Tài liệu process software and digital networks

INSTRUMENT ENGINEERS’ HANDBOOK Fourth Edition Process Software and Digital Networks BÉLA G. LI P TÁ K , E d itor-i n- Ch ief H A LI T Er En, Volu me E d itor INSTRUMENT ENGINEERS’ HANDBOOK Fourth Edition Process Software and Digital Networks VOLUME III INSTRUMENT ENGINEERS’ HANDBOOK Fourth Edition Process Software and Digital Networks VOLUME III Béla G. liPták, Editor-in-Chief Halit ErEN, Volume Editor MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB® software or related products does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular use of the MATLAB® software. CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2012 by Béla Lipták CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Version Date: 20110713 International Standard Book Number-13: 978-1-4398-6343-5 (eBook - PDF) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com Dedicated to our colleagues and to our profession of automation engineering. It is hoped that by applying the knowledge found on these pages, we will make our industries more efficient, safer, and cleaner, and thereby will not only contribute to a happier future for all mankind, but will also advance the recognition and respectability of our profession. Béla Lipták (http://belaliptakpe.com/) CONTENTS Preface   xiii Acknowledgments   xxiii Contributors   xxv Definitions   xxix Abbreviations   xlix Organizations   lix 1 2 Process Control and Automation   1 1 Distributed Control Systems and Process Plants   5 Mark Nixon 2 Networks in Process Automation: Hardware Structures and Integration of Process Variables into Networks   40 Peter G. Berrie and Klaus-Peter Lindner 3 Instrumentation in Processes and Automation   72 Donald Chmielewski and Miguel J. Bagajewicz 4 Programmable Logic Controllers   90 Sujata Agashe and Sudhir D. Agashe 5 SCADA—Supervisory Control and Data Acquisition System and an Example   101 Hiten A. Dalal 6 Intelligent Instruments and Sensors: Architecture, Software, Networks, Protocols, and Standards   130 Deniz Gurkan and Halit Eren 7 Calibrations in Process Control   141 Halit Eren 8 Standards in Process Control and Automation   150 Halit Eren 9 Automation and Robotics in Processes   158 Arif Sirinterlikci, Arzu Karaman, and Oksan Imamog̀lu Process-Control Methods   169 10 Batch-Process Automation   172 Asish Ghosh ix x   Contents 3 4 11 Plant-Wide Controller Performance Monitoring   200 Chris McNabb 12 Plant Optimization   228 Michael Ruel 13 Neural Networks in Process and Automation   254 R. Russell Rhinehart 14 Fuzzy Logic Control in Processes and Automation   265 R. Russell Rhinehart 15 Internet in Automation and Process Control Systems   275 Babu Joseph and Deepak Srinivasagupta 16 Telemetry Systems: Phone, Radio, Cellular, and Satellite   284 Curt W. Wendt Digital Techniques and Data Handling   301 17 Digital Technology Fundamentals, Microcontrollers, Microcomputers, and FPGAs in Processes   303 Cesar Ortega-Sanchez 18 Signal Processing in Process Control and Automation   313 Emre Deniz Eren and Halit Eren 19 Data Acquisition Fundamentals   330 David Potter and Halit Eren 20 Analog and Digital Signal Transmission in Processes: Protocols and Standards   342 Ian H. Gibson 21 Data Acquisition: Buses, Networks, Software, and Data Handling   351 David Potter 22 Data Reconciliation and Software Methods for Bias Detection   364 Miguel J. Bagajewicz and SR Derrick K. Rollins Software, Programming, and Simulations   383 23 Software Fundamentals   385 Halit Eren 24 Virtual Plants: Process Simulation and Emulation   391 Hironori Hibino 25 Virtual Reality Tools for Testing Control Room Concepts   406 Asgeir Droivoldsmo and Michael Louka 26 Model-Free Adaptive Control Software   415 George S. Cheng and Steven L. Mulkey 27 Operation Optimization with Sequential Empirical Optimization and Software Implementation   438 Carlos W. Moreno 28 Data Historian   454 Irena Yee and Halit Eren Contents   xi 5 6 7 Networks, Security, and Protection   461 29 Computer Networks: LANs, MANs, WANs, and Wireless   465 Hura Gurdeep 30 Internet Fundamentals and Cyber Security Management   484 Hura Gurdeep 31 Network Security, Threats, Authentication, Authorization, and Securing Devices   506 Wenbin Luo 32 VPN, CCN, and IT Support   517 Burhan Basaran and Cengiz Burnaz 33 Fiber-Optic Network Components   525 Jin Wei Tioh, Mani Mina, Robert J. Weber, and Arun K. Somani 34 Fiber-Optic Communications and Networks   546 Martin Maier 35 Network Access Protection   565 Dinesh C. Verma 36 Comments on Cyber Security in Industrial Control Systems and Automation   571 Jacob Brodsky and Joseph Weiss Fieldbus Networks   575 37 Fieldbuses   578 Lucia Seno and Stefano Vitturi 38 HART Networks   587 David S. Nyce 39 Foundation Fieldbus: Features and Software Support   598 Salvatore Cavalieri 40 PROFIBUS Networks   616 Peter G. Berrie and Jochen Müller 41 Industrial Ethernet and TCP/IP-Based Systems   636 Gianluca Cena, Stefano Scanzio, Stefano Vitturi, and Claudio Zunino 42 Niche Fieldbus Networks   654 Lucia Seno and Stefano Vitturi Process Management, Maintenance, Safety, and Reliability   671 43 Network Security Awareness, Management, and Risk Analysis   674 Helen Armstrong 44 Manufacturing Execution Systems   689 Zameer Patel 45 Auditing and Upgrading Plants, Control Rooms, and Networks   698 Bridget A. Fitzpatrick 46 Hazardous Areas: Classifications, Equipment, Purging, and Management   706 David S. Nyce xii   Contents 8 47 Safety in Processes: Rules, Standards, Certification, Culture, and Management   716 Asish Ghosh 48 Reliability, Redundancy, and Voting Systems   736 William Goble 49 Computerized Maintenance and Maintenance Management   746 Pete Peter W. Ralph Process Control and Automation Applications   757 50 Manufacturing, Plant, and Production Management: Applied in Automobile Industry   762 Paulina Golinska and Marek Fertsch 51 Control Systems and Automation in Steel Thixoforming Production   781 Ahmed Rassili and Dirk Fischer 52 Processes and Automation in Dairy Industry   797 Sudhir D. Agashe and Sujata Agashe 53 Software for Pharmaceutical Automation   803 George C. Buckbee 54 Process Automation in the Automotive Industry   809 Vivek Hajarnavis 55 Mine-Wide SCADA System   821 Erik Bartsch 56 Application of Artificial Intelligence and Fuzzy Logic in Mineral Processing: Hydrocyclones   840 Kok Wai Wong and Halit Eren 57 Computer Control in Mining and Extractive Metallurgy   847 Greg Baiden and Sirkka-Liisa Jämsä-Jounela 58 Telemetry Control and Management of Water Treatment Plants   858 Curt W. Wendt 59 Design and Implementation of a Safe and Reliable Instrumentation and Control System in Oil and Gas Industry   870 Harvindar S. Gambhir 60 Power Network Security   880 Nian Liu 61 Nuclear Plant Instrumentation and Control System Performance Monitoring   903 Hashem M. Hashemian 62 Alternative Energy I: Control Software Needs of Renewable Energy Processes   919 Béla Lipták 63 Alternative Energy II—SCADA System for Thermal Power Plant   930 Mihai Iacob, Gheorghe-Daniel Andreescu, and Nicolae Muntean 64 Alternative Energy III—Wind Energy   940 Gordon Smith Appendix   951 A.1 A.2 International System of Units   953 Engineering Conversion Factors   963 Contents   xiii A.3 A.4 A.5 A.6 A.7 A.8 Chemical Resistance of Materials   985 Composition of Metallic and Other Materials   993 Steam and Water Tables   997 Friction Loss in Pipes   1005 Tank Volumes   1013 Partial List of Suppliers   1017 PREFACE I thank Dr. Halit Eren for his fine work in editing the fourth edition of the third volume of the Instrument Engineers’ Handbook (IEH). I also thank his daughter, Pelin Eren and colleague Cesar Ortega-Sanchez, who assisted him in this major task, as well as the coauthors of this volume for their valuable contributions. This edition covers the software and the digital networks that are used to monitor, control, and automate the various industrial and nonindustrial processes. As to the other volumes of the three-volume IEH set, the fourth edition of Volume 1 covered measurement and analysis, while Volume 2 dealt with control and optimization. Now, when the fourth updating edition of all three volumes is completed and is available in both printed and electronic forms, we will start working on the fifth edition of IEH. Each updated edition of the three IEH volumes requires about a decade to prepare, so the fifth edition is expected to be completed only by 2020. We will start that effort by first updating Volume 1, which should be completed in about 3 years (in 2014). If you or one of your colleagues is knowledgeable about a sensor or an analyzer and would like to contribute to this fifth edition of Volume 1 by either updating an existing chapter or by preparing a new one that would describe an instrument that did not exist 10 years ago, please contact me at [email protected], http://belaliptakpe.com//. After some comments by the volume editor, Dr. Halit Eren, I will tell you about the birth of IEH and will share with you some observations concerning the present state of our profession and its future goals. NOTES BY THE EDITOR OF THIS VOLUME This volume provides an in-depth, state-of-the-art review of control software packages that are used in plant optimization, control, maintenance, and safety. Automation and control systems are evolving rapidly with more and more applications of intelligent instruments, enhanced networks, use of the Internet, virtual private networks, and integration of control systems with the main networks used by management, all of which operate in a global environment. This holistic approach is convenient and efficient, but it also introduces cyber and local network security problems that need to be addressed by effective technical solutions and proper management policies and practices. Conventional control networks, such as fieldbuses, do not operate as stand-alone and isolated entities. They are—or if not already can be made—a part of a larger network through the use of virtual networks. If given permission or access, they can be controlled and operated by anyone anywhere in the world. This volume highlights the technological, management, and cultural aspects of network security. Nowadays, almost every device is software driven. Some single chips comply with international standards, contain memory and intelligence, and are often embedded in the manufactured products. This introduces network security problems. This volume is organized in eight parts, each containing a number of related chapters. Part I discusses SCADA and PLC systems that are still at the heart of industrial control. Process automation–related national and international standards are also discussed together with topics like calibration. Part II covers software used in the automation and optimization of batch and continuous processes. It also covers artificial intelligence, the Internet, and telemetric operations. Part III focuses on digital systems and describes the fundamentals of signal processing, data acquisition, data handling, signal transmission, data reconciliation, and bias detection techniques used in digital hardware. Part IV explains the fundamentals of software used in process simulations, control room design, and virtual reality. In addition, modelfree adaptive control and optimization are also discussed. Part V discusses networks and network security as part of the World Wide Web. It also covers virtual networks and the fundamentals of Internet access protection and security from a technological, organizational, and cyber attack point of view. Part VI focuses on fieldbuses and on wireless technology used in the form of LAN, WAN, or satellite systems. In addition to HART, Foundation Fieldbus, Profibus, and the industrial Ethernet are also discussed. Part VII covers the xiii xiv   Preface use of networks as used by managers to make economic and financial decisions as well as production planning, maintenance, etc. It also covers network security from a managerial point of view and discusses MES, CMMS, reliability, redundancy, and voting systems. Part VIII gives 15 examples of software applications from a number of industries, including automobile, mining, renewable energy, steel, dairy, pharmaceutical, mineral processing, oil, gas, electric power, utility, and nuclear industries. HISTORY OF IEH In 1956, as a 20-year-old university student in Hungary, I was one of the freedom fighters who tried to end Soviet occupation. Our efforts were similar to the fight to liberate America in the war between 1775 and 1781. The difference was that George Washington won, while Imre Nagy was hanged and 3% of the Hungarians—mostly well-educated young people—were either killed or became refugees. I was one of them. My first job in America was to work for Sam Russel, who, during World War II, led the effort to develop a synthetic rubber industry after the Japanese cut off our natural rubber supply from Indonesia. By the end of the war, American trucks were rolling on synthetic rubber tires and Sam was already thinking about starting an engineering design firm (C&R) with the goal of developing a plastics industry. In the late 1950s, he hired me to work for him as an instrument engineer in a department that, at that time, had only three engineers. I tried to learn all that was available in this new profession of automation. I read all there was to read, attended every conference, and in a couple of years knew a lot about stuff like frequency domain analysis. As to practical knowledge of measurement and control, I had to learn on my own or from people like Greg Shinsky and Donald Eckman. One day, Sam Russel asked me if I would take over our fast-growing instrument department. I said that at the age of 25 and with my thick Hungarian accent I would not be able to lead experienced engineers twice my age. I volunteered instead to accept his offer only if I could hire the best graduates from the best universities and if Sam allowed me to teach them. This I proposed to do by taking every Friday and instead of working on the jobs, I would teach them the profession and bring them “up to speed” that way. He said OK. Later in the early 1960s, Nick Groonewelt—a Dutch American who still parted his hair in the middle and was running a publishing company—visited my office and noted the tall piles of paper on my desk. “What are these?” he asked. “My Friday notes, which I teach from,” I replied. “Why don’t you use a handbook of your profession?” he asked. “Because there is none that is any good!” was my reply. “Well, let’s publish these Friday notes then!” And this is how the IEH was born. In preparing the first edition that was published in 1969, I asked the Nobel laurate Edward Teller to write a preface for it. “Why should I? I know nothing about your profession!” he replied. “Because everybody knows that you are Hungarian and I want the readers to know that the IEH is a Hungarian contribution to American science!” I said. Ede bácsi (uncle Edward) went to the library, spent a whole weekend there, and wrote a preface that turned out to be much better than this one. AUTOMATION AND CONTROL ENGINEERING Ours is a very young profession. Compared to other fields of engineering, we were way behind. When the first edition of the IEH was published, Marks’ Mechanical Engineers’ Handbook was in its fifth edition and Perry’s Chemical Engineers’ Handbook was in its sixth edition! It is partially for this reason that people are more aware of what a mechanical engineer or a chemical engineer does, whereas they have no idea what we do. When I say that my field is process control or instrumentation, all I get is a blank stare. It is time for us to change that. The first step would be to use a name for our profession that people understand, such as “automation.” It is also time for our profession to develop a distinct identity. When I was teaching at Yale University, my course was offered under the chemical engineering department. This was not because Yale had anything against our profession, but because they did not even know that it exists. Even this handbook provides an example as to the confusion about our identity, because CRC/Taylor & Francis Group will be publishing this handbook in their electrical engineering series. Once again, this is not due to any bias against our profession, but rather it reflects our failure to develop a distinct identity! “Automation” is a term that the wider public understands, and I am glad that ISA changed its name to International Society of Automation (ISA). DEVELOPMENTS OVER THE LAST DECADES These days, computers are our main tools of control, and it is the software that makes computers useful. The chapters of this volume describe how computers are being used in optimizing our processes, providing self-diagnostics, and displaying status information in operator-friendly formats. Today, we can fully automate the safety of our processes, we can eliminate the need for manual actions by panicked operators following the instructions of greedy management, and thereby we can prevent oil spills, nuclear accidents, and the use of our air and waters as garbage dumps. During the last decade, the artificial separation between the plant’s control, logic, and business needs has disappeared and have gradually been integrated. The operating software of the future will support all needs of a process; not only will digital networks become wireless, digital bus protocols will also become the same around the world, eliminating the need for interfacing and the associated risk of mix-up, not to mention the creation of captive markets. This situation is being Preface   xv improved by the Open Network Foundation (ONF), which was established by the big technology companies who agreed to blur the distinctions between networks and to establish a common standard for all of them. In the following text, I will review some of the developments that have already occurred during the last decade: IS THE AGE OF THE PID OVER? Designating a valve on a flow sheet as a temperature control valve (TCV) will not suspend the laws of nature and will not, for example, prevent the valve from affecting the process pressure. Similarly, the number of available control valves manipulating a process does not necessarily coincide with the number of variables we need to control. Our plants do not produce pressures and temperatures; they manufacture products. Therefore, our goal must be to optimize the safety and efficiency of our plants and let the pressures and temperatures serve those goals. To realize this, multivariable herding or envelope control should gradually replace the uncoordinated single-loop controllers, while decoupling the interactions and considering the relative gains among the controlled variables. The result of this trend will be the unit operation controller (UOC), which will control such subsystems of the plant like a boiler, distillation column, or a compressor as a single system and treat their flows, temperatures, etc. only as constraints of the operating envelope. Most UOCs will use model-based control utilizing both steady-state and dynamic models to predict and correct unsafe process responses before they occur. In this regard, artificial neural networks (ANN), artificial intelligence, statistical process control, fuzzy logic, empirical optimization, and other future strategies will play an important role. Advances have also been made in the use of set point for different processes. This is often replaced by a set point gap, so that as long as the controlled variable is within that gap, the output is unaltered. This tends to stabilize sensitive loops such as flow, or even faster loops like missile guidance. Another aspect in which the set point is treated differently today is its effect on the controller output. In many algorithms, a change in set point does not change the proportional or derivative contributions to the output, because the P and D modes act only on the measurement. In other algorithms, while the set point change does affect the integral contribution to the output, it is “feed-forwarded” directly to the output to minimize reset windup. Reset windup is also minimized by external feedback taken from the slave measurement in case of cascade loops, from the valve signal in case of selective loops, and from the inverse model in case of feed-forward loops. DYNAMICS, DEAD TIMES, AND MODEL-BASED CONTROL For traditional quarter amplitude damping in PID control, we kept the gain product of the loop at about 0.5. This meant that if the process gain doubled, the controller gain had to be cut in half (Figure P.1 on page xiv). This understanding was important to the control of nonlinear processes like heat transfer, chemical reaction, pH, etc. Similarly, in the past we understood that in PID control the integral and derivative settings are a function of the dead time. Therefore, we reduced the loop dead time to a minimum and kept it constant. This was not always possible, because the dead time varied with load (e.g., because transportation lag varied with flow as it displaced fixed volumes); so when the dead time was large, we replaced the regular PID algorithm with sample-and-hold or predictor algorithms. Today, the trend is to use some type of control models when optimizing unit operations. A variety of software packages are available using model-based or model-free control. They can be model predictive control (MPC) or internal model control (IMC) packages, if they deal with well-understood processes such as heat transfer or distillation processes. If the process is not well understood and is very complex, model-free expert systems can be used. These systems operate and learn like tennis players who do not understand Newton’s laws of motion or the aerodynamic principles that determine the behavior of a tennis ball, yet learn by past experience. All the neural network–based “trainable” software packages mimic this method of learning. Neural networks, fuzzy logic, and statistical process control are all such model-free methods of control. The major difference between fuzzy logic and neural networks is that the latter can only be trained by data, but not with reasoning. In this respect, fuzzy logic is superior, because it can be modified both in terms of the gain (importance) of its inputs and in terms of the interactions of its inputs. The main limitation of all model-free expert systems is their long learning period (which can be compared to the maturing of a child) and the fact that their knowledge is based solely on past events. Consequently, they are not prepared to handle drastic changes in the controlled process. Therefore, if the nature of the process changes from what it used to be, they require “re-training.” ARTIFICIAL NEURAL NETWORKS AND HERDING CONTROL ANNs can either be applied under human supervision or can be integrated with expert and/or fuzzy logic systems. Figure P.2 on page xiv shows a three-layer ANN, which serves to predict the boiling point of a distillate and the Reid vapor pressure of the bottoms product of a column. Such predictive ANN models can be valuable, because they are not limited by either sensor error or by the dead time of analyzers. The “personality” of the process is stored in the ANN by the way the processing elements (nodes) are connected and by the importance assigned to each node (weight). The ANN is “trained” by example, and therefore it contains the adaptive mechanism for learning from example and the ability to adjust its parameters based on the knowledge that is gained xvi   Preface Gv Load Valve gain (Gv) m1 + Controller gain (Gc) Gc + For stable control: Gc× Gv×Gp×Gs =0.5 Gp Process gain (Gp) Load Load (U) m Load e + – c b Sensor gain (Gs) Set point (r) Gs Load FIG. P.1 After an upset that produces cycling, in order for a traditional PID loop to cut the amplitude of each succeeding cycle in half, the gain product of the loop had to be about 0.5. RVP in bootoms Distillate 95% bp Output layer (2 nodes) Hidden layer (4 nodes) Input layer (9 nodes) Bias Feed flow Bottoms flow Distillate flow Feed temp Steam flow Bottoms temp Top temp Pressure Reflux temp FIG. P.2 A three-layer ANN can be used to predict the quality of overhead and bottoms products in a distillation column. Preface   xvii ANN controller w + – u y Process ef + ANN model – em Filter FIG. P.3 ANN control with self-training of the ANN model. through adaptation. The hidden layers help the network to generalize and even to memorize. During the “training” of these networks, the weights are adjusted until the output of the ANN matches that of the real process. The ANN can “learn” both the input–output relationships between variables and the inverse of these relationships. Hence, ANN can be useful in building IMC systems using ANN-constructed plant models and their inverses (Figure P.3). When a large number of variables are to be kept within some limits, but only one is to be kept on set point, the use of model-free herding control can be considered. This approach to control can be compared to what the shepherd’s dog does when it goes after one animal at a time and changes its direction or speed, thereby keeping the herd together or moving it in the right direction. I applied this herding algorithm in designing the controls of the headquarters building of IBM (590 Madison Ave) in New York City. The goal in this case was to herd the warm air to the perimeter (to offices with windows) from the offices that even in the winter were heat generators (interior offices). Thereby, the building became self-heating. This was done by changing the destination of the return air from the interior offices by first directing it to the supply air header of the perimeter offices. In general, herding control is effective if there are thousands of manipulated variables and they all serve some common goal, in this case the reduction of the building’s energy consumption. THE AUTOMATION ENGINEER (AE) AND COMMON SENSE An AE’s best tool is common sense and his or her best teacher is Murphy who says that anything that can go wrong, will. In this regard, I will list here some advice and observations: • Before an AE can control a process, he or she must fully understand it. • Being progressive is good, but being a guinea pig is not. Bad design implemented by advanced hardware is still bad. • Management should be told what they need to hear and not what they like to hear. • Increased safety is gained through backup. In case of measurement, reliability is increased by the use of multiple sensors, which are configured through median selectors or voting systems. • If an instrument is worth installing, it should also be worth calibrating and maintaining. No device can outperform the reference against which it was calibrated. • Sensors with errors expressed as percent of the actual reading are preferred over those with percent of fullscale errors. • It is easier to drive within the limits of a lane than to follow a single line. Similarly, control is easier and more stable if the single set point is replaced by a control gap. • Constancy is the enemy of efficiency. Optimization requires efficient adaptation to changing conditions. • Trust your common sense, not the sales literature. Independent performance evaluation (SIREP-WIB) should be done before installation and “business lunches” should wait until after start-up, not before the issue of the purchase order. • Annunciators do not correct emergencies; they just throw the problems that designers do not know how to fix into the laps of the operators. The smaller the annunciator, the better the design. ROLE OF THE AE PROFESSION IN THE THIRD INDUSTRIAL REVOLUTION Here, I will describe what I hope our profession will accomplish in the next decade. I would not be surprised if by the end of the twenty-first century, we would be using self-teaching computers. If this comes to pass, some might argue that this would be a step forward, because machines do not forget; do not get tired, angry, or sleepy; and do not neglect their job to watch a baseball game or to argue with their wife on the phone. This might be so, yet I would still prefer to land in a human-piloted airplane. xviii   Preface In addition to new control tools, we will also have new processes to control. As we enter the third Industrial Revolution and convert to an inexhaustible and renewable energy economy, the AE will meet new challenges (see Chapter 62). One such challenge will be to control regular and reversible fuel cells. The fuel cell is like a battery, except that it does not need recharging, because its energy is the chemical energy of hydrogen, and hydrogen can come from an inexhaustible source, namely, the splitting of water by solar energy. The challenge involves not only the control of the electrolytic process in the fuel cell, but also the storage and transportation of solar energy in the form of hydrogen. In the past, it took decades to reach an agreement on the standard analog signal ranges of 3–15 PSIG (0.2–1.0 bar) and later on the 4–20 mA DC. Yet, when these signal ranges were finally agreed upon, the benefit was universal. Similarly today, the time is ripe for a single standard for digital communication protocols to link all the digital “black boxes,” software packages, and networks and to eliminate the need for all interfacing. Protocol is the language spoken by our digital systems. Unfortunately, there is no standard currently in place that allows all devices to communicate in a common language, but work on the creation (and universal acceptance) of a fieldbus standard has started. There is nothing wrong with, say, the Canadians having two official languages, but there is something wrong if a pilot does not speak the language of the control tower or if two black boxes in a refinery do not speak the same language. Yet the commercial goal of manufacturers was to create captive markets and this resulted in different control-oriented protocols, as follows: AS-Interface HART PROFIBUS Found. Fieldbus ControlNet MODBUS Ethernet TCP/IP www.as-interface.com www.hartcomm.org www.profibus.org www.fieldbus.org www.controlnet.com www.modbus.org www.industrialethernet.com During the last decade, HART has become the standard for interfacing with analog systems while Ethernet was handling most office solutions. SCADA served to combine field and control data to provide the operator with an overall view of the plant. While there was no common DCS fieldbus protocol, all protocols used Ethernet at the physical and TCP/IP at the Internet layer. MODBUS TCP was used to interface the different DCS protocols. The layers of the communication pyramid were defined in several ways. OSI defined it in seven layers, with layer 1 being the physical and layer 7 the application layer (layer 8 being used for the “unintegrated,” external components). The IEC61512 standard also lists seven levels, but they base their levels on physical size as follows: (1) control module, (2) equipment, (3) unit, (4) process cell, (5) area, (6) site, and (7) enterprise. As I noted earlier, in the everyday language of process control, the automation pyramid consists of four layers: layer 1 is the level of the field devices, the sensors, and actuators; layer 2 is control; layer 3 is plant operations; and layer 4 is the level of business administration. Naturally, it is hoped that in the next decade, uniform international standards will replace our digital Babel, so that we can once again concentrate on controlling and optimizing our processes. SOFTWARE OUTSOURCING AND CONNECTIVITY PROBLEMS The job of AEs can also be complicated by the commercial practices of some suppliers. For example, some suppliers sell their systems without including all the software that is needed to operate them. This is wrong. To treat software as an extra and not to include the unique control algorithms, faceplates, and graphic displays in the basic package can lead to serious problems. Another reoccurring problem is the quality of the instruction booklets. These instructions are often written in the language of programmers (“computerese”) that is not spoken or understood by plant operators. In some proposals, one might also read that the stated cost is for “hardware with software license.” This would suggest that the operating software for the DCS or other systems is included. In most cases, however, it is not and only its license is. Similarly, when the proposal states that an analyzer, an optimization, or a simulation package requires “layering,” or can be implemented in the “eighth layer,” one might think that the bid contains eight layers and that they are fully integrated. Yet, what this actually means is that the cost of integrating these packages into the overall control system is an extra. So, on the one hand, while the plant-wide digital networks and the advanced control strategies they support provide great opportunities for optimization, on the other hand, AEs must be very careful in making sure that all the pieces of the digital puzzle are provided and will conveniently fit together. The control systems of most newly built plants consist of four levels of automation. The first level supports the local instruments (sensors, valves, motors, safety devices), including the intelligent and self-diagnosing ones. This first level can be connected by a number of data highways or network buses to the second level, which is the level of control. The third level serves plant operations and the fourth the enterprise-wide business. In addition, wireless hand tools are used by the roving operators, and external PCs are available to engineers who introduce or modify control or simulation models and algorithms. NEED FOR MORE MEANINGFUL SPECIFICATIONS The professional organizations of AEs, such as ISA, should issue uniform performance testing and calibration criteria to Preface   xix be used throughout the industry. This is needed, because in the sales literature today, the meanings of the terms stating the instrument’s inaccuracy, repeatability, or rangeability are rarely based on individually calibrating and testing each device, and the statements seldom describe the calibration procedures and the reference devices used. Even such terms as “inaccuracy” or “error” are often given without stating if these statements are based on fullscale or on actual readings. Also the validity of the published performance data is seldom checked by reliable third parties. If third parties do check them, their findings are seldom published. Therefore, it is difficult for the user (the AEs) to determine or compare the expected performance of various manufacturers’ products. Also, it would be desirable that the rangeability of all sensors be stated and guaranteed. Rangeability is the range—the reading at full scale divided by the reading at the minimum reading of a detector (such as 10:1 or 20:1)—across which the inaccuracy of the sensor or other instruments is guaranteed. The sales literature of all manufacturers should always publish their accuracy (which is actually inaccuracy) statements on this basis and should always state the rangeability of their products. It would also be desirable if the sales literature stated the basis of the “accuracy statement” in terms of the testing that was used to arrive at them. This is important, because it makes quite a difference if the performance of all sensors were individually tested or only “sampling testing” was performed. Also, besides accuracy, rangeability, and calibration information, the performance statement should include linearity, hysteresis, drift, and the effects of variations in ambient temperature, supply voltage, humidity, RFI, vibration, or the maximum over-range that the detector can be exposed to. CONTROL VALVE PERFORMANCE In the next decade, much improvement is expected in the area of final control elements, including smart control valves. This is a welcome development, because the performance of the control valve is affected not only by trim wear, stem sticking, or packing friction but also by the misoperation of its actuator, limit switch, valve position detector, or positioner. The tuning and stability of the control loop also depends on the gain characteristics of the valve and the accuracy of the gain (linear or nonlinear) that the manufacturer guarantees. If the control valve is nonlinear (its gain varies with the valve’s opening), the loop will become unstable if the load changes (is different from the load that existed when the loop was tuned). For this reason, the loop gain must be compensated for by the gain characteristics of the valve and such compensation is possible only if the valve characteristics are accurately known. For these reasons, it is desirable that the manufacturers accurately test and state in their sales literature the characteristics of their valves. The other performance capabilities, such as the rangeability of control valves, should also be stated. For example, a valve should be called linear only if its gain (Gv) is constant across its stated range. The valve manufacturers should also publish the stroking range (minimum and maximum percentages of valve openings) within which the claimed valve gain is guaranteed (for a linear valve it is Fmax/100%). Similarly, valve rangeability should be defined as the ratio of the minimum and maximum valve Cvs at which the valve characteristic is what it is specified to be. Finally, and most importantly, the sales literature should state if the performance data are based on individual testing and if the testing was performed by measuring the actual flow through the valve and the stem position throughout its range. SMART CONTROL VALVES A traditional valve positioner serves only the purpose of keeping a valve at its intended opening. Intelligent (digital) control valves, on the other hand, provide the ability to collect and analyze the valve performance data, including valve characteristics and performance trends over time. Smart positioners also provide (wired or wireless) twoway communication to enable diagnostics of the entire valve assembly. Section 6.12 in Volume 2 of this three volume handbook set provides detailed information on the capabilities of intelligent control valves and their suppliers. The capabilities and use of smart valves are likely to increase with time and to include the ability to measure variables such as their own: • • • • • • Upstream pressure Downstream pressure Temperature Valve opening position Actuator air pressure, etc. Flow Smart valves will also be able to eliminate converter errors (digital-to-analog or analog-to-digital conversions) and will be able to update their data at least 10 times per second. They could also be equipped with filters to remove fluctuations caused by turbulence or other effects. Eventually, smart valves should be able to measure the flow passing through them if their stem positions and trim characteristics as well as the properties of the flowing fluid (up- and downstream pressure, density, viscosity, temperature, turbulence, etc.) are accurately known. In such a case, the smart valve can calculate and display its own flow by solving various equations used in valve sizing. For details, see Section 6.15 in Volume 2 of this three volume handbook set. The smart valves of the coming decades will hopefully be able to measure the flow over a range which exceeds the rangeability of most flowmeters, because they in effect are differential pressure flowmeters operating on variable