Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

TR 752502

Download as pdf or txt
Download as pdf or txt
You are on page 1of 40

TECHNICAL REPORT

ANSI/ISA–TR75.25.02–2000

Control Valve Response


Measurement from Step
Inputs

Approved 11 March 2001


TM

ISA–The Instrumentation,
Systems, and
Automation Society
ANSI/ISA–TR75.25.02–2000
Control Valve Response Measurement from Step Inputs

ISBN: 1-55617-743-7

Copyright  2000 by ISA–The Instrumentation, Systems, and Automation Society. 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
USA
—3— ANSI/ISA–TR75.25.02–2000

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.

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–The Instrumentation, Systems, and
Automation Society, 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 ADHERES TO THE POLICY OF THE AMERICAN NATIONAL STANDARDS


INSTITUTE WITH REGARD TO PATENTS. IF ISA IS INFORMED OF AN EXISTING PATENT THAT IS
REQUIRED FOR USE OF THE STANDARD, IT WILL REQUIRE THE OWNER OF THE PATENT TO
EITHER GRANT A ROYALTY-FREE LICENSE FOR USE OF THE PATENT BY USERS COMPLYING
WITH THE STANDARD OR A LICENSE ON REASONABLE TERMS AND CONDITIONS THAT ARE
FREE FROM UNFAIR DISCRIMINATION.

EVEN IF ISA IS UNAWARE OF ANY PATENT COVERING THIS STANDARD, THE USER IS
CAUTIONED THAT IMPLEMENTATION OF THE STANDARD MAY REQUIRE USE OF TECHNIQUES,
PROCESSES, OR MATERIALS COVERED BY PATENT RIGHTS. ISA TAKES NO POSITION ON THE
EXISTENCE OR VALIDITY OF ANY PATENT RIGHTS THAT MAY BE INVOLVED IN IMPLEMENTING
THE STANDARD. ISA IS NOT RESPONSIBLE FOR IDENTIFYING ALL PATENTS THAT MAY
REQUIRE A LICENSE BEFORE IMPLEMENTATION OF THE STANDARD OR FOR INVESTIGATING
THE VALIDITY OR SCOPE OF ANY PATENTS BROUGHT TO ITS ATTENTION. THE USER SHOULD
CAREFULLY INVESTIGATE RELEVANT PATENTS BEFORE USING THE STANDARD FOR THE
USER’S INTENDED APPLICATION.
ANSI/ISA–TR75.25.02–2000 —4—

HOWEVER, ISA ASKS THAT ANYONE REVIEWING THIS STANDARD WHO IS AWARE OF ANY
PATENTS THAT MAY IMPACT IMPLEMENTATION OF THE STANDARD 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 following people served as members of ISA Subcommittee SP75.25:

NAME COMPANY

C. Langford, Chairman Cullen G. Langford, Inc.


W. Weidman, Managing Director Parsons Energy & Chemicals Group
J. Beall Eastman Chemical Co.
D. Bennett Sampson Controls, Inc.
W. Bialkowski Entech Control Engineering, Inc.
W. Black Cashco, Inc.
M. Boudreaux Exxon Mobil Chemical
S. Boyle Neles Automation
D. Buchanan Union Carbide Corporation
F. Cain Flowserve Corporation
N. Cammy UOP LLC
R. Champagne Consultant
J. Jamison Bantrel Inc.
S. Kempf Harold Beck & Sons, Inc.
P. Maurath Procter & Gamble Company
R. McEver Bettis Corporation
N. McLeod Elf Atochem
G. McMillan Solutia, Inc.
J. Reed Norriseal

The following people served as members of ISA Committee SP75:

NAME COMPANY

D. Buchanan, Chairman Union Carbide Corporation


W. Weidman, Managing Director Parsons Energy & Chemicals Group
A. Abromaitis Red Valve Company, Inc.
H. Backinger J. F. Kraus & Company
G. Barb Retired
H. Baumann H B Services Partners LLC
H. Boger Masoneilan Dresser
G. Borden Consultant
S. Boyle Neles Automation
R. Brodin Fisher Controls International, Inc.
F. Cain Flowserve Corporation
C. Corson Fluor Daniel Inc.
M. Coughran Fisher Controls
—5— ANSI/ISA–TR75.25.02–2000

A. Engels Praxair, Inc.


H. Fuller Valvcon Corporation
J. George Richards Industries
A. Glenn Flowserve Corp.
L. Griffith Consultant/Retired
B. Guinon Shell Chemical
F. Harthun Retired
B. Hatton DeZurik Division Unit
J. Jamison Bantrel, Inc.
R. Jeanes TXU Electric
J. Kersh M. W. Kellogg Company
C. Langford Cullen G. Langford, Inc.
A. Libke DeZurik Valve Company
R. Louviere Creole Engineering Sales Company
O. Lovett Consultant/Retired
A. McCauley Chagrin Valley Controls, Inc.
R. McEver Bettis Corporation
H. Miller Control Components, Inc.
T. Molloy CMES Inc.
L. Ormanoski Frick Company
J. Ozol Commonwealth Edison
W. Rahmeyer Utah State University
J. Reed Norriseal
K. Schoonover Con-Tek Valves, Inc.
A. Shea Copes-Vulcan, Inc.
E. Skovgaard Leslie Controls, Inc.
H. Sonderegger Tyco Flow Control
R. Terhune Retired

This standard was approved for publication by the ISA Standards and Practices Board on
30 December 2000:

NAME COMPANY

M. Zielinski, Vice President Fisher-Rosemount Systems, Inc.


D. Bishop Consultant
P. Brett Honeywell, Inc.
M. Cohen Senior Flexonics, Inc.
M. Coppler Ametek, Inc.
B. Dumortier Schneider Electric
W. Holland Southern Company
A. Iverson Ivy Optiks
R. Jones Dow Chemical Co.
V. Maggioli Feltronics Corp.
T. McAvinew Bateman Engineering, Inc.
A. McCauley, Jr. Chagrin Valley Controls, Inc.
G. McFarland Westinghouse Process Control Inc.
D. Rapley Rapley Consulting Inc.
R. Reimer Rockwell Automation
J. Rennie Factory Mutual Research Corp.
H. Sasajima Yamatake Corp.
R. Webb Altran Corp.
W. Weidman Parsons Energy & Chemicals Group
ANSI/ISA–TR75.25.02–2000 —6—

J. Weiss EPRI
J. Whetstone National Institute of Standards & Technology
M. Widmeyer EG&G Defense Materials
R. Wiegle CANUS Corp.
C. Williams Eastman Kodak Co.
G. Wood Graeme Wood Consulting
—7— ANSI/ISA–TR75.25.02–2000

Contents

1 Purpose........................................................................................................................................... 9
2 Scope .............................................................................................................................................. 9
3 Definitions ....................................................................................................................................... 9
4 Control valve response ................................................................................................................. 14
4.1 Measurement of control valve response ............................................................................... 14
4.2 System response .................................................................................................................. 14
4.3 Test environments used to determine control valve response.............................................. 16
4.4 Size of input signal change – regions ................................................................................... 17
5 Maintenance and design issues affecting process control............................................................ 17
5.1 Stem seal .............................................................................................................................. 18
5.2 Valve seat shutoff.................................................................................................................. 18
5.3 Valve seat type...................................................................................................................... 19
5.4 Process fluid effects .............................................................................................................. 19
5.5 Mechanical tolerances .......................................................................................................... 19
5.6 Structural stiffness................................................................................................................. 19
5.7 Pneumatic positioner............................................................................................................. 19
5.8 Actuator size/type.................................................................................................................. 20
5.9 Electric and hydraulic actuators ............................................................................................ 20
5.10 Flow effects ......................................................................................................................... 20
5.11 Valve sizing and selection................................................................................................... 20
6 Process and control design issues................................................................................................ 20
6.1 Control loop process gain – range and variability ................................................................. 20
6.2 Over-sizing ............................................................................................................................ 22
6.3 Control valve inherent characteristic ..................................................................................... 22
6.4 Closed loop performance – control valve dynamic specification........................................... 23
6.5 Nonlinear regions .................................................................................................................. 24
7 Static behavior tests...................................................................................................................... 25
7.1 Important measures of static behavior .................................................................................. 25
7.2 Applications affected and classes of performance................................................................ 25
7.3 Testing considerations .......................................................................................................... 26
7.4 Data presentation.................................................................................................................. 28
7.5 Design and maintenance factors important to static behavior .............................................. 29
ANSI/ISA–TR75.25.02–2000 —8—

8 Small amplitude and medium amplitude dynamic response tests (regions 2 and 3) .................... 29
8.1 Important measures for regions 2 and 3 ............................................................................... 30
8.2 Applications affected and classes of performance................................................................ 32
8.3 Testing considerations for small amplitude and medium amplitude dynamic response
(regions 2 and 3) ................................................................................................................... 32
8.4 Data presentation for regions 2 and 3................................................................................... 33
8.5 Design and maintenance factors at small amplitude and medium amplitude
(regions 2 and 3) ................................................................................................................... 33
8.6 Oscillatory response.............................................................................................................. 34
8.7 Performance near the closed position................................................................................... 34
9 Large amplitude dynamic response tests (region 4) ..................................................................... 35
9.1 Important measures in region 4 ............................................................................................ 35
9.2 Applications affected and classes of performance................................................................ 36
9.3 Testing considerations for large amplitude dynamic response ............................................. 36
9.4 Design and maintenance factors important at large amplitude ............................................. 37
10 References.................................................................................................................................... 37
—9— ANSI/ISA–TR75.25.02–2000

1 Purpose
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, "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, "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). 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, typical examples include a positioner, transducer, signal booster relay, air set,
snubber, etc.

3 Definitions
This document and ANSI/ISA-75.25.01-2000 make use of terms as defined in ANSI/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
ANSI/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 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 [ANSI/ISA-51.1-1979 (R1993)]. Sometimes also referred to as slop,
lost motion, or free play.
ANSI/ISA–TR75.25.02–2000 — 10 —

Output

b
a c
d
a < resolution ≤ b

Input c ≤ dead band < d


Amplitude

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.

3.3 dead band:


the range through which an input signal may be varied, with reversal of direction, without initiating an
observable change in output signal [ANSI/ISA-S51.1-1979 (R1993)]. In this technical report and in
standard ANSI/ISA-75.25.01-2000 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 (Td):


the time after the initiation of an input change and before the start of the resulting observable response.

3.5 dead zone:*


a zone of input for which no value of the output exists [ANSI/ISA-S51.1-1979 (R1993)].

3.6 dynamic response:


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", fifth edition, 1994]. Dynamic means that the control valve is
— 11 — ANSI/ISA–TR75.25.02–2000

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.

3.7 gain ratio GZ / GZ02:


the response gain GZ divided by the response gain GZ02 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.

GR = GZ/GZ02

3.8 hunting:*
an undesirable oscillation of appreciable magnitude, prolonged after external stimuli disappears
[ANSI/ISA-S51.1-1979 (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.

3.9 limit cycle:


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 ed, Prentice Hall, 1990, p14]. See also
hunting [ANSI/ISA-S51.1-1979 (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.

3.11 nonlinear system:*


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 ANSI/ISA-
S51.1-1979 (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
ANSI/ISA-S51.1-1979 (R1993). Usually expressed as a percentage of the full change in steady state
value.

3.14 position Z:
the position of the closure member relative to the seated position. In this technical report and in standard
ANSI/ISA-75.25.01-2000 expressed as a percent of span.
ANSI/ISA–TR75.25.02–2000 — 12 —

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.

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, the step response can be stem position, flow, or another process variable.

3.17 response flow coefficient CvR:*


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.

3.18 response gain GZ:


the ratio of the steady state magnitude of the process change ∆Z divided by the signal step ∆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.

GZ = DZ / Ds

GZ02 = DZ02 / Ds02

3.19 sampling interval Dts:


the time increment between sampled data points. It is the inverse of the sampling rate, f0. ∆tS = 1/f0. As
used in this technical report and ANSI/ISA-75.25.01-2000, 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 ∆t.

3.20 sampling rate f0:


the rate at which data samples are taken or the number of samples per unit time. See sampling interval.

3.21 shaft windup:*


in rotary valve systems, the tendency of the drive shaft to twist under load while the closure member is
stuck at a given position.

3.22 sliding friction Fr or Tr:


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", fifth edition,
1994]; 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.
— 13 — ANSI/ISA–TR75.25.02–2000

3.24 steady state:


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.

3.25 step change:


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 ∆tsc:


the time between the start of a signal input step and when it reaches its maximum value.

3.27 step test:


the application of a step change to an input signal in order to test the step response dynamics.

3.28 step response time (T86):


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 size ∆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 [ANSI/ISA-S51.1-1979 (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).

3.32 stick/slip cycle:*


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.

3.33 time constant τ:


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 ANSI/ISA-75.25.01-2000 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.

GX = DX / Ds
ANSI/ISA–TR75.25.02–2000 — 14 —

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, T86. The use of the approximate time
constant in no way implies that the response of the control valve is first order. The step response of the
control valve is typically complex, having dead time initially, followed by potentially complex dynamics
before the steady state is achieved. T86 includes the dead time in the initial part of the response, as well
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 T86 that is dead time increases, this approximation becomes less ideal.

3.36 velocity limiting:*


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

3.37 wait time ∆tw:*


the time spent after a step input change waiting for the response to come to the new steady state value.

3.38 X-Y plot:*


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.

4.2 System response

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.
— 15 — ANSI/ISA–TR75.25.02–2000

4.2.1 Flow coefficient

The fundamental parameter is the valve flow coefficient. The flow coefficient, or Cv, is calculated from the
flow through the valve, the fluid density and the differential pressure across the valve [see ISA-75.01-1985
(R1995), “Flow Equations for Sizing Control Valves” and ANSI/ISA-75.02-1996, “Control Valve Capacity
Test Procedure”]. For in-process testing, a response flow coefficient C vR, is defined in this technical report.
The response flow coefficient CvR is calculated 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 Cv. This coefficient 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.

4.2.3 Stem position

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 stem1 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.

______
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 — 16 —

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

c) In-process test during plant operation

Table 1 — Control valve output measurements and test environments

Test Environment Parameter Recommended for Measurement


(listed order of preference within each environment)
Static Performance Dynamic Performance
Measures - Resolution, dead Measures - Dead time, response
band, and valve travel gain. time, and overshoot.

Bench test Stem position (use only if a Stem position


Tested on a bench, in a plant, a laboratory or in process test is
lab, a shop, etc. with no flow not available)
(may be under applied pressure),
with valve completely assembled,
and with recommended packing
load.

Laboratory test Flow coefficient for dead band Stem position


Tested in a laboratory flow loop and resolution
with flow Flow coefficient
Process response, where
possible Process response
In-process test Flow coefficient (if Stem position
Tested in the actual process measurements available)
application Flow coefficient (if
Process response, where measurements available)
possible
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.
— 17 — ANSI/ISA–TR75.25.02–2000

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 document, 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.

4.4.1 Region 1

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

4.4.2 Region 2

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.

4.4.3 Region 3

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.
ANSI/ISA–TR75.25.02–2000 — 18 —

5.1 Stem seal

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 16 through 19).

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

e) project costs.

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.
— 19 — ANSI/ISA–TR75.25.02–2000

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.

5.5 Mechanical tolerances

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.

5.6 Structural stiffness

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.

5.7.1 Tubing/fitting capacity

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.

5.7.2 Actuator volume

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
ANSI/ISA–TR75.25.02–2000 — 20 —

be increased. The times for opening and for closing the valve will probably differ from each other. See
ANSI/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.

5.8 Actuator size/type

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
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.

5.10 Flow effects

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.
— 21 — ANSI/ISA–TR75.25.02–2000

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

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)


ANSI/ISA–TR75.25.02–2000 — 22 —

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.

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
— 23 — ANSI/ISA–TR75.25.02–2000

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 13 through 15), 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.

70

65
Flow Rate
60
Actuator Position
55
(%)
50 Input Signal

45
0.5% Steps 1% Steps 2% Steps 5% Steps 10% Steps
40
4-Inch Segmented Ball Valve with Diaphragm Actuator & Positioner
35
0 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 86, approximate time constant t'

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
ANSI/ISA–TR75.25.02–2000 — 24 —

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 T86.

6.4.2 Impact of dead time on loop stability

The step response time (T86) is made up of two components, the dead time (Td) and the remainder of 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, T86, is used to represent the total step
response time of the control valve. It remains desirable to minimize the dead time portion of T86.

6.5 Nonlinear regions

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 86 is 40% of the closed loop time constant, or 4 seconds, hence the
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 T86 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 t' of 0.2 seconds, and a T86 of 0.4 seconds. Other
— 25 — ANSI/ISA–TR75.25.02–2000

applications exist where the speed of response of the control valve is not an issue, and it may be
acceptable to specify a t’ of as long as 10 seconds, and a T86 of 20 seconds. Such a specification may
allow the selection of a less expensive valve actuator. A t’ in the 1 to 3 second range, and a T86 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 for the standard testing procedures.

Clause 5 of ANSI/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 ANSI/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 systems2.

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,

______
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 thorough discussion of test methods in ANSI/ISA-51.1-1979 (R1993) and
ISA-75.13-1996.

3
The meaning of dead band – measured when reversing direction – is explicit in ANSI/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 ANSI/ISA-51.1-1979 (R1993).
ANSI/ISA–TR75.25.02–2000 — 26 —

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
affects only the period of the cycle, not the amplitude. The period can range from fractions of a minute to
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 establishing the desired
process gain also contribute to the importance of static behavior in these applications. Advanced Process
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.

7.3 Testing considerations

7.3.1 General considerations

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.
— 27 — ANSI/ISA–TR75.25.02–2000

6 0 .0 6 2 .0
V a lv e X / A c t u a t o r Y / P o s it io n e r Z , p a c k in g p e r in s t r u c t io n s
5 9 .5 6 1 .5
T e s te d a t n o m in a l 6 0 % o p e n , 6 0 0 g p m , 3 8 p s id
5 9 .0 6 1 .0
In p u t
5 8 .5 0 .1 % ≤ D e a d B a n d < 0 .3 % 6 0 .5
S ig n a l A c tu a to r
5 8 .0 n o re s p o n s e 6 0 .0
(% ) u p o n re v e rs a l P o s itio n
5 7 .5 5 9 .5 (% )
0 . 1 % < R e s o lu t io n ≤ 0 . 3 %
5 7 .0 5 9 .0
5 6 .5 n o r e s p o n s e w h ile c o n - 5 8 .5
t in u in g in s a m e d ir e c t io n
5 6 .0 5 8 .0
900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300

T im e ( s e c o n d s )

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 ISA-75.13-1996, "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.

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.
ANSI/ISA–TR75.25.02–2000 — 28 —

61.5

58.6
Stem Position

55.6

52.6 Input Signal

49.7

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-regulating4 (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.

7.4 Data presentation

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 ANSI/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.

The labels on figure 3 show data interpretation according to ANSI/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 ANSI/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.

______
4 Defined in ANSI/ISA-51.1-1979 (R1993)
— 29 — ANSI/ISA–TR75.25.02–2000

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 — 30 —

Time (seconds)
158 159 160 161 162 163 164 165 166 167 168
68
A

67 B
C

66

I/P Input Signal


65
2% Steps on 4-Inch Globe Valves with Standard Actuators and Positioners
o
Stem Nominal Q = 600 gpm, ∆P = 38 psid, water T = 120 F
Position Packing maintained per instructions
(%)
67
C
A
66

65 B

64
178 179 180 181 182 183 184 185 186 187

Time (seconds)

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% steps5. 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 standards ISA-75.13-1996 and
ISA-26-1968. 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 SP75.25, 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.
— 31 — ANSI/ISA–TR75.25.02–2000

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, Td, is commonly measured after a step input. Tests show that dead times can be
much longer with the slowly changing controller output signals typical of many control systems. Whether
the control valve will overshoot depends on controller signals.

100

Response, Seconds

10

0.1
0.1 1 10
Step Size, Percent

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. T98, for example, is
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 tolerance6 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 — 32 —

8.2 Applications affected and classes of performance

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 T86 to be in the range of 0.1 to 0.2 seconds. Control of liquid flows often results in desired closed
loop time constants of about 10 seconds. A control valve T86 of about 1 to 3 seconds would be 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 T86 of 2 to 12 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)

8.3.1 General considerations

See discussion in 7.3.


— 33 — ANSI/ISA–TR75.25.02–2000

80 80

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


75 75

Actuator Position
70 70

(%) (%)
65 I/P Input Signal 65

60 60

55 Flow Rate 55

3" Butterfly Valve / Diaphragm Actuator / Spool Valve Positioner

50 50
0 50 100 150 200 250 300 350 400

Time (seconds)

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 T86 for bench tests on 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 ISA-75.13-1996). The first condition excludes the desired realism, and the second
condition excludes regions 1 and 2.
ANSI/ISA–TR75.25.02–2000 — 34 —

T ime ( se conds)
6 7 8 9 10 11 12 13 14 15 16
1 10
1 00
90
80 I/P Input B
70 Signal C
60 A
50
40
30
20
St em 10
Po sit ion 0 10 0% S t ep s on 4" Glob e Val v es wit h
(% ) 1 00 S ta nda rd A ct uat or s and Po sit ion ers
Full pump sp eed with 8 4 psid at shuto ff
90
80
A
70
60
50
40
C
30 B
20
10
0
34 35 36 37 38 39 40 41 42 43 44

T ime ( se conds)

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 ANSI/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.

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
— 35 — ANSI/ISA–TR75.25.02–2000

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-50.1-1982 (R1992), "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 SP75.25, 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 — 36 —

9.2 Applications affected and classes of performance

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.

9.2.3 Batch control

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
— 37 — ANSI/ISA–TR75.25.02–2000

system performance may be limited. Some valve testing with the compressor not running is possible. For
batch control applications testing is usually more convenient.

9.3.4 Data presentation

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, Control Valve Terminology.

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

3) McGraw-Hill Dictionary of Scientific and Technical Terms, fifth edition, (1994).

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

5) ISA-75.13-1996, Method of Evaluating the Performance of Positioners with Analog Input Signals
and Pneumatic Output.

6) Shinskey, F.G., Process Control Systems, 4th ed., 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) ISA-26-1968 and ANSI MC4.1-1975, Dynamic Response Testing of Process Control


Instrumentation.

10) Van De Vegte, J., Feedback Control Systems, 2nd ed, Prentice Hall, 1990

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

12) Goldfarb, M, and Celanovic, N,. Modeling Piezoelectric Stack Actuators for Control of
Micromanipulation, IEEE Control Systems, June 1997, p69

13) Dahlin, E. B., Designing and Tuning Digital Controllers, Instruments & Control Systems- Vol 41,
June 1968, p77

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

15) Levine, W. S, (Editor) The Control Handbook, CRC Press & IEEE Press, 1966, -- On “Lambda
Tuning and IMC” Astrom, K. J. p821, Bialkowski, W. L. p 1232.
ANSI/ISA–TR75.25.02–2000 — 38 —

16) 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.

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

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

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

20) Langford, C. G. "A Users View of Process Control and Control Valve Positioners" ISA, 1996 - Paper
#96-076 1054-0032/721-728
This page intentionally left blank.
Developing and promulgating technically 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
Attn: Standards Department
67 Alexander Drive
P.O. Box 12277
Research Triangle Park, NC 27709

ISBN 1-55617-743-7

You might also like