs before any valve stem motion. Two possible cases exist and two sorts of tests are required to define these two cases.

9.1.1 Outside the throttling range

The output signal from some control systems is not limited to 0% to 100% as defined by

ANSI/ISA-50.00.01-1975 (R2002), "Compatibility of Analog Signals for Electronic Industrial Process

Instruments" as 4.0 to 20.0 mA dc. Actual outputs will range from something less than 4 mA up to something more than 20 mA if the control system becomes saturated. Pneumatic signals also may fall outside the standard 3 to 15 psig standard signal range, and the pneumatic signals intercepted by solenoid valves used for interlock control will typically supply 0 or 20 psig. If the control valve is driven to less than 0% or more than 100% for a sufficient time, the actuator pressure will reach either nearly 0% or almost 100% of the air supply pressure for the positioner. Valves without positioners will drive towards

0 psig or to controller or I/P (current to pressure converter) supply pressures. This situation frequently exists with interlock safety valves, and compressor anti-surge valves. Many ordinary applications will have the controller signal driven beyond the throttling range at times during operation. The dead time will be extended for any air added or vented beyond that required for stroke limits.

If the starting position of the valve test is specified to be "beyond fully open" or "beyond fully closed" then a considerable time may be required for the actuator to reach a fully saturated condition with no further flow of air into, or out of, the actuator.

9.1.2 Within the throttling range

The other situation is the stroking speed for changes within the normal throttling range. For many of the throttling control valve applications considered by ISA75.25, starting the large-amplitude tests slightly greater than zero travel and slightly below 100% travel may be more logical. Values in common use are

10% and 90% of signal.

Velocity-limited motion, by definition, tends to follow a straight line of motion with time. The term time constant, which is associated with exponential response, therefore has no meaning in region 4.






ANSI/ISA-TR75.25.02-2000 (R2010) - 40 -


Applications fall into three major categories. Typical stroking time requirements for each of these applications depend on the application.

9.2.1 Emergencies

In a process emergency causing a plant shutdown the control valve may be expected to move rapidly to the fully closed position to contain the process (e.g., to protect downstream equipment or the environment, or to avoid feeding a fire). Or, for another application, the valve may need to fully open to provide maximum coolant flow or vent a process.

9.2.2 Compressor surge control or pump minimum flow re-circulation valves

Effective surge control requires that the antisurge recycle valve respond to large disturbances quickly.

Typically, the control valve must move from full closed to the throttling position within a few seconds. The actual required time is a function of compressor design and installation. This information will come from the manufacturer or by test. With some anti-surge controllers, the valve must step open a predetermined amount when surge is detected. In addition to step responses, the surge controller may have a standard

PID control system to keep the compressor in operation beyond the surge control condition.

Pneumatic actuators may require volume boosters to increase the effective capacity of the positioner.

The volume boosters are installed between the positioner and actuator, have their own air supply/filter, and are adjusted to open only for fast changes in input signal. Volume boosters installed in parallel must be adjusted so that they all have the same characteristics. It is important to work closely with the compressor and anti-surge control vendor to understand and test the control valve for the particular application. Valves used to maintain a minimum pump re-circulation often operate closed until the forward flow is blocked (i.e., batch control or a safety valve). These valves must respond quickly to maintain a minimum pump flow in order to prevent pump damage.

Other examples of large step change requirements may be valves that perform a deluge function such as

“killing” an exothermic reaction by opening the valve. Applications using on/off valves for this purpose are not considered in this report.

9.3 Testing considerations for large amplitude dynamic response

9.3.1 Tests without process load in a plant instrument shop or control valve manufacturing site

Large step tests can be done to verify correct installation of all the pneumatic hardware, including sizing of solenoid valves and tuning of volume boosters.

9.3.2 Laboratory testing with flow, simulating a plant process

Large step tests can be done easily in the laboratory. Addition of process flow may reveal surprising behavior near the seat (e.g., control valve C in Figure 8).

9.3.3 In-process testing during plant operation

The commissioning of surge control startups often includes a very cautious approach to the predicted surge conditions in order to determine the actual surge point characteristics. Data collected on the control system performance may be limited. Some valve testing with the compressor not running is possible. For batch control applications testing is usually more convenient.






- 41 - ANSI/ISA-TR75.25.02-2000 (R2010)

See, for example, Figure 8. Velocity limiting is difficult to quantify; for example, control valve A followed a straight line when closing but not when opening. Because of this non-symmetrical response, quantifying the size of region 4 is not recommended.

9.4 Design and maintenance factors important at large amplitude

For pneumatic actuators, in region 4 the actuator volume and the air flow in the pneumatic system become the dominant features. Various factors involved are discussed in Clause 5.

10 References

NOTE — For the referenced standards the user should refer to the current versions.

1) ANSI/ISA-75.05.01-2000 (R2005), Control Valve Terminology.

2) ISA-51.1-1979 (R1993), Process Instrumentation Terminology.

3) McGraw-Hill Dictionary of Scientific and Technical Terms, sixth edition, 2002.

4) Control Valve Dynamic Specification, published by EnTech Control Inc., Toronto.

5) ANSI/ISA-75.13.01-1996 (R2007), Method of Evaluating the Performance of Positioners with Analog

Input Signals and Pneumatic Output.

6) Shinskey, F.G., Process Control Systems, 4th edition, McGraw-Hill, 1994.

7) Coughran, M.T., “Measuring the Installed Dead Band of Control Valves,” ISA TECH/97-1114.

8) Champagne, R.P. and Boyle, S.J., “Optimizing valve actuator parameters to enhance control valve performance,” ISA Transactions 35 (1996) 217-223.

9) Van De Vegte, J., Feedback Control Systems, 2nd edition, Prentice Hall, 1990.

10) Gibson, J.E., Nonlinear Automatic Control, McGraw-Hill, 1963.

11) Goldfarb, M. and Celanovic, N., Modeling Piezoelectric Stack Actuators for Control of

Micromanipulation, IEEE Control Systems, June 1997, p. 69.

12) Dahlin, E.B., Designing and Tuning Digital Controllers, Instruments & Control Systems, Vol. 41, June

1968, p. 77.

13) Morari, M. and Zafiriou E., Robust Process Control, Prentice Hall, 1989.

14) Levine, W.S., Editor, The Control Handbook, CRC Press & IEEE Press, 1966, on “Lambda Tuning and IMC,” Astrom, K.J. p. 821, Bialkowski, W.L. p. 1232.

15) Senior, K.A. "Technical Guidelines and Design Information, Using KVSP Packing Systems For

Improving Process Control and Minimizing Fugitive Emissions," white paper, DuPont Dow L.L.C.

16) Senior, K.A. "Valve Packing Systems Improve Process Control," Chemical Processing, June 1997.

17) Brestal, R. et al, "Control Valve Packing Systems," technical monograph 38, 1992. Fisher Controls,

Marshalltown, IA.




--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---



ANSI/ISA-TR75.25.02-2000 (R2010) - 42 -

18) "Packing Selection Guidelines for Sliding Stem Valves", Bulletin 59.0:062, Fisher-Rosemount,

Marshalltown, IA.

19) Langford, C.G. "A Users View of Process Control and Control Valve Positioners" ISA, 1996, Paper

#96-076 1054-0032/721-728.

--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---









--`,`,`````,,``,,,`,`,`,,`,`,``,-`-`,,`,,`,`,,`---



Developing and promulgating sound consensus standards, recommended practices, and technical reports is one of ISA’s primary goals. To achieve this goal the Standards and Practices Department relies on the technical expertise and efforts of volunteer committee members, chairmen and reviewers.

ISA is an American National Standards Institute (ANSI) accredited organization. ISA administers United

States Technical Advisory Groups (USTAGs) and provides secretariat support for International

Electrotechnical Commission (IEC) and International Organization for Standardization (ISO) committees that develop process measurement and control standards. To obtain additional information on the

Society’s standards program, please write:

ISA

Research Triangle Park, NC 27709

ISBN: 978-1-936007-50-9






ISA-TR75.25.02-2000 (R2010)

ANSI/ISA-TR75.25.02-2000 (R2010)

Control Valve Response Measurement from Step Inputs

Preface

Contents

Foreword

Abstract

Key Words

2 Scope

3 Definitions

b a a < resolution

b

c d c

dead band < d

Figure 1 —Dead band and resolution

4 Control valve response

Table 1 — Control valve output measurements and test environments

5 Maintenance and design issues affecting process control

6 Process and control design issues

Figure 2 — Example of step response test with lost motion

7 Static behavior tests

Figure 3 — Example of a test using a series of small steps

Figure 4 — Step response test 0.5% steps 7 - 8% dead band

8 Small amplitude and medium amplitude dynamic response tests (regions 2 and 3)

Figure 5 — Example graph of time series tests showing step response times

Figure 6 — Step response for four valves

Figure 7 — Control valve Oscillations

Figure 8 — Example of large amplitude or “stroking time” testing actuator

9 Large amplitude dynamic response tests (region 4)

10 References

Products

Support

ISA-TR75.25.02-2000 (R2010)

ANSI/ISA-TR75.25.02-2000 (R2010)

Control Valve Response Measurement from Step Inputs

Preface

Contents

Foreword

Abstract

Key Words

2 Scope

3 Definitions

b a a < resolution

b

c d c

dead band < d

Figure 1 —Dead band and resolution

4 Control valve response

Table 1 — Control valve output measurements and test environments

5 Maintenance and design issues affecting process control

6 Process and control design issues

Figure 2 — Example of step response test with lost motion

7 Static behavior tests

Figure 3 — Example of a test using a series of small steps

Figure 4 — Step response test 0.5% steps 7 - 8% dead band

8 Small amplitude and medium amplitude dynamic response tests (regions 2 and 3)

Figure 5 — Example graph of time series tests showing step response times

Figure 6 — Step response for four valves

Figure 7 — Control valve Oscillations

Figure 8 — Example of large amplitude or “stroking time” testing actuator

9 Large amplitude dynamic response tests (region 4)

10 References

Add this document to collection(s)

Add this document to saved

Suggest us how to improve StudyLib