Valve Positioner Principles and Maintenance Guide

Valve Positioner Valve Positioner Principles Principles and Maintenance Guide

N

S E D

E

WARNING: WARNING: Please read the License Agreement A on the back cover before removing R the Wrapping Material. E L

C I

Effective December 6, 2006, this report has been made publicly available in accordance with Section 734.3(b)(3) and published in accordance with Section 734.7 of the U.S. Export Administration Regulations. As a result of this publication, this report is subject to only copyright protection and does not require any license agreement from EPRI. This notice supersedes the export control restrictions and any proprietary licensed material notices embedded in the document prior to publication.

I

L

M

A T

Reduced Cost

Technical Report

Plant Maintenance Support

Equipment Reliability

Valve Positioner Principles and Maintenance Guide 1003091

Final Report, December 2001

EPRI Project Manager L. Loflin

EPRI  3412 Hillview Hillview Avenue, Palo Alto, Alto, California 94304 • PO Box 10412, Palo Alto, California California 94303 • USA 800.313.3774 • 650.855.2121 • [email protected] • www.epri.com

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM: (A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR (B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT. ORGANIZATION(S) THAT PREPARED THIS DOCUMENT EPRI

ORDERING INFORMATION Requests for copies of this report should be directed to EPRI Customer Fulfillment, 1355 Willow Way, Suite 278, Concord, CA 94520, (800) 313-3774, press 2. Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Electric Power Research Institute, Inc. Copyright © 2001 2001 Electric Power Resea Research rch Institute, Inc. All rights reserved.

CITATIONS This report was prepared by Nuclear Maintenance Applications Center (NMAC) EPRI 1300 W.T. Harris Boulevard Charlotte, NC 28262 This report describes research sponsored by EPRI. The report is a corporate document that should be cited in the literature in the following manner: Valve Positioner Principles and Maintenance Guide, EPRI, Palo Alto, CA: 2001. 1003091.

iii

REPORT SUMMARY

This guide provides information for personnel involved with the maintenance of valve positioners, including the principles of operation, applications, calibration, condition monitoring, preventive maintenance, and troubleshooting. It provides insights for experienced personnel as well as basic information, guidance, and instructions for personnel assigned to maintain valve positioners.

Background A valve positioner is a device in the control loop of a flow, pressure, or level control process that improves valve response to changes in the demand signal from a process controller. The positioner is used to limit control valve dead band, mitigate friction-induced nonlinearities, change valve flow characteristics, permit double-acting actuator operation, increase shutoff forces, allow for split-ranging, and add loop gain to decrease the effects of process lag and dead band. In 1999, NMAC conducted a survey of unplanned capacity loss factors. The survey identified control valves as the number four cause of such losses. An EPRI survey identified the existence of information on control valves, but none for control loops and positioners. Another survey performed in the paper and pulp industry showed that, of 31 valve control problems, 71% were attributable to the positioner. The next most frequent problem cause was bench set at 38%. Based on this information, the positioner as part of the control loop was selected for further research. To provide additional information, a tutorial on control loops was also developed.

Objectives To help power plant maintenance personnel understand the basic principles of positioner designs and application To provide technical information for plant maintenance personnel on proper calibration, condition monitoring/preventive maintenance, and troubleshooting To provide additional technical information on control loops and how the positioner acts within the control loop

Approach A detailed review of industry literature, product information, and standards was conducted to identify the various designs, applications, and maintenance practices associated with valve positioners. Utility and industry personnel were surveyed to determine specific problems and commonly encountered failure mechanisms. Based on this information, recommendations were made on proper calibration, condition monitoring/preventive maintenance, and troubleshooting.

v

Results This guide presents a thorough discussion of valve positioner design and principles of operation to provide a technical background for personnel performing valve positioner maintenance and information about how positioners are applied. Subsequent chapters contain the maintenance portion of the guide and discussions on proper calibration, condition monitoring/preventive maintenance, and troubleshooting. The emphasis of the maintenance section is not only on good practices, but also on how each of these maintenance areas is closely related and mutually supportive. The contents of this guide are intended to assist plant personnel in reducing costs and equipment unavailability and in improving equipment reliability and performance.

EPRI Perspective Based on industry studies, the valve positioner is the component that causes most control loop problems. Nearly all positioner problems are the result of improper setup and maintenance. This guide provides maintenance personnel with details of the basic principles of positioner design and application as well as calibration, condition monitoring/preventive maintenance, and troubleshooting. An appendix is also included that provides technical information on control loops and the function of the positioner within the control loop.

Keywords Maintenance Control loop Control valve Positioner Calibration Troubleshooting

vi

EPRI Licensed Material

ABSTRACT A valve positioner is a device in the control loop of a flow, pressure, or level control process that improves valve response to changes in the demand signal from a process controller. In 1999, the EPRI Nuclear Maintenance Applications Center (NMAC) conducted a survey of unplanned capacity loss factors. The survey identified control valves as the number four cause of such losses. A subsequent survey identified the existence of information on control valves, but none for control loops and positioners. Based on these surveys and other industry information, the valve positioner as part of the control loop was selected for further research. This guide presents a thorough discussion of valve positioner design and principles of operation in the context of control loop principles. It provides a technical background for personnel performing valve positioner maintenance and information about how positioners are applied. Subsequent chapters contain the maintenance portion of the guide with discussions on proper calibration, condition monitoring/preventive maintenance, and troubleshooting. The emphasis of the maintenance section is not only on good practices, but also on how each of these maintenance areas is closely related and mutually supportive. The contents of this guide are intended to assist plant personnel to reduce costs and equipment unavailability and to improve equipment reliability and performance.

vii


ACKNOWLEDGMENTS The following individuals were active members of the Positioner Maintenance Guide Task Advisory Group. They made significant contributions to the development of this document by attending Task Advisory Group meetings and reviewing and providing feedback on various drafts of the guide. *Jim Allan Steve Ball *Bill Bowyer *Harry Cole *Randy Croxton *Scott Dill George Farley Scott Ladd Chuck Linden Marie Murphy *Bill Muscia *Mike Sawaya *Bill Slover Fred Wiens

Diablo Canyon Seabrook Vogtle Point Beach Palo Verde Salem Susquehanna Prairie Island Ft. Calhoun Cooper Beaver Valley Carolina P&L EPRI NMAC South Texas Project

*Attended TAG meeting July 18–19, 2001. Also acknowledged is George Gassman, Senior Research Specialist, Final Control Systems, Fisher Controls International, a division of Emerson Processes. Gassman provided technical insight in the principles of positioner operations. NMAC was supported in this effort by Bill Slover.

ix


CONTENTS

1 INTRODUCTION.................................................................................................................. 1-1 1.1

Background............................................................................................................... 1-1

1.2

Electronic/Digital Positioners..................................................................................... 1-1

1.3

Approach................................................................................................................... 1-1

1.4

Highlighting Key Points ............................................................................................. 1-2

1.5

Glossary.................................................................................................................... 1-2

1.6

References................................................................................................................ 1-2

2 INTRODUCTION TO THE CONTROL LOOP ...................................................................... 2-1 2.1

Purpose .................................................................................................................... 2-1

2.2

Overview................................................................................................................... 2-1

2.3

Description ................................................................................................................ 2-1

2.4

Summary................................................................................................................... 2-3

2.5

Additional Information................................................................................................ 2-4

2.6

Reference ................................................................................................................. 2-4

3 POSITIONER DESIGN AND APPLICATION....................................................................... 3-1 3.1 Introduction................................................................................................................... 3-1 3.2

Two Black Boxes....................................................................................................... 3-1

3.2.1 Black Box One...................................................................................................... 3-1 3.2.2 Black Box Two...................................................................................................... 3-2 3.2.3 Discussion............................................................................................................ 3-2 3.3

Positioner Design ...................................................................................................... 3-3

3.3.1 Function ............................................................................................................... 3-3 3.3.2 Building Blocks ..................................................................................................... 3-4 3.3.3 Input Signal Conversion........................................................................................ 3-4 3.3.3.1 Nozzle-Flapper .............................................................................................. 3-5 3.3.4 Output Signal Generator....................................................................................... 3-7

xi


3.3.4.1 Connecting the Input Converter to Output Generator..................................... 3-7 3.3.4.2 Types of Output Signal Generators ............................................................... 3-7 3.3.5 Output Signal Correction (Feedback).................................................................. 3-11 3.3.5.1 Purpose....................................................................................................... 3-11 3.3.5.2 Interface Between Valve and Positioner ...................................................... 3-11 3.3.5.3 Positioner Input ........................................................................................... 3-11 3.3.5.4 Motion-Balance and Force-Balance (Balancing Mechanisms) ..................... 3-13 3.4

Classification of Commonly Used Positioners.......................................................... 3-14

3.5

Positioner Application.............................................................................................. 3-15

3.5.1 Limiting Control Valve Dead Band ...................................................................... 3-15 3.5.2 Mitigating Stiction or Stick-Slip............................................................................ 3-15 3.5.3 Change Valve Response .................................................................................... 3-16 3.5.4 Control Double Acting Actuator........................................................................... 3-16 3.5.5 Increase Shutoff Forces...................................................................................... 3-16 3.5.6 Split-Ranging...................................................................................................... 3-16 3.5.7 Delays Due to Distance Between Controller or I/P Converter and Valve............. 3-17 4 CALIBRATION .................................................................................................................... 4-1 4.1

Calibration and Condition Monitoring/Preventive Maintenance.................................. 4-1

4.2

Basic Calibration ....................................................................................................... 4-1

4.2.1 Bench Set Confirmation........................................................................................ 4-1 4.2.2 Feedback Alignment............................................................................................. 4-2 4.2.3 Zero and Span Adjustment ................................................................................... 4-3 4.2.3.1,Purpose......................................................................................................... 4-3 4.2.3.2,Static Band.................................................................................................... 4-3 4.2.3.3 Adjustments .................................................................................................. 4-4 4.3

Functional Check ...................................................................................................... 4-5

4.4

Data Acquisition Systems.......................................................................................... 4-6

4.4.1 Calibration Using the Data Acquisition System ..................................................... 4-6 4.4.2 Hysteresis/Dead Band (Dynamic Error) ................................................................ 4-8 4.5

References................................................................................................................ 4-8

5 CONDITION MONITORING/PREVENTIVE MAINTENANCE............................................... 5-1 5.1

Condition Monitoring ................................................................................................. 5-1

5.1.1 Continuation of Calibration ................................................................................... 5-1

xii


5.1.2 Data Acquisition Systems ..................................................................................... 5-2 5.1.2.1 Other Recommended Performance Curves Using a Data Acquisition System...................................................................................................................... 5-2 5.2

Preventive Maintenance............................................................................................ 5-3

5.2.1 Air Supply Check.................................................................................................. 5-3 5.2.2 Walkdown Inspection............................................................................................ 5-4 5.2.3 Internal Inspections (as Applicable) ...................................................................... 5-4 5.3

References................................................................................................................ 5-5

6 TROUBLESHOOTING......................................................................................................... 6-1 6.1

Introduction ............................................................................................................... 6-1

6.2

Approach................................................................................................................... 6-1

6.2.1 Use of Data Acquisition System............................................................................ 6-1 6.2.2 Symptom-Based Approach................................................................................... 6-2 6.2.3 Positioner ............................................................................................................. 6-3 A CONTROL LOOP DETAILS ...............................................................................................A-1 A.1

Purpose .................................................................................................................... A-1

A.2

Control Loop Block Diagram......................................................................................A-1

A.3

Loop Elements .......................................................................................................... A-2

A.3.1 Sensor/Transmitter ............................................................................................... A-2 A.3.2 The Controller (Including the Comparator)............................................................ A-3 A.3.2.1Proportional Control....................................................................................... A-3 A.3.2.1.1 Offset..................................................................................................... A-5 A.3.2.2Integral (Reset) Control ................................................................................. A-5 A.3.2.3Derivative (Rate) Control ............................................................................... A-5 A.3.2.4 Controller Problems and Tuning .................................................................... A-5 A.3.3 The Final Control Element .................................................................................... A-6 A.4

Control Loop Problems.............................................................................................. A-6

A.4.1 Dead Time............................................................................................................ A-6 A.4.1.1Sources of Dead Time................................................................................... A-7 A.4.1.2Component Dead Time ................................................................................. A-7 A.4.1.3Identifying Dead Time Problems.................................................................... A-7 A.4.1.4Dealing with Dead Time ................................................................................ A-8 A.4.1.5Lag ................................................................................................................ A-9

xiii


A.4.2 Linearity................................................................................................................ A-9 A.4.2.1 Inherent and Installed Flow Characteristics ................................................. A-10 A.4.2.2 Hysteresis/Dead Band/Stem Friction ........................................................... A-11 A.4.2.3Dynamic Error ............................................................................................. A-11 A.5

Scaling Calculations................................................................................................A-11

A.6

Reference ............................................................................................................... A-12

B AOV POSITIONER CHECKLIST ........................................................................................B-1 C GLOSSARY ........................................................................................................................C-1 D DIGITAL POSITIONERS.....................................................................................................D-1 D.1

Introduction ...............................................................................................................D-1

D.2

Reason for Digital......................................................................................................D-1

E POSITIONER PROBLEM CASE HISTORIES..................................................................... E-1 E.1

Introduction ............................................................................................................... E-1

E.2

Case Histories........................................................................................................... E-1

E.3

Reference ................................................................................................................. E-9

F MINIMUM INTEGRATED ABSOLUTE ERROR TUNING.................................................... F-1 G SUMMARY OF KEY POINTS .............................................................................................G-1

xiv


LIST OF FIGURES Figure 2-1 Example of a Control Process................................................................................ 2-2 Figure 3-1 Positioner Block Diagram....................................................................................... 3-4 Figure 3-2 Input Signal Converters.......................................................................................... 3-5 Figure 3-3 Nozzle/Flapper Operation ...................................................................................... 3-6 Figure 3-4 Balance Beam........................................................................................................ 3-7 Figure 3-5 Spool Valve (Pneumatic)........................................................................................ 3-8 Figure 3-6 Double Three-Way Poppet Directional Control Valve ............................................. 3-9 Figure 3-7 Throttle Directional Control Valve......................................................................... 3-10 Figure 3-8 Pressure Control Directional Control Valve .......................................................... 3-11 Figure 3-9 Feedback Cam Characterization.......................................................................... 3-12 Figure 3-10 Pneumatic Transmission Lag: Time to Reach 63.2% Final Value (Time Constant ) .................................................................................................................... 3-17 Figure 4-1 Calibration of Valve Positioner ............................................................................... 4-7 Figure A-1 Level Control Block Diagram ................................................................................ A-1 Figure A-2 Proportional Band................................................................................................. A-3 Figure A-3 Controller Output Response to Square Pulse Showing Gain ................................ A-4 Figure A-4 Stability................................................................................................................. A-4 Figure A-5 Process Variable Versus Controller Output Showing Stiction (Limit Cycle) ........... A-8 Figure A-6 Robustness Plot ................................................................................................... A-9 Figure A-7 Valve Flow Characteristics.................................................................................. A-10 Figure E-1 Severely Deformed Positioner Bellows (Input Signal Converter) —Travel Versus Positioner Input Pressure .................................................................................... E-1 Figure E-2 Double Acting Cylinder —Stability Problem due to Insufficient Air Pressure (I/P Input Signal/Actuator Position Versus Time) ............................................................. E-2 Figure E-3 Double Acting Cylinder —Stability Problem due to Insufficient Air Pressure (Cylinders Differential Pressure [Low Cyl]/Actuator Position Versus Time) ...................... E-3 Figure E-4 Double Acting Cylinder —Corrected Stability Problem due to Insufficient Air Pressure (Cylinders Differential Pressure [Low Cyl]/Actuator Position Versus Time) ....... E-4 Figure E-5 Double Acting Cylinder —Stability Problem due to Insufficient Air Pressure (As-Found Versus As-Left Actuator Position Versus Control Signal Input)....................... E-5 Figure E-6 Positioner with No Air Leak —Supply Pressure/Actuator Pressure/Actuator Position Versus Time Plot ............................................................................................... E-6

xv


Figure E-7 Positioner with Internal Air Leak —Supply Pressure/Actuator Pressure/Position Versus Time Plot ................................................................................ E-7 Figure E-8 Positioner with Internal Air Leak —Supply Pressure Versus Actuator Pressure Plot................................................................................................................... E-7 Figure E-9 Air Leak Downstream of Positioner —Supply Pressure/Actuator Pressure/Position Versus Time Plot ................................................................................ E-8 Figure E-10 Air Leak Downstream of Positioner —Increasing/Decreasing Supply Pressure Versus Time Plot ............................................................................................. E-9 Figure F-1 Area Representation of Integrated Absolute Error................................................. F-1

xvi


LIST OF TABLES Table 2-1 The Positioner in the Control Loop ......................................................................... 2-3 Table 3-1 Classification of Commonly Used Positioners....................................................... 3-14 Table 4-1 Suggested Static Band Pressure Values................................................................ 4-5 Table 4-2 Suggested Intermediate Median Values for Selected Pressure Inputs ................... 4-6 Table 5-1 Condition Monitoring Performance Curves ............................................................. 5-2 Table 6-1 Final Control Element Problem Symptoms/Causes ................................................ 6-2 Table 6-2 Positioner Problem Symptoms/Causes .................................................................. 6-3 Table F-1 Post-Test Settings to Achieve Minimum Integrated Absolute Error Tuning.............. F-2

xvii


1

INTRODUCTION

1.1

Background

In 1999, the EPRI Nuclear Maintenance Application Center (NMAC) conducted a survey of unplanned capacity loss factors. The survey identified control valves as the number four cause of such losses. To address this issue, an additional survey was completed to find out which areas had not been previously addressed by earlier EPRI documents. It was determined that Control Valve Guidelines [1] provided a detailed treatment of the control valve proper and, to some extent, addressed some of the accessories. What appeared to be missing was information on control loops. Further study determined that to provide a document meaningful to experienced technicians and their supervision, the document should focus on components that have historically received little attention. In addition, it was beneficial to survey other industries to determine the leading causes of control loop problems. One of the most useful surveys was performed in the paper and pulp industry. This survey showed that of 31 valve control problems, 71% were attributable to the positioner. The next most frequent problem cause was bench set at 38%. Based on this information, the positioner as part of the control loop was selected for further research.

1.2

Electronic/Digital Positioners

This guide gives details only for pneumatic positioners. The population of electronic and digital positioners used in the nuclear industry is still small, and experience is limited. Appendix D contains a brief summary of digital positioners.

1.3

Approach

In many cases, the positioner is mounted by the control valve manufacturer and has been adjusted as required to give satisfactory results. However, when positioner troubleshooting begins in earnest, a routine maintenance action can turn into a frustrating learning experience. This occurs partly because the positioner is part of a complex group of components called a control loop. When a problem is encountered, the focus tends to be upon the controller or the valve. In addition, fixing these components may result in satisfactory operation even though it seems that something else is wrong. Therefore, this guide provides the basis for a systematic approach to troubleshooting that distinguishes between the component that causes a problem and the component that is affected by that problem. The approach is to briefly discuss the control loop and then focus on the function of the positioner in the control loop. Subsequent sections address positioner design and 1-1

EPRI Licensed Material Introduction

application , calibration, condition monitoring/preventive maintenance, and troubleshooting. The appendices contain control loop details, checklists for positioner checkout, a glossary, information about digital positioners, and positioner problem case histories.

1.4

Highlighting Key Points

Throughout this report, key information is summarized in key points. Key points are boldlettered boxes that succinctly restate information addressed in detail in the surrounding text, making the key point easier to locate. The primary intent of a key point is to emphasize information that enables individuals to take action for the benefit of their plant. The information included in these key points was selected by NMAC personnel, consultants and utility personnel who prepared and reviewed this report. The key points are organized according to the three categories: O&M costs, technical, and human performance. Each category has an identifying icon, as shown below, to draw attention to the specific category when quickly reviewing the guide. Key O&M Cost Point Emphasizes information that will result in reduced purchase, operating, or maintenance costs.

Key Technical Point Targets information that will lead to improved equipment reliability.

Key Human Performance Point Denotes information that requires personnel action or consideration in order to prevent injury or damage or ease completion of the task.

Appendix G contains a listing of all key information in each category. The listing restates each key point and provides reference to its location in the report. A review of this listing can help users of this guide determine if they have taken advantage of key information that the authors believe would benefit the users ’ plants.

1.5

Glossary

A glossary of terms used in this guideline is contained in Appendix C.

1.6

References

1. Control Valve Guidelines, EPRI, Palo Alto, CA: 1994. TR-102051-R1. 1-2


2

INTRODUCTION TO THE CONTROL LOOP

2.1

Purpose

Positioners are part of a control loop. If not specified properly or if not maintained properly, positioners can have an unacceptable effect on process control, costing both maintenance time and materials. These effects may result in valve “hunting, ” in a condition called limit cycle, or they can result in plant shutdown due to internal leakage. In addition, positioners are control loops themselves and subject to control loop problems. Therefore, when dealing with positioner problems, you are really dealing with control loop problems at both system and component levels. This section provides information about control loop basics.

2.2

Overview

The control loop is fundamentally nothing more than a group of components, normally in a series, each of which responds to input from a previous component by supplying output to the next component. The goal of the loop is to work together to control/maintain a process as desired when challenged by some disturbance. To operate as desired means that there is a process variable —for example, flow or temperature —that will be maintained at a value. To do this automatically without operator action, some type of feedback is necessary to provide this maintenance. In the simplest case, we may only want to control a flow, and we do so by measuring the flow and causing a valve to close or open based on whether the flow value is high or low when compared to the desired value. In another case, we may want to control tank level by controlling how much flow is going into or out of the tank. Alternatively, we can control the temperature of a fluid exiting a heating unit by varying the voltage applied to the heating element used as the source of heat.

2.3

Description

Each process control loop may be described in terms of process variables and control elements. The control elements include a sensor, transmitter (usually lumped with the sensor), controller, and final control element. In the following example of a control process (see Figure 2-1), we are trying to maintain a process variable, the tank level.

2-1

EPRI Licensed Material Introduction to The Control Loop

Figure 2-1 Example of a Control Process

The tank level is measured by the level sensor, and the level transmitter sends a corresponding signal to the controlling device. In most cases, this signal is electrical and has been scaled to correspond to a maximum and minimum level of interest. In the controlling device, the level signal is compared to a signal that corresponds to the desired tank level (the setpoint). In some cases, this comparison may take place before the controlling device provides a signal. The result is an error signal that has a direction (high/low) or polarity (plus/minus) that is desired to indicate which way the tank level is from the desired level. The controlling device has been programmed (adjusted) to provide a signal to the positioner of the final control element (FCE), an air operated flow control valve (AO FCV) that causes it to open or close so that the error signal becomes zero. If everything functions correctly, the system is in equilibrium with the flow-in and the flow-out essentially the same. However, the flow-in is subject to disturbances, and as a result, the level changes. As an example, suppose that the flow-in becomes less. The tank level begins to drop because the flowout has become greater than the flow-in. The level sensor detects this, and the corresponding signal is sent by the transmitter and immediately compared to the desired value. The controller generates a signal to the positioner, which will cause the AO FCV to move in the closed direction. The positioner compares the valve position with the signal and causes the valve to move as required to a new position. As the flow-out is reduced to below the flow-in, the tank level finally begins to recover. Depending on the controlling device, the level either returns to the desired level or to one that is slightly lower (with an offset). In most power plant controlling devices, the process returns to the desired level without offset. The preceding example shows how a control loop functions. As with every technology, understanding the language is essential to understanding the concepts. Information about control loops includes terms like feedback and manipulated variable. Table 1 contains some terms and how they apply to the example. These terms are also defined in Appendix C and in Process Instrumentation Terminology [1].

2-2


Table 2-1 The Positioner in the Control Loop Terminology

Physical Representation

Controlled variable

Tank level

Manipulated variable

Flow-out

Setpoint

The (electrical equivalent of the ) desired tank level

Summing point

Where (the electrical equivalents of) the actual tank level (from the transmitter) and the set point are compared. A comparator.

Feedback element

The level transmitter

Disturbance

Flow-in

Error

Output of the summing point

The process above is what is called a closed loop because a feedback occurs through the level transmitter. On a control panel, closed-loop operation is known as auto control. If the feedback path is broken (for example, the transmitter fails as-is), then the control loop becomes an open loop and the system is in manual control. An operator then has to directly manipulate the flowout in response to level changes. The enemies of all closed loops are dead time and non-linearity. Dead time is the delay associated with a control loop response. Non-linearity is the failure of a control loop to respond to an error with the same magnitude of response over the range of control (for a detailed discussion of these topics, see Appendix A). Positioners can help to mitigate these problems. One of the positioner ’s functions is to help combat dead time by supplying additional gain, or amplification, to the control loop. Gain increases response time and, therefore, improves the control loop’s ability to deal with disturbances. However, too much gain can also be a problem, causing instability in the loop. To mitigate process non-linearity, positioners can be adjusted or characterized to produce an output that algebraically subtracts the non-linearity over the range of control. This is done by characterizing a cam. Characterization is discussed in Chapter 3.

2.4

Summary

Positioners are part of a control loop. If not specified properly or if not maintained properly, they can have an unacceptable effect on process control, costing both maintenance time and materials. The control loop is fundamentally nothing more than a group of individual components in a series, each of which responds to input from the previous component by supplying output to the next component The goal of the loop is to work together to control a process as desired when challenged by some disturbance.

2-3


To operate control loops, use feedback to monitor a controlled variable , and send signals to a manipulated variable to keep the controlled variable at some setpoint. Control loops using feedback are called closed loops. All closed loops experience time delay and non-linearity in processing changes to the controlled variables. The delay of the most concern is dead time, which positioners can improve by supplying gain. To mitigate process non-linearity, positioners can be adjusted or characterized, using a cam, to produce an output that subtracts the deviations over the range of control.

2.5

Additional Information

For additional information on control loops, see Appendix A, Control Loop Details.

2.6

Reference

1. ANSI / ISA S51.1-1979 (R1993), “Process Instrumentation Terminology, ” Instrument Society of America, Research Triangle Park, NC: 1993.

2-4


3

POSITIONER DESIGN AND APPLICATION

3.1 Introduction Every manufacturer supplies a technically accurate service manual that describes the construction, operation, and calibration of the positioner at various levels of detail. Therefore, an understanding of the basic designs and applications is useful for interpreting the manual during calibration, maintenance and troubleshooting. This section will review basic positioner operation, break a positioner up into discrete components, describe various design principles for these components and provide a table of commonly used positioners classifying them according to the principles. Using these principles of operation, the section will describe how positioners are applied to solve many problems that occur in control valve applications, including the problems of dead band and linearity discussed in the previous section.

3.2

Two Black Boxes

Before discussing positioner design, it may be helpful to think about the positioner as a black box. This will help to understand positioner capabilities and function. In fact, two black boxes are discussed in this section, and both look the same on the outside. Each has two pneumatic inputs and one pneumatic output. The inputs come from a supply of air, for example, at 20 psig (137.9 kPa), and a pressure regulator that we can vary from 0 psig to 30 psig (206.8 kPa). The inputs are called supply and signal, respectively. The output of each is routed to the diaphragm of an air actuator. The actuators begin stroking at about 3 psig (20.7 kPa) and completely stroke with about a 12 psi (82.7 kPa) change of air pressure. Three pressure gages measure the input from the regulator and the output to each actuator. A mechanical linkage from the actuator stem to the black box corresponds to the position of the stem.

3.2.1 Black Box One The signal to black box one (BB1) is increased. The actuator stem begins to move when the pressure is just over 3 psig (20.7 kPa). It continues to move until it reaches some limited position —for example, the backseat in the attached valve —at just under 15 psig (103.4 kPa). The distance traveled is proportional to the amount of pressure at the input. In addition, the output pressure closely follows the signal pressure, but it may not be the same. As the signal pressure varies up and down, the position and the output pressure respond up and down. As the signal is increased to just below 15 psig (103.4 kPa), the valve stops moving, and the pressure goes to 20 psig (137.9 kPa) and stops.

3-1

EPRI Licensed Material Positioner Design and Application

The signal is set to 9 psig (62.1 kPa), and the actuator stem is gagged at about the mid-position of travel. The signal is increased to 9.1 psig (62.7 kPa). However, instead of going to 9.1 psig (62.7 kPa), the output pressure increases to 20 psig (137.9 kPa), or the supply pressure. Correspondingly, when the signal is decreased by 0.1 psig (0.7 kPa) —that is, to 8.9 psig (61.4 kPa), the output goes to 0 psig. Repeating the change from 9.0 psig (62.1 kPa) but going to 9.2 psig (63.4 kPa) only makes the pressure change faster. Next, the stem is disconnected from the box, and the linkage is left in the same position as when the stem was locked. The signal is increased to 9.1 (62.7 kPa), and the output pressure responds as it did with the locked stem, that is, going to 20 psig (137.9 kPa). The actuator, free to move, shifts to a fully open position. Similar locked-stem responses are obtained at greater pressures or smaller pressures. In effect, the response of the box is no different than when the stem was locked.

3.2.2 Black Box Two With the valve stem free to move, black box two (BB2) responds the same as BB1. However, when the stem is gagged as before, the response is somewhat different. When the signal is increased to 9.1 psig (62.7 kPa), the output pressure goes quickly to 14 psig (96.5 kPa) —a 5 psig (34.5 kPa) increase —and stops. Increasing the signal to 9.2 psig (63.4 kPa) results in 19 psig (131 kPa) at the output. At 9.3 psig (64.1 kPa) signal input, the output is 20 psig (137.9 kPa). Correspondingly, by decreasing the signal by 0.1 psig (0.7 kPa) —that is, to 8.9 psig (61.4 kPa), the output goes to 4 psig (27.6 kPa). Decreasing to 8.8 psig (60.7 kPa) results in 0 psig output. The same response is obtained when the stem is disconnected and the linkage fixed as before.

3.2.3 Discussion Both of these black boxes have responded as positioners. When they are hooked up to the stem, the stem position corresponds to the signal received. As long as the stem is free to move, there is virtually no difference between having a positioner or having the signal connected directly to the valve actuator diaphragm/piston. The response is the same. The action of a positioner is very clear when the stem is prevented from moving. This corresponds to factors such as friction and inertia. The positioner responds by immediately supplying gain to the signal. The process does not have to drift further from the setpoint to obtain an error signal large enough to cause movement. This gain quickly overcomes the resistance to movement. Because the black box attempts to drive the stem to some desired position (as observed by an increased/decreased output pressure to the actuator) when the stem is prevented from moving (a disturbance), the black box must contain components that are acting together as a control loop. The black boxes illustrate two types of gain. In the case of BB1, the gain is called flow gain because the flow increases with increased signal pressure and output continues to rise as long as an input signal is present. In the case of BB2, the gain is called locked-stem pressure gain

3-2


(sometimes called open-loop gain) because output pressure has a fixed relationship to the inlet or signal pressure. In reality, the difference between the black boxes (pressure versus flow gain) is not seen. When pressure changes are made to a pressure-gain type of unit, the changes are so large (that is, 1 psig [6.9 kPa]) or more) that when multiplied by the gain, the output pressure exceeds the remaining margin between the initial pressure and either zero pressure or the supply pressure. As a result, pressure changes appear the same as seen in flow amplification. In normal maintenance activities, the values or type of gain is unimportant, but knowledge of their existence may provide understanding of a problem or the result of some test. The locked-stem result and disconnected fixed-linkage-position response are the same. This means that it is the linkage that is providing feedback to the positioner to null or correct the input signal. Stem position is the desired response; therefore, the linkage provides feedback to the positioner. Because there is feedback, the positioner is a closed control loop. The boxes have a maximum or minimum output pressure equal to the supply pressure (20 psig [137.9 kPa]) or zero. This means that at the extremes of travel, a positioner can provide additional force to the stem because it is not limited to signal range, for example, 3 –15 psig (20.7–103.4 kPa), but to the larger range of zero to the supply pressure.

3.3

Positioner Design

3.3.1 Function Based on the capabilities described in the black box discussion, a positioner can have the following functions: Provide an output pressure that tracks the input signal closely. This does not necessarily mean that the input and output pressures are the same. The actuator characteristics or even the positioner characteristics may introduce some difference. What this does mean is that for any input signal, there is a corresponding actuator position that is always the same for any given pressure. Provide an output pressure that increases (or decreases) rapidly whenever there is a difference between the desired position and the corresponding input pressure. In other words, provide a gain. This may be either a flow gain or a pressure gain . When the position is correct, the positioner reduces the gain rapidly. Provide for tracking in which the output pressure increases with the input pressure increase (direct acting), or in which the output pressure decreases with the input pressure increase (reverse acting). An additional function not described in the black box discussion is the output characterization that mitigates process non-linearity (see Appendix A, Section A.4.2 Linearity).

3-3


3.3.2 Building Blocks Within any positioner is a set of discrete components that function together as described above. In all cases, a motion of one of these components causes a series of events that result in a counter-motion. This motion may be the simple movement of components or a result from the application of force. There are a limited number of ways that these component can be designed and assembled to make this happen. The block diagram in Figure 3-1 illustrates this assembly. These components perform the following functions within the positioner: 1. Input signal conversion: The pneumatic input is converted to a mechanical motion. Positioner gain may be developed here using a pneumatic amplifier. In such cases, the positioner is referred to as a two-stage device. 2. Output signal generation: A mechanical motion causes a directional control valve to change position and supply air to (or exhaust air from) the actuator. Positioner gain will normally be developed here, usually through the use of spring or spring-like devices. 3. Output signal correction: The gain developed earlier is reduced to zero.

Figure 3-1 Positioner Block Diagram

3.3.3 Input Signal Conversion The input signal can be sent to the positioner is one of two ways: A pneumatic signal (3 –15 psig [20.7 –103.4 kPa], 6–30 psig [41.4–206.8 kPa] 3–27 psig [20.7–186.2 kPa], etc) directly from the controller An electrical signal (4 –20ma, 10–50ma, 0–10v) from the controller that is converted to a pneumatic signal by a current to pneumatic (I/P) converter or a voltage to pneumatic (E/P) converter that is either external or internal to the positioner.

3-4


The pneumatic signal must be converted to a mechanical motion. This conversion process always begins with a diaphragm or bellows chamber. The changing pressure is transformed to a linear motion. Simplified sketches of these components are shown in Figure 3-2.

Figure 3-2 Input Signal Converters

The linear motion is now used directly to position a directional control valve within the positioner, or it is used to modulate the flapper of a nozzle-flapper.

3.3.3.1

Nozzle-Flapper

Input signal conversion may employ a “preamplifier ” to enhance positioner response. The nozzle-flapper is a device that produces an amplified pneumatic signal. The flapper is sometimes called a vane. Figure 3-3 shows the basic layout of the nozzle-flapper. In this example, the bellows expands/contracts, moving the flapper toward or away from the nozzle. During a steady state operation, supply air passes through two restrictions. The first restriction is a fixed orifice and sized to permit adequate flow for nozzle-flapper operation without affecting the supply pressure. The second restriction is caused by the flapper moving toward/away from the nozzle. Moving the flapper toward the nozzle increases the pressure in the nozzle chamber and vice versa. In other words, the orifice restriction size is smaller than the nozzle restriction size, which allows the supply pressure to bleed to atmosphere faster than it enters the unit through the fixed restriction when the flapper is away from the nozzle.

3-5


Figure 3-3 Nozzle/Flapper Operation

Amplification or gain comes about because it takes very little motion of the flapper and hence very little pressure change in the input signal converter to give wide variations in the control pressure. The input bellows (or diaphragm in some cases) is designed to produce relatively large movement to the flapper relative to its pressure changes. Thus, a small pressure change in the input will produce a large change in the nozzle chamber pressure. The nozzle chamber pressure is then piped to a diaphragm-type signal converter that is connected to the output signal generator. A nozzle-flapper/diaphragm-like signal converter combination is commonly called a two-stage device and is characterized by high gain and rapid response. In some processes, this may be desirable, such as feedwater control. The nozzle-flapper mechanism is used as a preamplifier and not as a primary signal generator for the following reasons: If the restrictions are made large enough to make it possible to rapidly change the backpressure, excessive amounts of air are required. Large restrictions result in an increase of the “blast effect ” (momentum of the air stream through the nozzle). This requires that the flapper be more robust and makes it necessary to use larger forces to overcome the blast. This robustness results in feedback mechanisms that are less sensitive to changes in control pressure. If the restrictions are kept small to reduce the air consumption, it takes too long to change the back-pressure. As above, the sensitivity is reduced. Therefore, it is necessary to route the control pressure to a second device, generically called a relay, which then controls the supply pressure to generate the output signal.

3-6


The fixed flow restriction can become fouled, reducing flow rate and causing loss of sensitivity, especially when movement is demanded in one direction. In the extreme, this may result in no valve movement from one end of the stroke. All positioners using nozzle/flapper mechanisms have a means to clean this fixed flow restriction.

3.3.4 Output Signal Generator

3.3.4.1

Connecting the Input Converter to Output Generator

The input signal converter sends a linear motion that is used by the output signal generator in one of two ways: The linear motion is used directly to cause the output signal generator to move and to transmit an output signal. For example, an input diaphragm linear motion is connected directly to the stem of a sliding spool directional control valve. The linear motion is used to move a balance beam that causes the output signal generator to move and to generate an output signal. An example of a balance beam is shown in Figure 3-4.

3.3.4.2

Types of Output Signal Generators

The term output signal generator describes the function of several different types of mechanisms that are used to provide the output signal. The manufacturers ’ terms used to describe the output signal generator function vary widely and include pilot valve, pneumatic relay, pneumatic amplifier, and relay. In other words, there is no common functional nomenclature for these or, for that matter, any of the other parts of the positioner. Therefore, the names of these devices are meant to describe the particular make-up of the part by using nomenclature commonly encountered in various industries.

Figure 3-4 Balance Beam

3-7


Sliding Spool Directional Control Valve

The sliding spool valve (see Figure 3-5) finds use in many pneumatic and hydraulic applications. It is very simple in construction and usually more rugged but slower than the two-stage device. A typical sliding spool valve is shown in Figure 3-5. Neutral spool position is shown. If the spool is moved to the right, then supply is ported to Cylinder A; if it is moved to the left, then supply is ported to Cylinder B. Simultaneously, Cylinder B (A) is vented through a groove (identified as exhaust in Figure 3-5) or other such passage and hence to the atmosphere. The input to the spool is usually supplied by the mechanical motion of a diaphragm. The spool may be configured to provide two outputs for use in double-acting actuator applications or with one output plugged for spring-operated (single acting) actuators.

Figure 3-5 Spool Valve (Pneumatic)

Double Three-Way Poppet Directional Control Valve

The double three-way poppet directional control valve (Figure 3-6) describes a device that provides the same output capability as the spool valve. For example, it can be used for doubleacting or single-acting actuator applications. However, the internal design varies considerably. It usually combines the input conversion and output signal generation in a single unit. Referring to Figure 3-6, control pressure is contained between two diaphragms —an isolation diaphragm (shown to the right of the control chamber in the figure) and a force balance diaphragm (shown to the left of the control chamber). On the other side of the force balance diaphragm is a supply pressure chamber (shown to the left of the supply pressure chamber) also with an isolation diaphragm. The control pressure is varied by using a nozzle/flapper device as previously described. Variation in the control pressure allows the hollow center structure to shift right or left, thus opening poppet valves at both ends. The figure shows that when the input signal is lowered, it causes the flapper to move away from the nozzle and thus lowers the control pressure, allowing movement to the right. This causes supply air to be ported to Cylinder A and allows Cylinder B to exhaust through the hollow central structure. As the central structure begins to move, however, immediate feedback to the flapper controls the positioner pressure gain. As the actuator stem begins to move, the feedback is received (output signal correction), and the flapper is brought back to its original position, thus returning the central structure to a neutral position with all poppet valves closed. 3-8


Figure 3-6 Double Three-Way Poppet Directional Control Valve

Throttle Valve

The throttle valve is a simple device (see Figure 3-7) that allows the air supply to go in two directions. As the throttle moves toward the output end, more air is bled off and the output pressure is reduced. Movement of the throttle is usually through a balance beam device as shown.

3-9


Figure 3-7 Throttle Directional Control Valve

Pressure Control Directional Control Valve

The last commonly used output signal generator is sometimes called a relay or pneumatic relay. It operates like a pressure control valve, except the adjustable coil spring is replaced by a pressure (control) chamber (see Figure 3-8). The pressure in the chamber is the control pressure created by a nozzle/flapper device. When in equilibrium, the pressure on the other side of the diaphragm is the output pressure. When the control pressure increases, it pushes down the upper diaphragm, which is connected to the lower diaphragm, thus pushing the poppet down. Supply pressure is then admitted through the lower end of the poppet to the actuator until the output pressure value provides enough force on the diaphragm to counter the control pressure and close the poppet. Similarly, when the upper diaphragm moves up, the seat at the upper end of the poppet moves up and allows venting from the actuator diaphragm to the exhaust. The lower seat prevents the poppet from moving up. Again, the output pressure decays sufficiently to allow the upper diaphragm to move down.

3-10


Figure 3-8 Pressure Control Directional Control Valve

3.3.5 Output Signal Correction (Feedback)

3.3.5.1

Purpose

The output signal correction is to use the feedback to nullify or balance the output when the desired position is achieved. This relationship has been identified in Figures 3-3, 3-4, 3-6, and 3-7. The output signal correction is also the point in the positioner loop in which the output can be modified, or characterized to mitigate process non-linearity.

3.3.5.2

Interface Between Valve and Positioner

To begin the balancing process, a rod, bracket, or other suitable device is attached to the stem. The device is mechanically routed to the positioner (for example, using linkages) to provide the appropriate motion used by the positioner: rotation or linear push-pull. If the motion is rotation, the input to the positioner is a lever.

3.3.5.3

Positioner Input

A linear feedback motion always results in a proportional feedback to provide a balance force to nullify the output. In other words, no matter where the stem is, any amount of movement of the stem always results in the same but proportional amount of movement at the positioner. There may be an interface that allows the input range to be adjusted, but the motion remains linear. Sometimes it is necessary to modify this input. This is where the rotary motion comes in. Rotary motion is used in the positioner to modify the feedback in a non-linear way if desired. The reason for making it non-linear is to compensate for some non-linearity in the control loop. The universal method for doing this is by using a cam. Input from the stem causes the input lever to rotate the cam (see Figure 3-9). The cam follower rides on the edge of the cam and follows the

3-11


cam contour. This allows the original input lever motion to be changed, or characterized, before it is used to correct the output through the balance beam.

Figure 3-9 Feedback Cam Characterization

Normally, three cams are available as standard —square root or quick opening, linear, and square or equal percentage —because these cams handle some of the more common non-linearities. Also available are blank cams that can be modified by the user to obtain other characterizations.

3-12


3.3.5.4

Motion-Balance and Force-Balance (Balancing Mechanisms)

Output signal correction uses feedback to nullify or balance the output when the desired position is achieved. Balancing involves one of two principles. The motion that initiated the action must be met with a counter-motion, or the force that initiated the action must be met with a counterforce. These counter-actions are referred to as motion balance and force balance. The motion-balance positioners use nozzle-flapper devices (however, this does not mean that using a nozzle flapper makes it a motion balance positioner). These positioners use a beam, commonly called a balance beam, that moves about a pivot. One part of the beam is usually moved by a bellows in response to the input signal. This motion moves the flapper toward/away from the nozzle. The feedback from the valve position is also applied as a motion to the beam. This motion counters the original input in such a way that a different output pressure is seen at the relay. In Figure 3-3, the flapper acts as a motion-balance device. Most positioners use the force-balance concept. In the force-balance device, the input signal creates a force that is resisted by an opposing force created by the feedback mechanism. Force balance devices are commonly associated with spool valves in which the spool is moved back and forth in response to the equilibrium position of the opposing forces. One of the easiest ways to identify a force-balance device is the existence of stem feedback being applied through a spring, which is called a range spring.

3-13


3.4

Classification of Commonly Used Positioners

Table 3-1 classifies commonly used positioners according to the principles described in this chapter. While the actual designs may vary somewhat in appearance when compared to the chapter figures, the underlying principles of operation remain the same. Table 3-1 Classification of Commonly Used Positioners Manufacturer Model

3-14

Input Signal Conversion

Output Signal Generator

Output Signal Correction

Type of Gain

Bailey AP

Diaphragm, nozzle flapper

Double three-way poppet directional control valve

Force-balance, cam characterization

Pressure

Bailey AV, 5311450, 5321030

Bellows, balance beam

Sliding spool directional control valve


Flow

Conoflow GP 50

Bellows, nozzle flapper, diaphragm


Motion-balance, cam characterization

Pressure

Fisher 3582

Bellows, nozzle flapper, diaphragm (in pressure control valve)

Pressure control valve

Motion-balance, cam characterization

Pressure

Masoneilan 4600

Diaphragm



Pressure

Masoneilan 7000

Bellows, balance beam

Throttle valve

Force-balance

Pressure

Moore 74

Diaphragm, nozzle flapper, diaphragm (two-stage)


Force-balance

Pressure

Moore 750P

Diaphragm, nozzle flapper, diaphragm (two-stage)



Pressure

Valtek Beta

Diaphragm, balance beam



Flow

Valtek System 80 and XL Series

Diaphragm, nozzle flapper



Pressure


3.5

Positioner Application

Positioners are used to solve many problems that occur in control valve applications. The following sections describe the major uses.

3.5.1 Limiting Control Valve Dead Band Dead band is a characteristic of any mechanical device, and the control valve is no exception. Dead band refers to the delay in motion of a control valve that occurs after a reversal in demand. In a sense, it is “free play. ” It can involve part looseness (backlash), inertia, friction, or friction. To overcome dead band, the signal to the valve is sufficiently amplified (gain is added) by the positioner so that its effect can be minimized. Dead band adds to dead time, the enemy of control loop stability. Dead band as much as 5% is not uncommon for a control valve without a positioner. With a positioner, dead bands can be reduced to less than 1%. There are limitations on the amount of gain that can be supplied to limit dead band. For additional details, see the discussion on dead band in Appendix A.

3.5.2 Mitigating Stiction or Stick-Slip Stiction (from sticking and friction) is a problem caused by the existence of two types of friction: static and dynamic. Static friction is nearly always greater than dynamic friction, which becomes a problem with packing friction because it is the most dominant form of friction found in most control valves. It can also be the result of cage style control valves that use piston rings for sealing and guiding. Stiction is recognized quickly if valve position and controller demand are plotted on the same graph (see Appendix A, Figure A-7). The following scenario is observed: The controller sends a pneumatic signal to the valve to correct a process deficiency. The actuator pressure changes, but the control valve does not move because of the static packing friction. Therefore, the process continues to drift away from the set point causing the controller to continue to change the input to the control valve. Finally, the actuator pressure is sufficient to apply enough force for the stem to overcome the static friction. It overshoots the desired position, however, because the smaller dynamic friction dominates. Of course, the over-correction becomes a process problem; the controller sends the opposite signal; and an overshoot occurs again. What is seen on the graph is a square wave valve position with a saw-tooth controller output. The two plots will be 180 out of phase. This situation can go on “forever. ” Another name for this kind of plot is limit cycle. Because some processes are very sensitive to striction, it can result in excessive instability and equipment trips. A positioner is one of the most effective ways to deal with it. The packing friction effects are not eliminated, but controlled for two reasons: A control valve with a positioner moves to the position demanded by the controller. Rapid and accurate positioning is assured by immediate stem feedback at the positioner. A control valve without a positioner receives only a pressure and relies on the process feedback loop to detect the change. Due to integral control action, the signal continues to increase until movement is detected through the feedback loop. For many processes, this delay in detection results in overshoot. 3-15


Some sources refer to the positioner as a proportional controller. This means that it will provide an output pressure to the actuator greater than the demand. This input pressure will continue to increase until either the proper movement is obtained or the pressure reaches supply pressure. This gain quickly overcomes the friction and, together with the feedback, mitigates the amount of jump in valve position associated with stiction.

3.5.3 Change Valve Response The positioner cam may be used as follows: To change the installed valve flow characteristic (for example, from linear to equal percentage) in order to optimize the process To linearize the valve flow characteristics instead of changing the valve trim To redistribute the gain of the valve (ratio of valve output to controller input) to achieve a better controllability in certain operating ranges.

3.5.4 Control Double Acting Actuator A controller has only one output, but the double acting actuator requires two inputs (and exhaust paths). Some positioners are designed (or can be) to provide two output pressures and two exhaust paths.

3.5.5 Increase Shutoff Forces The positioner can be used to drive the output pressure to zero or to full supply pressure (for example, 20 psig [137.9 kPa]) at the end of the valve travel, rather than just the signal pressure. In other words, without the positioner, only the signal pressure (for example, 3 –15 psig [20.7– 103.4 kPa]) would be available. If the valve uses a direct acting actuator (fail open), then the full supply pressure is available to shut the valve. In the case of reverse acting (fail close), the pressure is completely removed, allowing full spring force to load the seat at closure.

3.5.6 Split-Ranging Sometimes, it is desirable to control a process using two or more valves in parallel instead of one. The advantage of this is to reduce the range of flow controlled by any one valve and, thus, gain better control. To accomplish this method of control requires that the valves be opened in sequence using one control signal. A positioner can be adjusted to open and stroke a valve using any range of input signal pressures. Therefore, for a two-valve arrangement, the first valve fully opens after the input signal has changed only half of the full range. The second valve then opens and is fully opened at the full range. It is obvious that this can be extended to a three- or fourvalve arrangement if it is desirable.

3-16


3.5.7 Delays Due to Distance Between Controller or I/P Converter and Valve A pneumatic signal can take too long to get to the valve that it is controlling. This delay is referred to as lag, because the response begins immediately but takes time to reach the necessary pressure. For example, if a step change in pressure is sent down a 200-foot (60.96-m), 0.375-inch (0.9525-cm) inside diameter line, it takes 0.43 seconds for the value at the other end to reach 63.2% of the input value (see Figure 3-10). In some applications, this is a problem. Adding a positioner to the valve will greatly improve the situation.

Figure 3-10 Pneumatic Transmission Lag: Time to Reach 63.2% Final Value (Time Constant

)

3-17


4

CALIBRATION

4.1

Calibration and Condition Monitoring/Preventive Maintenance

The goal of calibration, or alignment, is to ensure that the positioner is functioning correctly within design tolerances, for example, start and stop points and travel. It ensures that the fidelity between the input signal and the position of the valve is established. Calibration also starts the condition monitoring/preventive maintenance (CM/PM) process (see Section 5). Finally, it sets the stage for more efficient and rapid troubleshooting. Basic calibration is an alignment process. When the process is completed, the actuator should operate in accordance with the manufacturer ’s specifications. This is also the opportune time to obtain baseline data that will be used for CM/PM and troubleshooting.

4.2

Basic Calibration

The basic calibration process consists of three steps. These steps may be performed together, but the method used must be able to discriminate problems in each step. The steps may be completed manually or by using data acquisition systems. A discussion of the latter is included at the end of this chapter. The three steps are: Bench set confirmation Feedback linkage alignment, including cam alignment Zero and span adjustment

4.2.1 Bench Set Confirmation Technically, the bench set is not part of the positioner alignment, but an improper bench set can affect a valve operation that might be associated with the positioner operation. Bench set is the nameplate specification used to verify a proper actuator operation. Bench set is expressed as a pressure range from the start of the actuator stroke to the valve ’s rated travel. Because actuator spring rates are not consistent, it is reasonable to assume that only one of the bench set points can be met. Therefore, spring adjustments must be made to meet the most important set point. On air-to-open valves, the start pressure is important for a valve that requires positive shutoff by the spring (the spring provides the seat load). On air-to-close, the end pressure value is important to have enough force to overcome the spring force and the valve friction and to seat the valve.

4-1

EPRI Licensed Material Calibration

When conducting the bench set test, the configuration of the actuator and valve is defined by the valve manufacturer. The most common configuration is to have the actuator disconnected from the valve stem. Some manufacturers require that the packing load be removed but that the valve plug and stem remain attached to the actuator. This method accounts for the weight of the stem and plug in the proper operation and sizing of the valve and actuator assembly. Because the valve forces are not present in many cases, the bench set pressure range is not the same as the pressure required to stroke the valve in actual service. Making the adjustment while the valve is in service may result in failure of the valve to fully stroke. Key Technical Point The bench set pressure range is not the same as the pressure required to stroke the valve in actual service. Making the adjustment while the valve is in service may result in unsatisfactory performance and/or make the valve inoperable.

4.2.2 Feedback Alignment The purpose of this alignment is to ensure that the zero and span of the feedback mechanism position correspond to the fully open (closed) and fully closed (open) valve travel positions. The positioner is not in service during the performance of this alignment. Either pressure is supplied directly to the actuator, or the hand-wheel is used to position the valve. Failure to perform this alignment may result in calibration problems, positioner performance issues, or component damage. Feedback alignment consists of the following: 1. Verifying that the installation is in accordance with appropriate vendor information and plant documents to determine specific requirements 2. Verifying that the positioner is mounted rigidly to the valve 3. Inspecting to determine that linkage is tight and that appropriate washers are installed 4. Checking and adjusting feedback linkage so that it is within allowable limits 5. Making sure there is freedom of movement throughout the range of travel so that the positioner is not “in the stops” at the ends of travel 6. Verifying that the appropriate cam is installed or, if the cam is a multiple type, that the feedback mechanism is on the correct range, for example, linear, square, or square root 7. Setting the feedback mechanism to the zero position on the cam 8. If applicable, setting the stroke lever parallel to the spring lever at midstroke 9. Stroking the valve and adjusting the appropriate linkages to ensure that the cam rotates in the correct direction, stops at the 100% position, and returns to zero as appropriate

4-2


Key O&M Cost Point Failure to perform feedback alignment may result in calibration problems, positioner performance issues, and/or component damage.

4.2.3 Zero and Span Adjustment

4.2.3.1

Purpose

The purpose of the zero and the span adjustments is to synchronize the valve position to positioner demand. At first, this seems to be just varying the input pressure from the bottom to the top of the control range (for example, 3 –15 psig [20.7–103.4 kPa]) and making sure that the valve strokes appropriately. However, a positioner can provide one additional advantage over an operation without a positioner. A positioner can ensure that at the extremes of the control range, the maximum force of the actuator is available. In other words, the positioner can be calibrated to give zero or full supply pressure rather than 3 psig or 15 psig (20.7 kPa or 103.4 kPa). This can be a distinct advantage in providing tight closure of a control valve to prevent seat damage.

4.2.3.2

Static Band

To obtain the maximum actuator force, pressure static bands are established at both ends of the pressure range. In reality, they are intentional dead bands, but the term dead band is not used here to avoid confusion with the dead band associated with control loops. The magnitude of these static bands does not have to be large. A static band of about 0.3 psi (2.1 kPa) allows for both instrument inaccuracies and positioner gain. This means that instead of expecting the valve to start opening at 3 psig (20.7 kPa) (or 15 psig [103.4 kPa ]), it should start opening at about 3.3 psig (22.8 kPa) (or 14.7 psig [101.4 kPa]) The result of this is as follows when the valve is in service. The valve is given a close signal by the controller. Assume that this is 3 psig (20.7 kPa). As the pressure reaches 3.3 psig (22.8 kPa) during the reduction to 3 psig (20.9 kPa), the valve is just shut. As the pressure goes to 3.2 psig (22.1 kPa) and below, the valve is not moving. In a sense, the stem is “locked” from going any further closed. Recall the black box discussion in Section 3. The error of 3.3 minus 3.0 psig = 0.3 psig (22.8 minus 20.7 kPa = 2.1 kPa) is multiplied by the gain of the positioner. The result is that the pressure to the actuator drops to 0 psig trying to make the valve respond to the lower signal. Hence, the valve is shut with maximum actuator force. For direct acting valves, the positioner is calibrated to provide seat closure at about 0.3 psig (2.06843 kPa) below maximum, for example, 14.7 psig (101.353 kPa). The same action occurs; only the actuator reaches the supply pressure value, for example, 20 psig (137.895 kPa). Key O&M Cost Point Establishing pressure static bands at the ends of the normal control range of valve operation results in tighter valve shutoff. Depending on application, this will lessen or eliminate seat damage and/or reduce megawatt losses.

4-3


4.2.3.3

Adjustments

Before beginning the zero and span adjustment, determine if the positioner is a motion balance type, for example, Fisher 3582 or Conoflow GP 50. If the motion beam has been repaired or otherwise affected by maintenance, then beam or flapper leveling is required before the zero and span adjustments. In addition, some models require an adjustment if the unit is changed to a split-range operation. The primary purpose is to ensure that the nozzle flapper approaches the nozzle to provide uniform variation of the control pressure with the flapper movement. Failure to do so results in the inability to properly calibrate the unit. Follow the manufacturer ’s technical manual for this leveling process. Key Human Performance Point Before zero and span adjustments, motion balance positioners require beam or flapper leveling if the motion beam has been affected by maintenance or if the unit is changed to split-range operation. Failure to do so will result in the inability to properly calibrate the unit.

Zero adjustments are always done in conjunction with span adjustments. In other words, after the span has been adjusted, the zero adjustment must be re-verified to ensure that the zero adjustment has not changed. The purpose of this guideline is not to provide zero and span adjustment instructions. These must be performed in accordance with the manufacturer ’s technical manual. However, because not all manufacturers provide for the establishment of static band, the steps below are considered supplementary guidance to be used in conjunction with the technical manual. If a cam is installed, it is assumed that the cam has been adjusted as described in the feedback alignment section. Key Technical Point Zero adjustments are always done in conjunction with span adjustments. In other words, after the span has been adjusted, the zero adjustment must be re-verified to ensure that the zero adjustment has not changed.

To obtain the static pressure bands, the following steps are recommended. It is assumed that in all other respects, the valve is ready for service. For example, the packing has been adjusted to the specified value; the stem/actuator adjustments have been made; the valve is free to stroke; and the travel is correct. 1. Verify that the supply pressure to the positioner is at least 5 psi (34.4738 kPa) higher than the upper signal range. However, do not exceed the 5 psi (34.4738 kPa) margin without referring to the technical manual. Some positioners may not operate properly and/or may be damaged by excessive supply pressures. 2. If required by the technical manual, stroke the valve to mid-travel to make the adjustments as required.

4-4


3. Slowly increase the signal from 0 psig according to Table 4-1 below, and make the adjustments as instructed by the technical manual. Note that these are suggested values to illustrate the establishment of static bands. They are not intended to replace values already in use that establish static bands, nor are they intended to be universal for all applications. 4. Repeat step 3, but decrease the signal. If necessary, continue until the values are obtained. Table 4-1 Suggested Static Band Pressure Values Signal Input (psig)* Action

Direct Acting Positioner

Reverse Acting Positioner

Positioner output pressure equals zero

3 (6)

15 (30)

Valve movement detected

3.1-3.5 (6.1-6.5)

14.5-14.9(29.5-29.9)

Valve fully stroked (no movement detected)

14.5-14.9 (29.5-29.9)

3.1-3.5 (6.1-6.5)

Positioner output pressure equals supply

15 (30)

3 (6)

Note: These are suggested values to illustrate the establishment of static bands. They are not intended to replace values already in use that establish static bands, nor are they intended to be universal for all applications.

4.3

Functional Check

When the calibration has been completed, functional checks should be performed. These consist of verifying the following: The operation is smooth and stable. All other accessories operate correctly. The valve intermediate positions correspond to the appropriate input signal.

4-5


Table 4-2 contains suggested intermediate median values for selected pressure inputs. Tolerances should be applied in accordance with the plant doctrine. Table 4-2 Suggested Intermediate Median Values for Selected Pressure Inputs Positioner Input (psig)

Percent of Span

Positioner Output Median Value (psig)*

3.3

0

3.3

6.3

3.3

6.3

3.3

6.3

4.4

10

4.4

8.6

3.4

6.5

6.9

13.7

5.6

20

5.6

11.0

3.8

7.2

8.4

16.8

6.7

30

6.7

13.3

4.3

8.4

9.5

19.1

7.9

40

7.9

15.7

5.1

10.0

10.5

21.1

9.0

50

9.0

18.0

6.2

12.2

11.4

22.8

10.1

60

10.1

20.3

7.4

14.7

12.1

24.4

11.3

70

11.3

22.7

8.9

17.8

12.8

25.9

12.4

80

12.4

25.0

10.6

21.3

13.5

27.2

13.6

90

13.6

27.4

12.5

25.3

14.1

28.5

14.7

100

14.7

29.7

14.7

29.7

14.7

29.7

Linear

Square Root

Square

Note 1: These are suggested values to illustrate the establishment of static bands. They are not intended to replace values already in use that establish static bands, nor are they intended to be universal for all applications. Note 2: 1 psig = 6.9 kPa

4.4

Data Acquisition Systems

The introduction of air-operated valve (AOV) data acquisition systems has revolutionized calibration, condition monitoring, and troubleshooting of AOVs and their accessories. AirOperated Valve Maintenance Guide, Section 6 [1] contains a detailed explanation of data acquisition system usage and the benefits that can be obtained from these systems. The calibration of positioners using a data acquisition system is vital if the condition monitoring and troubleshooting capabilities of these systems are to be fully realized (see Sections 5 and 6). Key O&M Cost Point The calibration of positioners using a data acquisition system is vital if the condition monitoring and troubleshooting capabilities of these systems are to be fully realized.

4.4.1 Calibration Using the Data Acquisition System The calibrations discussed above can be done using procedures that require the use of pressure gages, inputting control signals, recording information, and evaluating the information to compare it to some baseline value. Normally, maintenance personnel, aware of the importance of 4-6


gathering accurate information but lacking the proper tools, resort to such techniques as using thumbnails to detect valve motion and to determine if it is moving smoothly. It works but lacks consistency from person to person, and in today ’s environment of high capacity and high reliability, it may not be enough. In the hands of trained and experience personnel, the data acquisition system not only gathers data quickly and efficiently but does it consistently and accurately. All condition monitoring processes rely on the ability to trend the right data. The data acquisition systems available provide that ability. Figure 4-1 shows how a calibration plot generated by a data acquisition system looks after zero and span adjustments have been made.

Figure 4-1 Calibration of Valve Positioner

The typical ramp starts out at a minimum signal, A, and increases to the point when the valve starts to move, B. The output travel follows the input signal in a linear fashion until the valve plug or actuator contacts the open stop, C. From this point, the signal continues to increase without any change in the valve ’s position until point D is reached. When the input signal starts to decrease, it continues until it reaches point E, at which point the valve starts to close. The valve follows the signal down until the valve plug contacts the seat, F. Again, the signal continues to decrease until point A is reached, without any valve movement. Note that it is the valve travel (inches) that is being compared to the positioner input (pressure). While positioner output (pressure) is of interest, final calibration is always valve position versus positioner input; that is, the input signal controls valve position and is consistent with Tables 4-1 and 4-2. Figure 4-1 reveals that static bands have been used in setting up the positioner. Notice that movement does not take place until pressure is above 3 psig (20.6843 kPa), and all movement is completed before 15 psig (103.421 kPa )(static bands). It also shows two other problems 4-7


encountered: hysteresis and dead band. They have been lumped into the gap between the traces called dynamic error .

4.4.2 Hysteresis/Dead Band (Dynamic Error) Hysteresis results from the fact that all real devices bend, twist, or otherwise deform when they are stressed. Therefore, if they are forced in one direction, they will bend in one direction; and if forced in another direction, they will bend in that direction. In Figure 4-1, hysteresis is a result of the deformation of both the positioner and the valve actuator. However, hysteresis does not dominate the dynamic error. The dominant factor in the dynamic error is dead band. Dead band is the range that the input changes before the output begins to change after the input has been reversed. One of the causes of dead band is backlash, which is the result of looseness and free-play. In the above example, most of the travel from F to B and C to E is dead band. While dead band cannot be cured, it can be minimized by good design and attention to looseness and free-play. Dynamic error is a term that has been coined to describe hysteresis and dead band under dynamic conditions. Before the advent of the data acquisition system, the classic method had the pressure increased to a predetermined value and maintained until a reading was taken, that is, written down. The process continued until all data points were obtained. Such a method is described in ISA-75.13-1996, “Method of Evaluating the Performance of Positioners with Analog Input Signals and Pneumatic Output ” [2]. Hence, the hysteresis/dead band was obtained under static conditions, and these results may be different from dynamic conditions. Therefore, when describing data acquisition system results, the term dynamic error should be used. It is recommended that this practice should also be extended to the constituents, that is, dynamic hysteresis or dynamic dead band.

4.5

References

1. Air-Operated Valve Maintenance Guide, EPRI, Palo Alto, CA: 1996. NP-7412-R1. 2. ISA-75.13-1996, “Method of Evaluating the Performance of Positioners with Analog Input Signals and Pneumatic Output, ” Instrument Society, America, Research Triangle Park, NC: 1996.

4-8


5

CONDITION MONITORING/PREVENTIVE MAINTENANCE

5.1

Condition Monitoring

5.1.1 Continuation of Calibration Condition monitoring for positioners is primarily the continuation of the calibration process. The goal of calibration or alignment was to ensure that the positioner is functioning correctly within design tolerances, for example, start and stop points and travel. It ensured that the fidelity between the input signal and the position of the valve was established. Calibration must be done carefully and consistently for condition monitoring/preventive maintenance to be effective. The reason for this is that positioners particularly do not communicate impending failures well. Their performance on-line is masked by other parts of the control loop, and unless the positioner failure is catastrophic, degradation does not attract attention. What this means is that in addition to normal routine items discussed below, condition monitoring actions rely heavily on repeating calibrations. Persistent, significant calibration drifts can be used to identify the need for positioner repair or replacement. In fact, no other information might be available for this purpose because it is not cost-effective to internally examine the condition of accessories. Key O&M Cost Point Condition monitoring for positioners is primarily the continuation o f the calibration process. Calibration must be done carefully and consistently for condition monitoring/preventive maintenance to be effective. Persistent, significant calibration drifts can be used to identify the need for positioner repair or replacement.

5-1

EPRI Licensed Material Condition Monitoring/Preventive Maintenance

5.1.2 Data Acquisition Systems

5.1.2.1

Other Recommended Performance Curves Using a Data Acquisition System

Figure 4-1 shows only one of the numerous curves that can provide condition monitoring trends. As a minimum, the following inputs should be acquired when using a data acquisition system. Supply pressure Positioner input signal pressure (that is, I/P or E/P output pressure) Positioner output pressure Valve displacement I/P or E/P input (ma or v) Since a data acquisition system collects all this information on a time basis, it can easily correlate any or all of these parameters on a single plot. The first four of these parameters provide a set of graphs that give sufficient information to detect and, in many cases, identify the cause of unsatisfactory positioner performance. The last item, I/P, allows personnel to review the performance of the overall final control element (that is, the control valve and all attachments). The performance curves that are useful using pairs of the above inputs include the following: Table 5-1 Condition Monitoring Performance Curves Performance Curve

X Axis

Y Axis

Overall final control element

I/P (E/P) input

Valve displacement

I/P (E/P) performance

I/P (E/P) input

Positioner input signal pressure

Positioner performance

Positioner input signal pressure

Positioner output pressure

Valve performance

Positioner output pressure

Valve displacement

Note: The selection of the axes is arbitrary.

The inputs may also be displayed individually versus time. These curves are most useful when analyzing unsatisfactory or indeterminate problems found in the performance curves shown in Table 5-1. Key O&M Cost Point All condition-monitoring processes rely on the ability to trend the right data accurately. In the hands of trained and experience personnel, a digital acquisition system not only gathers data accurately but does it quickly, efficiently and consistently.

5-2


5.2

Preventive Maintenance

In a coordinated condition monitoring/preventative maintenance program, condition monitoring actions should not be used to detect failures that could have been prevented by planned maintenance actions. Therefore, PMs should be used to complement condition monitoring actions. On the other hand, the condition monitoring program can also be used to reliably predict problems and to allow proper planning and repairs when system conditions can be established without interrupting production. In addition, the condition monitoring program can be used to establish preventive maintenance actions and allow reviews to ensure that these actions are effective and timely. The following are PM actions that manufacturers and utilities have found effective in preventing failures: Air supply check Walkdown inspection Internal inspections (as applicable)

5.2.1 Air Supply Check 1. Verify air quality. a. Air quality affects air positioner performance. Some positioner manufacturers specify that the air quality should be instrument (IA) grade in accordance with ISA S7.3 [1] (replaced by ISA7.0.01 [2]). The particulate size that is permitted by this standard for IA is 40 microns. However, other manufacturers specify particulate as 3 microns. Therefore, even though most nuclear power plants meet the IA standard, supplementary filtering may be required and should be considered. b. The dew point may also be a factor and should be reviewed to ensure that the air supply is per the manufacturer ’s specification. c. The oil or hydrocarbon content can be detrimental to elastomers, thus increasing the likelihood of air leaks. 2. Blow down the regulator, and inspect the effluent for contaminants. 3. Verify the air supply pressure to detect regulator set point drift. 4. Inspect the regulator, tubing and fittings, and positioner for air leaks using a soap solution or equivalent. 5. Inspect tubing for cracks and kinking.

5-3


5.2.2 Walkdown Inspection The walkdown inspection is done on-line and confirms that the material condition of the positioner —including cleanliness, physical damage, mounting integrity, and loose or missing parts—is satisfactory. This inspection is performed about halfway through an operating cycle. A recommended checklist may contain the elements inspected in the following tasks: 1. Observe the exterior to detect signs of physical damage to the housing, connections, pipe/tubing, hoses, feedback and linkage. 2. Feel the housing, and carefully determine if it is firmly attached to the valve. 3. Inspect for air leaks. If possible, check all pneumatic connections with a soapsuds solution to detect any leakage. 4. Observe the pressure gages, and note if pressures are consistent with required supply pressures, expected signal pressure for valve position, and expected output pressure for the valve position. 5. If the feedback cam position is visible, determine if its position is consistent with the valve position. 6. Observe the valve motion to see if the response is consistent with the changes in signal/output pressure variation.

5.2.3 Internal Inspections (as Applicable) It is recommended that the air supply be isolated when performing the following internal inspections: 1. Inspect the pilot valves for sign of scarring and wear; clean or replace them as necessary. 2. Ensure that the pilot valve is properly aligned. 3. Inspect the flapper for signs of pitting or damage. 4. Inspect the nozzle for signs of erosion or damage. 5. Inspect the cam and cam roller for signs of wear. 6. Ensure that the cam roller rotates freely and is properly oriented. For example, for the Bailey AV series positioners, the etched side of the cam roller must be opposite the snap ring. 7. Ensure that the cam roller snap rings are installed properly. 8. Ensure that the proper cam is installed and correctly oriented.

5-4


9. Inspect the springs, and verify that all are connected and of the proper type. For example, some manufacturers use a color code to identify the spring. 10. Inspect the gaskets, O-rings, and filters. Clean or replace them as necessary. 11. Replace the over-ranged gauges. 12. Ensure that the bypass block, relay and gauge block manifold are tightened properly. 13. Clean the restrictions; for example, clean out the plunger.

5.3

References

1. ISA S7.3-1975 (R1981), “Quality Standard for Instrument Air, ” Instrument Society of America, Research Triangle Park, NC: 1981 2. ISA7.0.01-1996, “Quality Standard for Instrument Air, ” Instrument Society of America, Research Triangle Park, NC: 1996.

5-5


6

TROUBLESHOOTING

6.1

Introduction

Troubleshooting is the systematic approach to data collection, failure analysis, and the use of a test/measurement plan that collectively result in high confidence that the complete cause of system/equipment degradation has been corrected and that the system/equipment has been restored to normal operation. Ideally, troubleshooting is a continuation of condition monitoring/preventive maintenance. In other words, a problem is discovered by condition monitoring actions before it becomes a failure. However, even if it does not, the information captured in calibration and in setting up the condition monitoring program puts the troubleshooting effort in a much better state.

6.2

Approach

Positioners particularly do not communicate impending failures well. Other parts of the control loop mask their performance on-line. On the other hand, other control loop components may be the problem. In most cases, however, it is the final control element (consisting of the control valve, actuator and positioner) that is not working properly. Therefore, the approach to troubleshooting discussed below assumes that the final control element (FCE) has been acting erratically and that the positioner may or may not be the problem.

6.2.1 Use of Data Acquisition System The use of a data acquisition system may be of some use while the plant is on-line and the FCE is still in service. Some FCEs are equipped to have test equipment connected non-intrusively to monitor calibration points. Although the full range of FCE operations is not available, obtaining single point data or a small range of data may, after analysis, provide information on the proximate cause of the problem. This is particularly easy and efficient because with the data acquisition system, the troubleshooter can take the new data, plot it on the original calibration curves, and note any differences. On the other hand, troubleshooting with the FCE out of service may be the only option, in which case full dynamic curves will be available (see Appendix E, Positioner Problem Case Histories).

6-1

EPRI Licensed Material Troubleshooting

6.2.2 Symptom-Based Approach Walkdown and data gathering are key components of the sympton-based approach. Table 6-1 is a matrix of symptoms and their causes at the FCE level. The FCE level, not the entire control loop, is selected because FCE performance is clearly visible and positioner failure is always seen at the FCE. Table 6-1 provides guidance in two ways: It allows maintenance personnel to gather all data and information that is relevant. It allows these same personnel to analyze this data and focus on the most likely causes. Table 6-1 Final Control Element Problem Symptoms/Causes CAUSE

PROBLEM Erratic or Jerky Throttling

Actuator Supply Pressure Low

Failure to Fully Retract

Failure to Fully Extend

X

X

Cycling

X

Actuator Supply Pressure High

X

Actuator Supply Erratic

X

X

Unsteady Signal

X

X

Improper Bench Set Positioner

Sluggish/ Slow

Slow in Increasing Air Pressure Direction

Slow in Decreasing Air Pressure Direction

X

X

X 1

X

X

X

X

Wrong Travel Stops/ Calibration

Failure to Stroke

X X

X

X

Increased Packing Friction

X

Actuator Spring Too Large

X

Actuator Spring Too

X

Small Air Leak (Diaphragm, Stem Seal, or Case Joint) Leaks

X X

X

Solenoid Valve Failure

X

X

X

X

Air Supply Tubing Crimped, Too Small

X

Actuator Too Large

X

2

6-2

X

Piston Lubrication

X

X

Damage Cylinder or Piston Rings/ Seals

X

X

X

X

X

X

X

1

Can cause positioner to fail.

2

Lack of lubrication permits the actuator supply air to migrate t o the opposite side of th e piston.


6.2.3 Positioner When the positioner seems to perform unsatisfactorily, Table 6-2 provides a symptom-based approach to the positioner problem cause based on the outcome of Table 6-1. Because positioner designs vary, the table has been generalized; therefore, it must be used in conjunction with the positioner technical manual. The table also includes symptoms frequently encountered in calibration. Table 6-2 Positioner Problem Symptoms/Causes Symptom Erratic or jerky throttling

Possible Cause Feedback linkage from control valve element loose Excessive packing friction Positioner output pressure too low Internal air leaks, including pressure gauges Output signal generator spool valve sticking

Failure to stroke (not moving from minimum travel position)

Leak in signal circuit Cam reversed Airlines reversed (double acting actuator) or airline in wrong port (single acting actuator) Air supply starved (line too small or crimped) Output relay or pilot stuck/not functioning properly Actuator handwheel in wrong position

Failure to stroke

Output signal generator (pneumatic relay) leaking

(not moving from one end of stroke)

Supply breakdown orifice to nozzle-flapper chamber plugged Air line(s) in wrong port Actuator handwheel in wrong position

Slow in increasing or decreasing air pressure direction

Output pressure too high or too low Blockage in output lines or output valves Pneumatic relay not responding correctly Booster not adjusted properly.

6-3


Table 6-2 (cont.) Positioner Problem Symptoms/Causes Symptom Positioner not calibrating for full range.

Possible Cause Leak in signal circuit Improper feedback alignment Motion balance positioner not properly balanced Pre-alignment problem Actuator air leakage Air supply pressure incorrect Internal air leaks External handwheel incorrectly positioned Positioner in bypass Worn cam and/or cam roller Pilot valve or output relay not responding properly Starved air supply Vent paths blocked Cam not installed properly Wrong range spring

Control Valve Oscillating During Checkout or Calibration

Feedback linkage from control valve element loose Positioner gain too high Positioner output pressure too low Booster not adjusted properly Excessive packing friction Valve plug/stem separation Internal air leaks in pressure gauges Worn cam and/or cam roller

Excessive Air Consumption

6-4

Leakage at joints of manifold assembly or pneumatic relay assembly


A

CONTROL LOOP DETAILS

A.1

Purpose

Section 2, Introduction to the Control Loop, was intended to provide sufficient background to understand the role of the positioner in the control loop and to understand the positioner as a control loop in its own right. Appendix A is a continuation of Section 2. It is fundamentally a tutorial to give additional information of sufficient detail to persons who wish to understand the control loop better.

A.2

Control Loop Block Diagram

The level control process introduced in Section 2 can be explained better with a block diagram. Figure A-1 is a block diagram representation of Figure 2-1.

Figure A-1 Level Control Block Diagram

The control loop that is shown in Figure A-1 is a closed loop because there is a feedback path (through the feedback elements). On the control panel, the closed-loop operation is known as auto control. If the feedback path is broken, then the control loop is referred to as an open loop and the system is in manual control.

A-1

EPRI Licensed Material Control Loop Details

A.3

Loop Elements

The following paragraphs provide a summary discussion of the key elements of the control loop (see Figure A-1) that affect loop performance. A brief control loop symptomatic troubleshooting follows this discussion.

A.3.1 Sensor/Transmitter The sensor/transmitter signal is the feedback in the control loop and is therefore compared with the set-point signal to determine the error to be fed to the controller. Sensor/transmitter failures can be broken down into three types: 1. The signal sent to the comparator fails high, low, or as-is. 2. The signal does not accurately reflect the value of the controlled variable. 3. The signal has increasing time delays associated with it. The first type of failure is usually sudden and is identified by the operator as a loss of automatic control. Depending on the controller adjustments, failing high and low can be challenging to the operator because suddenly the error will be relatively large and the controlled variable will start going to a maximum value. The only recourse for the operator is to put the loop in manual as quickly as possible. Failing as-is normally appears the same except it is slower. Therefore, type 1 failures are seen as loss of control. The second type is characterized as a loop that does not operate at the set-point value. There is a signal, but it is wrong. It may be more difficult to detect at first if it involves a problem that develops slowly, such as a calibration drift. It may also occur if the calibration of the sensor or the zero/span of the transmitter is not correct. Therefore, type 2 failures are seen as operating in automatic but not at the set point. The third type of failure normally shows up as change in the loop response time or existence of the limit cycle. Because the response time of the loop can be the result of other components in the loop, this problem has no unique pointer and can be found only by the process of elimination. Problems that are associated with sensor/transmitters are: Improperly calibrated sensor or transmitter (zero, span) Transmitter filters that increase time delays Partially plugged pressure sensing line that causes time delays Fully plugged pressure sensing line Temperature sensor not located properly in thermowell Buildup of material (corrosion) of thermowell

A-2


A.3.2 The Controller (Including the Comparator) As discussed before, the set point and the feedback signal are combined in the comparator and an error signal transmitted to the controller. Because the output of a comparator is either plus, zero, or minus some value, the controller could be nothing more than a pneumatic servo that sent air to or bled air from an air operator. However, the controller is also used to “condition ” the error signal in such a way that the loop responds more rapidly and with greater accuracy to changes in the process disturbances. The way that a controller conditions (or processes) the error signal determines how a controller is classified. The classifications are proportional control, integral control, and differential control. All controllers in a power plant control loop use proportional control as a minimum.

A.3.2.1

Proportional Control

Proportional control provides an output that is proportional to the input. (In reality, the output is some fixed value when the error is zero.) The constant of proportionality is called gain, or Kc (see Figure A-3). An equivalent term is proportional band , which is equal to 100/Kc. The relationship between these terms is shown in Figure A-2. Proportional band is the span of the input that results in a complete or full change of the output. Span in this case refers to the difference between the highest and lowest input values.

Figure A-2 Proportional Band

In a way, proportional control acts like the simple servo discussed above, but has the added characteristic of having a gain adjustment . Since gain determines how large a signal goes to the final control element, gain has the effect of increasing the speed of the response of the manipulated variable to the disturbance of the controlled variable. This means that control is A-3


better by minimizing the error between the controlled variable and the set point. However, too much gain results in too large (rapid) a response, which causes problems in the controlled variable reaching the steady state value desired. The response will result in a loop that has unacceptable oscillations. These oscillations may be large but not increasing in amplitude, or they may increase in amplitude with time (see Figure A-4). On the other hand, too little gain results in sluggish performance and the inability to maintain the controlled variable at the desired value. Usually, the process engineer cannot predict the optimal gain setting; it must be done using the results of loop tests. Methods for tuning are identified later. The key point is that the gain settings must be reviewed anytime that the loop is experiencing control problems involving unacceptable cycling or sluggish response. Key Technical Point Gain settings must be reviewed anytime that the loop is experiencing control problems involving unacceptable cycling or sluggish response. Usually, the process engineer cannot predict the optimal gain setting; it must be done using the results of loop tests.

Figure A-3 Controller Output Response to Square Pulse Showing Gain

Figure A-4 Stability

A-4


A.3.2.1.1

Offset

One of the problems associated with proportional control is offset. Offset refers to the difference between the controlled variable and the set point, that is, a steady state error. The reason for this steady state error is that at zero input error, all controllers are adjusted (biased) to provide a fixed output (value) that has been adjusted to maintain the controlled variable at the desired value for the original conditions, for example, disturbances. To provide any other output requires the existence of an error at the input. The required error can be made smaller by increasing controller gain, but this may cause instability or other problems.

A.3.2.2

Integral (Reset) Control

The solution to the offset problem is to modify the output of the proportional controller so that it accommodates the required controlled variable but with a zero error. In effect, the fixed output referred to above is reset with an integral controller action in which the output is proportional to the time integral of the error as long as the error is present. In other words, for any particular error, the output of the controller continues to increase (or decrease) at a constant rate. When the error is zero, the increase (decrease) stops, and the new reset value exists at the output (see Figure2-4). Proportional-integral (PI) controllers are very common in power plant applications. To change the integral control, adjust the reset time, (Ti) or its reciprocal repeats per minute (1/Ti).

A.3.2.3

Derivative (Rate) Control

Derivative control, or rate control, is not frequently used in power plant applications. It is always used with proportional controllers (PD) and usually in combination with integral control (PID). In derivative action alone, the output of the controller is proportional to the rate of change of the error signal and returns to zero when the error rate is zero. To adjust the derivative control, use the rate adjustment. Therefore, if the error signal increases (decreases) rapidly, the controller output changes by an additional amount that will rapidly change to the steady state value (see Figure A-3). Use derivative control when the controller gain cannot achieve satisfactory process control without instability. This usually occurs when dead time is too large. Derivative control also improves slow loop response that is caused by lag.

A.3.2.4

Controller Problems and Tuning

Most improperly tuned controllers result in control loops that are unstable, cyclic, and/or have a slow recovery rate. Most controller problems result from trying to make the adjustments of gain, reset, and rate by trial and error. Not only can this be unnecessarily time consuming, it can result in totally incorrect controller settings. Trial and error is not necessary either. Proper tuning methods that achieve reasonable success have been in existence for years. These include:

A-5


Ultimate sensitivity method: performed in automatic control (closed loop) Reaction curve (Ziegler-Nichols): performed in manual (open loop) Minimize integrated absolute error method: performed in manual (see Appendix F) However, even using the “right” method may simply disguise another problem in the loop and result in de-tuning the controllers to compensate for this deficiency. Therefore, tuning should proceed only after verifying the satisfactory performance of the other components in the loop.

A.3.3 The Final Control Element The final control element (FCE) is the component or group of components that adjusts the manipulated variable to satisfy the controlled variable set point. For example, it could be an electric heating element along with power source, relays, or other components that maintains the temperature of a fluid. In most cases in a power plant, it is the valve, actuator, and positioner. The FCE for the example of Figure A-1 is the control valve, actuator, and positioner. This grouping of components is effectively a system of components that counters the effects of the process disturbance. It is appropriate that the FCE is the last component of the control loop to be discussed. Control loop problems are discussed in the next section. It is generally agreed in the industry that if the FCE is a control valve, it is the single largest contributor to poor control loop performance and destabilization of the process. The control loop problems that are discussed in the next section and the problems that make a difference are the result of the FCE. One of the most important jobs that a positioner does is to mitigate some of these problems.

A.4

Control Loop Problems

A.4.1 Dead Time Dead time is the time it takes a process to start to produce an output in response to an input. For the level control example in Figure A-1, it is the time that it takes for the flow (manipulated variable) to start to change in response to a level change (controlled variable). In a flow control loop, it is the time for the flow to start to change when the flow controller detects an error. All closed-loop processes have dead time. Some are worse than others. Dead time is the primary reason why even a properly tuned closed loop still has a tendency to have some oscillations in the controller output. Dead time causes these oscillations by delaying the response to a disturbance. The time delay allows the process to go further away from the set point before the correction takes effect and starts to reduce the error. This results in an over-correction and causes the process to deviate from the set point in the opposite direction. This can go on endlessly. Just as the dead time varies from loop to loop, so does the magnitude of the oscillations. Unacceptable oscillations are referred to as instability.

A-6


A.4.1.1

Sources of Dead Time

The causes for dead time are the process and the components that make up the process. The major factors in process dead time are the distance from the sensor that monitors the controlled variable to the final control element, and the time that it takes for the manipulated variable to start to affect the controlled variable. Therefore, process dead time is inherent in the design of the process.

A.4.1.2

Component Dead Time

Assuming that process dead time has been minimized by proper design, then it is the component dead time that needs to be minimized. Most component dead time is the result of device dead bands. In particular, dead bands of any consequence are the result of mechanical devices. Dead bands in electronic devices are virtually non-existent although they are subject to lag (see section 4.1.5, Lag). Dead band is the range that the input changes before the output begins to change after the input has been reversed. Backlash caused by dead band in mechanical devices, such as positioners and other final control elements, is the result of looseness and free-play. In Figure 5-1, most of the travel from F to B and C to E is backlash, which is referred to as dead band. While dead band, or backlash, cannot be cured, it can be minimized by good design and attention to looseness and free-play. Another source of dead time is friction. Friction-induced dead time is caused by the difference in static and dynamic coefficients of friction for the valve assembly packing. This problem is so common that it has been given such names as stiction, or stick-slip. Friction-induced dead time is not dead band. It occurs any time that motion is demanded, regardless of direction. In addition, there is no “taking up the looseness ” motion as with backlash. It is a pure time delay until the forces are great enough to cause the motion. It is sometimes called a non-linearity. Regardless of this distinction, it can be just as destructive to process control; and sometimes, it can be more destructive if packing tightening practices are not controlled.

A.4.1.3

Identifying Dead Time Problems

For processes in which the final control element is a control valve, dead time shows up in comparing the trace of manipulated or process variable (the process feedback) to the controller output. What occurs is termed limit cycle. Figure A-5 illustrates the characteristic controller sawtooth output and square wave process variable variation. The cause of this is as follows: When the valve receives a signal to move, nothing happens at first because of dead band. With integral control, the controller increases the signal as long as the error is present. At some point, the valve begins to move; but because of the integral action, it goes further than necessary and stops. Now the controller detects the error or the overshoot (through the feedback loop) and starts to correct in the opposite direction, and the process continues. With friction induced dead time, or stickslip, the jump can be substantial, hence the concern.

A-7


A.4.1.4

Dealing with Dead Time

Obviously friction-induced dead band can be dealt with through proper packing tightening practices, and stem lubrication . Reviewing the packing system to determine if lower friction packing systems are available may be of some help. Backlash is handled through attention to looseness and free play. Once these causes are corrected, then the only option remaining is to reduce the component response time, which is called lag. To do this, turn up the gain. The gain can be increased at each of the control loop components, that is, the feed back sensor/transmitter, the final control element, and the controller. If the sensor relies on mechanical components, then replacement with electronic versions, such as level sensors, may help. If it is a pressure or level sensor that uses pneumatic lines, then shortening them (the lines? )may help. Perhaps the gain of the transmitter can be improved. However, most of the time, the sensor/transmitter gain cannot be improved enough to make a difference.

Figure A-5 Process Variable Versus Controller Output Showing Stiction (Limit Cycle)

Because the final control element is a valve, ways to improve gain involve speeding up the response of the valve and speeding up the response of the process-to-valve motion. Almost the only way to speed up valve response time is to add a positioner. If a positioner is installed, then increasing the gain of the positioner is also possible. Obviously, from a maintenance standpoint, if the valve has a positioner, then it should be verified as operating correctly. Finally, turning up the controller gain provides some relief from dead band problems. This causes larger outputs that causes a larger response at the final control element. In addition, if the controller has integral processing, then increasing the reset helps, too. Increasing gain helps by speeding up the process to compensate for the delay. Nevertheless, gain can go only so far. Too much gain causes the system to over-respond and results in instability that can vary from unacceptable oscillations to outright system crash. A-8


One of the best examples that show the relationship between dead time, process gain, and stability is shown in a “robustness ” plot (see Figure A-6). It shows the stability region for various values of process gain versus process dead time. Each system has its own unique plot, although they always appear as a decreasing, exponential-like, curve. These plots are generated from tests made with the system in service. The plot shows the limitation of gain. The greater the dead time, the less gain helps.

A.4.1.5

Lag

It is important to differentiate between dead time and lag. Lag is how long a piece of equipment takes to respond. Dead time is how long it takes before the equipment begins to respond. Even though both appear as a time delay, this difference is crucial (for the effect of gain on dead time, see the section above). For lag, however, gain is the answer; that is, to solve a lag problem, gain can be increased without limit. The reason for this is that the beginning of the response is immediate. The error begins to be reduced instantly. Perhaps not by much, but it happens now. To improve this situation, just make it happen faster until it is satisfactory. Faster means gain.

Figure A-6 Robustness Plot

A.4.2 Linearity All control loop designs and analysis assume that the loop is linear. A linear control loop responds to its input signal with the same dynamic response —gain, time constants (lag), and dead time—regardless of the size of the change in the input signal. Sensor/transmitters, comparators, and controllers are electrical devices and by design can be made essentially linear. A well-known example is the use of the square-root extractor that converts a delta P of a flow orifice to an output directly proportional to flow.

A-9


If the loop is not linear, the controller settings are optimal for only a particular system response, that is, a particular dead time and lag. If the process conditions differ (for example, the disturbance changes), the response to the output of the controller is different. If this means longer dead time or lag, it may result in instability and loss of control. The three primary sources of non-linearity in a control valve system are: Inherent and installed flow characteristics Hysteresis/dead band/stem friction Dynamic error

A.4.2.1

Inherent and Installed Flow Characteristics

All control valves can be described by a flow characteristic curve that relates flow (as a percentage of full flow) to the stroke of the valve (as a percentage of full stroke). Figure A-6 shows three of the more common types that are commercially available. It is important to understand that these inherent characteristics curves are based on the valve as the only source of pressure drop in the line. When other sources of pressure drop become larger with flow (or with a centrifugal pump supplying motive force), these curves change for the installed valve and tend to look more like the curves that are to the left of the inherent curve. For example, equal percentage tends toward linear. It is these installed characteristic curves that must be considered when determining the linearity of the flow versus travel. In a properly designed system, the installed characteristic curve is linear.

Figure A-7 Valve Flow Characteristics

To ensure uniform control loop stability through the anticipated valve-operating ranges, valve characteristics should be specified so that the valve response match the dynamics of the process. A quick opening valve characteristic is desirable for on-off control. A linear characteristic (flow rate directly proportional to valve travel) is good for most flow control and liquid level control A-10


systems. An equal percentage characteristic is best suited to pressure control or flow control systems with wide-swinging pressure drops.

A.4.2.2

Hysteresis/Dead Band/Stem Friction

Hysteresis results from the fact that all real devices bend, twist, or otherwise deform when they are stressed. Therefore, if they are forced in one direction, they bend in one direction; if forced in another direction, they bend in another direction. Hysteresis is part of the dynamic error and is a result of the deformation of both the positioner and the valve actuator (see Figure 4-1). However, hysteresis does not dominate the dynamic error unless something is wrong. In fact, hysteresis is usually quite small. The dominant factor in the dynamic error is dead band. Dead band and stem friction effects not only cause dead time but also cause non-linearity. Dead band is defined as the range that the input changes before the output begins to change after the input has been reversed. The non-linearity occurs because the response to the changes is different (though predictable) each time that the stem changes direction. Because stem friction also causes the same induced dead time, it is a cause of non-linearity and, thus, similar to dead band.

A.4.2.3

Dynamic Error

Dynamic error is a term that describes hysteresis and dead band under dynamic conditions. Before the advent of the data acquisition system, the classic method of plotting such traces had the pressure increased to a predetermined value and maintained until a reading was taken, that is, written down. The method continued until all data points were obtained. Such a method is described in ISA-75.13-1996, “Method of Evaluating the Performance of Positioners with Analog Input Signals and Pneumatic Output ” [1]. Hence, the hysteresis/dead band was obtained under static conditions, and the result might be different from dynamic conditions. Therefore, when describing data acquisition system results, the term dynamic error should be used. It is recommended that this practice should also be extended to the constituents, that is, dynamic hysteresis or dynamic dead band.

A.5

Scaling Calculations

Scaling calculations are a good place to find out about the control loop and the basis for the setup. These calculations were developed to provide the values for calibrating plant instruments, constants for function cards, and electrical equivalents for set points. These calculations provide a wealth of information about zero, span, sensor location, and corrections to sensor values. While calibration procedures are normally written around the scaling calculations, it is good to review them as an aid to troubleshooting. They may also contain diagrams or provide the location of the input documents used in the calculations.

A-11


A.6

Reference

1. ISA-75.13-1996, “Method of Evaluating the Performance of Positioners with Analog Input Signals and Pneumatic Output, ” Instrument Society of America, Research Triangle Park, NC: 1996.

A-12


B

AOV POSITIONER CHECKLIST

The checklist on the following pages is courtesy of Progress Energy ’s Harris Nuclear Plant and is reproduced without modification.

B-1

EPRI Licensed Material AOV Positioner Checklist

Using Digital Acquisition System EQUIPMENT TAG POSITIONER MANUFACTURER POSITIONER SERIES / TYPE / MODEL NO. General Notes:

1.

a.

Circle as found conditions as Sat (Satisfactory) or UnSat (Unsatisfactory) N/A items not applicable Refer to the attached figures as needed

Collect the following As Found data: Input (mA) Point where valve begins to move Mid-position of valve travel

Positioner Input (PSI)

Positioner Output (PSI)

Stroke (Inches)

b. c. d.

Full Open or Full Closed position Point where valve begins to move Mid-position of valve travel

*

*

**

Point where valve returns to it's starting position

*

*

**

e. f.

* This is a critical reading, which should f all within the allowable ranges listed in step 10 ** This reading should equal either zero or supply pressure, depending on starting position of valve and if direct or reverse acting

2.

B-2

Cycle valve 3 times.

EPRI Licensed Material AOV Positioner Checklist 3.

Collect the following data: Input (mA)

a.



Stroke (Inches)

Point where valve begins to move Mid-position of valve travel

b. c. d.


*

*

**

Point where valve returns to it's starting position

*

*

**

e. f.

* This is a critical reading, which should fall within the allowable ranges listed in step 10 ** This reading should equal either zero or supply pressure, depending on starting position of valve and if direct or reverse acting

4.

Circle the failure position of the valve: Fails Closed

Fails Open

Fails As-Is

5.

Check overall physical condition of the Positioner (inside and outside): a. Securely mounted Found Sat / UnSat Notes: b. Physical damage Found Sat / UnSat Notes: c. Foreign material such as moisture, oil, insects Found Sat / UnSat Notes: d. Loose, missing, damaged parts Found Sat / UnSat Notes:

6.

Inspect tubing to and from Positioner for leaks and kinks. Found Sat / UnSat Notes:

7.

Snoop all fittings and Positioner components (inside and outside) looking for leaks other than from vents. Found Sat / UnSat Notes:

8.

Measure and record As-Found output of regulator supplying Positioner. Adjust as needed and record As-Left value. As Found Measured Value: psig Desired Value (see below): psig As Left Measured Value: psig The desired value should be determined from the following, in order of preference: PassPort Equipment Data Plant Drawing Vendor Manual Set equal to 5 to 7 psi greater than the Positioner operating pressure provided the pressure is less than the maximum pressure rating of actuator. HNP AOV Program Manager. Notes:

9.

Inspect and replace any worn or damaged parts: a. Cam (if used) Found Sat / UnSat Notes: b. Cam Follower (if used) Found Sat / UnSat Notes: c. Nozzle (if used) Found Sat / UnSat Notes: d. Flapper (if used) Found Sat / UnSat Notes: e. Pilot Valves (if used) Found Sat / UnSat Notes:

B-3

EPRI Licensed Material AOV Positioner Checklist f.

Relay (if used) Found Sat / UnSat Notes:

10.

If data collected in steps 1 and 3 are acceptable (for points c and f Inputs; see table below) and no components were found unsatisfactory, adjusted or replaced; perform the following: a. N/A steps 11 through 14. b. Proceed to step 15. Which Data to use for points c and f Inputs will depend on the type of signal, starting position of the valve and if direct or reverse acting Type Signal Desired Input Allowable Range 4 to 20 mA 4.66 4.34 to 4.98 4 to 20 mA 19.44 19.12 to 19.76 3 to 15 psi 3.50 3.26 to 3.74 3 to 15 psi 14.50 14.26 to 14.74 6 to 30 psi 6.50 6.02 to 6.98 6 to 30 psi 29.50 29.02 to 29.98

11.

Verify feedback mechanism/linkage as follows: Component condition a. Found Sat / UnSat Notes: b. No lost motion Found Sat / UnSat Notes: c. Proper orientation at valve mid position Found Sat / UnSat Notes:

12.

Perform the following: a. Apply required input for start of travel and adjust applicable Positioner setting (zero or span/range). b. Apply require input for end of travel and make any necessary adjustments. c. Repeat steps a and b until acceptable setup values are obtained.

13.

Verify all setscrews and locking fasteners are tightened securely.

14.

Collect the following As Left data: Input (mA)

a.



Stroke (Inches)

Point where valve begins to move Mid-position of valve travel

b. c. d.


*

*

**

Point where valve returns to its starting position

*

*

**

e. f.

* This is a critical reading which should f all within the allowable ranges listed in step 10 ** This reading should equal either zero or supply pressure, depending on starting position of valve and if direct or reverse acting

15.

If step 3 values are acceptable AND step 1 values are not acceptable; send a copy of Data Sheets to HNP AOV Program Manager.

COMMENTS:

B-4


C

GLOSSARY

amplifier. A component that allows an input signal to control a power source that is independent of the signal. back lash. The movement in the output that is observed after a motion has been reversed. It is result of looseness and free-play. beam. The summing point of a control valve positioner. bellows. A flexible cylinder that expands or contracts to convert pressure to force or motion. It is often used in positioners to position the beam. bench set. A specification that is used to verify proper actuator operation. Bench set is expressed as the pressure range from the start of the actuator stroke to the valve ’s rated travel. calibration. The act of adjusting the output of a device so that it corresponds to the value of the input of the device. In positioner calibration, zero, span, and crossover (balance) pressure are the primary calibration adjustments. closed-loop control. Pertaining to a system with a feedback type of control in which the output is used to modify the input. controller. The part of the control loop that takes the output of the summing point (error) and produces an output signal response for the final control element that will minimize the error. controlled variable. The variable of the process that is to be controlled. It may or may not be the manipulated variable. dead band. Dead band is the range that the input changes before the output begins to change after the input has been reversed. One of the causes of dead band is backlash, which is the result of looseness or free-play. dead time. The time required for a system to begin to respond to an input. Compare to lag. derivative mode. One of the three basic control modes. It produces a change in the output signal from a controller as a function of the time rate of change of the measured variable with respect to the set point. It acts only when the variable is changing and ceases to act when the variable stops moving. It is also know as rate action or rate response.

C-1

EPRI Licensed Material Glossary

direct-acting actuator. An actuator in which the actuator stem extends toward the control valve in response to an increasing input signal. It is used for fail-open valve applications because the loss of the air supply results in the valve going open. direct-acting positioner. A positioner that provides increasing output pressure in response to an increasing input signal. directional control valve. A valve that provides porting that allows it to direct flow to an actuator that causes the associated valve to move in a certain direction. Sometimes known as a pilot valve. double-acting positioner. A positioner that has two relays. It is typically used with doubleacting and spring bias designs. dynamic error. A special term that equals the sum of hysteresis and dead band when they are measured during dynamic or moving conditions. dynamic response. The response characteristics of an element, including phase shift and dynamic attenuation, observed as the frequency of the input is varied. E/P converter. A transducer that uses voltage (typically 0 –10v) for an input and pressure for an output (see I/P converter, transducer). equal percentage characteristic. The inherent flow characteristic, which for equal increments of rated travel gives equal percentage changes of the flow coefficient Cv (see Figure A-7). error. The output of the summing point; the difference between the set point value and the measured value. fail close. See reverse-acting actuator . fail open. See direct-acting actuator . feedback. A feature of control loops in which the output signal of a process is fed back through a mechanism that modifies the input to the controller. The feedback may be negative if it tends to decrease the output or positive if it tends to increase the output. Feedback element. The device in a control loop that acts to modify the input to the controller. final control element. The device that directly changes the value of the manipulated variable of a control loop. flow characteristic. The relation between the flow through a valve and the percent rated travel as the latter is varied from 0 to 100%. This is a special term. It should always be designated as inherent flow characteristic or installed flow characteristic. Common flow characteristics are linear, equal percentage, and quick opening. See inherent flow characteristic and installed flow characteristic. (See Figure A-7)

C-2

EPRI Licensed Material Glossary

flow coefficient, fluid (Cv). The number of US gallons per minute (gpm) of 60 °F (15.6°C) water that will flow through an orifice (valve) with a 1-psi (6.9-kPa) pressure drop. gain. An expression used to denote device output as related to input. Gain is usually a dimensionless value. hysteresis. The phenomenon exhibited by a system whose state depends on its previous history. It is seen as a non-linearity of an output resulting from application of an input in a different direction or sense. Particularly, it is the different device response that occurs when the force is reversed or applied in a different direction. inherent flow characteristic. The flow characteristic when pressure drop across a valve varies by the flow and related conditions in the system in which the valve is installed. input. The signal applied to a device, that is, a controller, that results in an output. integral mode. One of the three basic control modes. It produces a continuous change in the output signal from a controller as long as the measured variable deviates from the set point. The rate of change in the output signal depends on the amount of deviation, the proportional bandwidth, and the setting of the integral mode adjustment. The output ceases to change when the variable returns to the set point. It normally results in a new final output different from what existed before the action began. By so doing, it eliminates offset. This mode is also known as reset response. The adjustment for integral or reset may be calibrated either in repeats per minute or minutes per repeat. I/P converter. A transducer that uses current (typically 4 –20ma) for an input and pressure for an output (see E/P converter, transducer). lag. The time it takes for a device to achieve some output value after it begins to respond. Normally, this is expressed as a time constant, (tau), which is the time to reach 63.2% of the final value. Compare to dead time. limit cycle. A nonlinear control loop instability. Can be caused by friction-related dead band in the control valve, for example, stiction. linear characteristic. For control valves, a characteristic that produces a percentage of maximum Cv that is directly proportional to valve stem position as a percentage of full travel. linear flow characteristic. An inherent flow characteristic that can be represented ideally by a straight line on a rectangular plot of flow versus percent rated travel (equal increments of travel yield equal increments of flow at a constant pressure drop) (see Figure A-7). linear process. A process in which gain is constant regardless of load. manipulated variable. The process input (quantity, property, or condition) that is adjusted to maintain the controlled variable at a desired set point.

C-3

EPRI Licensed Licensed Material Material Glossary

measured variable. The physical quantity, property, or condition that is to be measured, for example, temperature, pressure, flow, level, speed, or weight. It is a function of the controlled variable. non-linear process. A process in which gain changes at different loads or flow conditions. nozzle-flapper amplifier. A pneumatic signal amplifier in which a change to the position of the flapper, relative to the nozzle, results in a change to the output pressure. offset. The difference between the controlled variable and the set point that occurs when a control loop operates at a point different from the point for which it was calibrated. It occurs with control loops that do not use integral mode control. output. A particular process quantity or variable that has been identified as the result of the values or actions of one or more other process values defined as input. phase shift. For a device, the amount of time that an output lags behind an input. One of the causes is dead band. pilot valve. See directional control valve. pneumatic. Pertaining to or operated by air or other gas. positioner. A device that compares the actual valve position with the desired position with respect to an input signal and adjusts actuator-loading pressure until the desired valve position is attained. The desired position has a predetermined relationship to the input signal. pressure drop. The difference between the upstream pressure and the downstream pressure. It represents the amount of flow stream energy that the control valve must be able to absorb or withstand. pressure drop, maximum allowable. The maximum flowing or shutoff pressure drop that a control valve can absorb or withstand. While maximum inlet pressure is commonly dictated by the valve body, the maximum allowable pressure drop is generally limited by the internal controlling components (for example, liner, disk, shaft, bearings, or seals). The maximum allowable pressure drop may apply to the pressure drop while flowing process fluids are at shutoff. process gain (static). The dimensionless ratio of the magnitude of change in the process variable (PV) to the magnitude of the change in input to the process. The change in input is usually measured at the controller output (CO); therefore, process gain is equal to PV/CO where PV and CO are in the same units. proportional band. The change in input required to produce a full range in output due to proportional control action.

C-4


proportional control mode. The most basic control mode in which the controller output signal is proportional to the amount of deviation between the measured variable and the set point, that is, error. The adjustment for proportional response is either in terms of gain or proportional band (equals reciprocal of gain times 100). quick opening. A characteristic that, for equal changes in stem position, provides a large change in Cv at low lifts and smaller changes in Cv at high lifts (see Figure A-7). quick opening characteristic. An inherent flow characteristic in which there is near maximum flow with minimum travel. range. The region between the upper and lower values of the variable in question, expressed in terms of the lower and upper range values, for example, 3 –15 psig (20.7 –103.4 kPa), 4 –20 ma, etc. (see span). rangeability. The ratio of maximum to minimum flow within which the deviation from the specified inherent flow characteristic does not exceed some stated limit. For example, for a linear flow characteristic, the range of flows for which the flow characteristic remains linear. This is usually somewhere in the central part of valve lift. Rangeability is expressed as the ratio of the maximum controllable flow coefficient to the minimum controllable flow coefficient of a control valve. rate action (rate response). See derivative mode. rated Cv. The value of Cv at the rated full-open position. relay. A device that receives one or more signals, modifies these signals, and/or changes their form (see transducer ) to produce one or more outputs. relay, pneumatic. A pneumatic power amplifier in which change in input pressure result in changes in the position of exhaust and supply valves that control a separate supply valve. reset (response). See integral mode. reverse-acting actuator. An actuator construction in which the actuator stem retracts from the control valve with increasing pressure. Used for fail-close valve applications because the loss of air supply results in the valve going closed. saturation. The point in an operating range at which a change in input no longer causes an output change. sensor. A transducer that is used to measure or monitor a physical variable. set point. The desired value for the controlled variable. The set point is an input variable that can be manually or automatically set and is expressed in the same units as the controlled variable, for example, degrees F or psig.

C-5


single-acting positioner. A positioner with only one relay, typically used with spring-return actuator designs. span. The difference between the upper and lower range values variable in question, expressed in terms of a single value, for example, 12 psig (82.7 kPa) or 16 ma (see range). span adjustment. The calibration procedure that establishes the control valve position desired when the input signal is at the maximum value of the input signal range (see zero adjustment ). ). split range. A technique in which one controller is used to operate two or more final control elements. static gain. The ratio of change of steady-state output to a change in input. static open-loop gain. The static gain of an entire system, measured by opening the loop and determining the ratio of the change in output to a change in input. It is the product of the gains of all system elements. stiffness. The resistance of the actuator to forces that tend to destabilize its position. summing point. Any point at which the signal are added algebraically. In a control valve positioner, summing is usually accomplished with a beam. It can also be a diaphragm. transducer. A device that accepts an input in one form (for example, pressure or electrical current) and provides a corresponding output in another form. zero adjustment. The calibration procedure performed to establish the desired valve position when the input signal is at the minimum value of the input signal range (see span adjustment ). ).

C-6


D

DIGITAL POSITIONERS

D.1

Introduction

As pneumatic controls revolutionized the control industry by providing better process control, digital applications will allow an even greater advance. The digital positioner has gained sufficient recognition that most of the pneumatic positioner manufacturers now offer digitalbased equipment.

D.2

Reason for Digital

In addition to the abilities of the pneumatic positioners discussed in the guideline, digital positioners extend these abilities and benefits in such a way that future replacements may be almost irresistible. Listed below are some of these advantages.

Valve Control Improvements Characterization is provided directly in the output signal. Just as the pneumatic positioner characterizes using the feedback elements (cam) to deal with non-linearities, digital does so in the output. Since the characterization take place at the beginning, it controls the process directly and, therefore, more quickly. In addition, characterization setup is simply pushing a button. Valve stroke speeds can be limited. This may be useful for instances in which the fast valve can cause water hammer. Valve travel limits can be programmed. If a valve is not supposed to close, the positioner can be programmed to provide this limitation.

Improved Calibration Zero and span calibration can be performed in a few minutes. With installation of databases, remote automatic calibration can be performed. With the pneumatic positioners, test equipment hook ups and data taking could result in calibrations taking several hours.

D-1

EPRI Licensed Material Digital Positioners

On-Line Condition Monitoring Positioners do not communicate impending failures well (see Chapter 5). With digital technology, positioners can generate alarms to an operator interface. These alarms can be local, with the addition of a data bus, these alarms can be sent to a central operator interface. Maintenance can assign positioner-based alarms such as valve travel deviation from the input signal, over-travel, etc. Maintenance related data can also be provided. Digital positioners can track valve reversal and total stem travel data that can be correlated with time and actual maintenance events to improve predictive maintenance forecasting. Installed valve performance can be monitored and trended. All the positioner parameters that were captured by the DCS can be captured by the digital positioner. Because they are almost completely electronic in nature, they are less susceptible to vibration. On the other hand, they would be susceptible to electrically induced failures associated with all digital devices.

D-2


E

POSITIONER PROBLEM CASE HISTORIES

E.1

Introduction

The case histories included here were selected from Appendix C of EPRI Report NP-7412R1, Air-Operated Valve Maintenance Guide [1]. They are based on a number of sources, and although the text may have been edited, the conclusions or results were solely those of the contributors and were not verified by the authors of NP-7412R1. The histories that follow here are only those identified as positioner-related.

E.2

Case Histories

Case 1 - Positioner Bellows

Figure E-1 Severely Deformed Positioner Bellows (Input Signal Converter)—Travel Versus Positioner Input Pressure

E-1

EPRI Licensed Material Positioner Problem Case Histories

This graph shows the positioner performance (pressure versus travel). A Fisher 546 I/P has a 6 – 30 psig (41.4–206.8 kPa) output pressure feeding a 3582 positioner with a 3 –15 psig (20.7–103.4 kPa) bellows. The positioner bellows was severely deformed. The valve position is distorted and is not uniform for any positioner pressure over 15 psig (103.4 kPa). Compare these results to the response for pressures less than 15 psig (103.4 kPa).

Case 2 - Stability Problem Double Acting Cylinder The valve is a globe valve with a double acting cylinder actuator. The reported problem was unstable control during plant operation. The data acquisition system testing was done using 150second increasing and decreasing ramps, with a 10 –50 milliampere control signals input into the I/P (see Figure E-2).

Figure E-2 Double Acting Cylinder —Stability Problem due to Insufficient Air Pressure (I/P Input Signal/Actuator Position Versus Time)

E-2


Figure E-3 shows the differential actuator pressure (Trace 3) that was obtained by measuring the upper and lower cylinder pressures with a differential pressure transducer. The pressure ranges are about 10 psid (approximately plus 10 psi (68.9 kPa) when the valve is opening and minus 10 psi (68.9 kPa) when it is closing). Notice that when the valve is opening, the lower cylinder is reading 10 psig (68.9 kPa), which means that the upper cylinder must be reading zero pressure. When the valve is closing, the lower cylinder is reading zero, which means that the upper cylinder must be reading 10 psig (68.9 kPa). Therefore, at any time during normal controlling operation, the largest cushion of air is 10 psi (68.9 kPa).

Figure E-3 Double Acting Cylinder —Stability Problem due to Insufficient Air Pressure (Cylinders Differential Pressure [Low Cyl]/Actuator Position Versus Time)

The positioner ’s preload was increased so that the nominal pressure would be between 40 (275.8 (kPa) and 50 psig (344.7 kPa), about 60% of the minimum operating supply pressure.

E-3


Figure E-4 shows that the pressure in the lower cylinder during the opening stroke is between 60 psig (413.7 kPa) and 70 psig (482.6 kPa). Because the differential pressure is still 10 psi (68.9 kPa), the upper cylinder must be between 50 (344.7 kPa) and 60 psig (413.7 kPa). During the closing stroke, the upper cylinder must be about 65 psig (448.2 kPa).

Figure E-4 Double Acting Cylinder —Corrected Stability Problem due to Insufficient Air Pressure (Cylinders Differential Pressure [Low Cyl]/Actuator Position Versus Time)

E-4


Increasing the nominal pressure from an average of 0 to an average of 60 psig (413.7 (kPa) definitely increases the stability of the valve. It is similar to installing a stronger spring or a much stiffer shock absorber. Increasing the nominal actuator pressure away from zero prevents the actuator from having to completely dump one side of the actuator and add pressure to the other every time that the valve direction is changed. Now the actuator only has to add or dump one side of the actuator to move the valve. Figure E-5 shows the improvement in dynamic error, from 9.4% to 1.8%, or more than five times better in controllability.

Figure E-5 Double Acting Cylinder —Stability Problem due to Insufficient Air Pressure (As-Found Versus As-Left Actuator Position Versus Control Signal Input)

Case 3 - Air Leak, Internal to the Positioner Figure E-6 is typical of a normal regulated supply pressure (Trace 2) with a high volume requirement. That is, this valve has a 280-square inch (1290.3-square centimeters) spring return actuator that tends to drag down the air supply whenever it moves. Note that when the valve is opening, the supply pressure drops, but it stays constant as the actuator is demanding a large volume of air to build up the air pressure. On the down stroke there is a slight, but again constant, reduction in air pressure as the positioner controls the volume of air in and out of the actuator to get the valve stroke to follow the control signal.

E-5


Figure E-6 Positioner with No Air Leak —Supply Pressure/Actuator Pressure/Actuator Position Versus Time Plot

In Figure E-7, the regulated supply pressure is low when the actuator pressure is low and reaches the highest values when the actuator pressure is at it highest. The lower regulated supply pressure must be due to either a demand or a leak (or possibly both). Because it is demanding the most when there is no pressure in the actuator, the leak cannot be in the actuator. Because the leak is highest when the demand is lowest in the actuator, the leak must be in the positioner where the delta P is highest at this time and lowest when the actuator is pressurized.

E-6


Figure E-7 Positioner with Internal Air Leak —Supply Pressure/Actuator Pressure/Position Versus Time Plot

When the actuator pressure (X axis) is plotted versus the regulated supply pressure (see Figure E-8), there is a direct correlation. The greatest demand occurs at the lowest actuator pressure.

Figure E-8 Positioner with Internal Air Leak —Supply Pressure Versus Actuator Pressure Plot

E-7


Case 4 - Air Leak Downstream of the Positioner In Figure E-9, the demand is the greatest when the actuator pressure is the highest. The high actuator pressure would occur any place downstream of the positioner from the tubing up to and including the actuator.

Figure E-9 Air Leak Downstream of Positioner —Supply Pressure/Actuator Pressure/Position Versus Time Plot

Although not as dramatic as Figure E-8, there is a definite slope to the lines in Figure E-10, which decrease with increasing actuator pressure.

E-8


Figure E-10 Air Leak Downstream of Positioner —Increasing/Decreasing Supply Pressure Versus Time Plot

E.3

Reference

1. Air-Operated Valve Maintenance Guide, EPRI, Palo Alto, CA: 1996. NP-7412-R1.

E-9


F

MINIMUM INTEGRATED ABSOLUTE ERROR TUNING

In the Minimize Integrated Absolute Error Method, the objective is to minimize the Integrated Absolute Error which total error represented by shaded area shown in Figure F-1 below. The curve represents a process variable as it goes to steady state.

Figure F-1 Area Representation of Integrated Absolute Error

System Conditions: Manual. Procedure: 1. Place system in manual. 2. Set controller output (CO) to somewhere between 10 and 90%. 3. Let process reach steady state. (In manual there will be no oscillations) 4. Measure following: a. Dead time ( ): time it takes for process variable (PV) to change. PV is same as manipulated variable. b. Lag time ( ): time it takes for the process variable (PV) to go 63% of the way to where it eventually ends up. For example, if temperature is the PV, and it finally increased from 100 to 200 , the lag time would be the time it took to go from 100 to 163 . c. Process gain (Kp): the total change in the PV divided by the change in the CO or 5. Compute following depending on situation: a.

= 2 used for aggressive but less robust tuning or F-1

EPRI Licensed Material Minimum Integrated Absolute Error Tuning

b.

= 2( + ) used for more robust tuning

6. Compute settings in accordance with following table: Table F-1 Post-Test Settings to Achieve Minimum Integrated Absolute Error Tuning Controller Type

PI control

Controller Gain Kg (No units) t

Integral Time (Seconds)

Derivative Time (Seconds)

Not applicable

Kp( + ) PID control

t Kp( + /2)

/2

For controllers that use proportional band instead of gain, proportional band is equal to 100 divided by gain (K).

F-2


G

SUMMARY OF KEY POINTS

The following list provides the location of key “pop out” information in this report. Key O&M Cost Point Emphasizes information that will result in reduced purchase, operating, or maintenance costs.

Referenced Section

Page Number

Key Point

4.2.2

4-3

Failure to perform feedback alignment may result in calibration problems, positioner performance issues and/or component damage.

4.2.3.2

4-3

Establishing pressure static bands at the ends of the normal control range of valve operation will result in tighter valve shutoff. Depending on application, this will lessen or eliminate seat damage and/or reduce megawatt losses.

4.4

4-6

The calibration of positioners using a data acquisition system is vital if the condition monitoring and troubleshooting capabilities of these systems is to be fully realized.

5.1.1

5-1

Condition Monitoring (CM) for positioners is primarily the continuation of the calibration process. Calibration must be done carefully and consistently for condition monitoring/preventive maintenance to be effective. Persistent, significant calibration drifts can be used to identify the need for positioner repair or replacement.

5.1.2.1

5-2

All condition-monitoring processes rely on the ability to trend the right data accurately. In the hands of trained and experience personnel, a digital acquisition system can not only gather data accurately, but it does it quickly, efficiently, and consistently.

G-1

EPRI Licensed Material Summary of Key Points

Key Technical Point Targets information that will lead to improved equipment reliability. Referenced Section

Page Number

Key Point

4.2.1

4-2

The bench set pressure range is not the same as the pressure required to stroke the valve in actual service. Making the adjustment while the valve is in service may result in unsatisfactory performance and/or make the valve inoperable.

4.2.3.3

4-4

Zero adjustments are always done in conjunction with span adjustments. In other words, after the span has been adjusted, the zero adjustment must be re-verified to ensure that the zero adjustment has not changed.

A.3.2.1

A-4

Gain settings must be reviewed anytime that the loop is experiencing control problems involving unacceptable cycling or sluggish response. Usually, the process engineer cannot predict the optimal gain setting; it must be done using the results of loop tests.

Key Human Performance Point Denotes information that requires personnel action or consideration in order to prevent injury or damage or ease completion of the task.

Referenced Section 4.2.3.3

G-2

Page Number

4-4

Key Point

Before zero and span adjustments, motion balance positioners require beam or flapper leveling, if the motion beam has been affected by maintenance or if the unit is changed to split-range operation. Failure to do so will result in inability to properly calibrate the unit.

Valve Positioner Principles and Maintenance Guide

Recommend Documents