Control of Batch Processes E-Book

Cecil L. Smith

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 562

Veröffentlichungsjahr: 2014

Beschreibung

Gives a real world explanation of how to analyze and troubleshoot a process control system in a batch process plant * Explains how to analyze the requirements for controlling a batch process, develop the control logic to meet these requirements, and troubleshoot the process controls in batch processes * Presents three categories of batch processes (cyclical batch, multigrade facilities, and flexible batch) and examines the differences in the control requirements in each * Examines various concepts of a product recipe and what its nature must be in a flexible batch facility * Approaches the subject from the process perspective, with emphasis on the advantages of using structured logic in the automation of all but the simplest batch processes. * Discusses the flow of information starting at the plant floor and continuing through various levels of the control logic up to the corporate IT level

Leseprobe

Copyright © 2014 by John Wiley & Sons, Inc. All rights reserved

Published by John Wiley & Sons, Inc., Hoboken, New Jersey

Published simultaneously in Canada

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, [phone], fax [phone], or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, [phone], fax [phone], or online at http://www.wiley.com/go/permissions.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at [phone], outside the United States at [phone] or fax [phone].

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Smith, Cecil L.

    Control of batch processes / Cecil L. Smith.

            pages cm

    Includes index.

    ISBN 978-0-470-38199-1 (hardback)

    1.  Chemical process control.    2.  Mass production.    I.  Title.

    TP155.75.S584 2014

    660'.2815–dc23

                                                                      2014007295

Preface

This book is based on the premise that the approach to automate a batch process involves three essential components:

Automation equipment.

  Many alternatives are available, but all are digital. The natural focus is on recipes and sequence control, but neglecting continuous control is a big mistake.

Instrumentation.

  Resistance temperature detectors (RTDs) for temperatures and coriolis meters for true mass flows address two essential requirements for most batch processes: close temperature control and accurate material charges.

Process modifications.

  With heat transfer rates varying by 50 : 1 during a batch, changes such as upgrading a once-through jacket to a recirculating jacket are not just desirable, they are essential to attain good control performance throughout the batch.

Especially when automating a batch process, this book takes the position that the following statements are good advice:

Process modifications are often not just desirable, they are essential.

  In an early project where process modifications were totally off the table, the project manager correctly observed “we are spending a million dollars to automate a hunk of junk” with a few expletives added for emphasis. Actually, I developed a great respect for this guy. He could express things in ways that the plant manager could understand, which helped save this project from the axe.

Priority must be to achieve a high degree of automatic control.

  For automating a batch process, much emphasis is placed on recipes, sequence control, and the like. Although definitely important, equally important is close control of key process variables such as reactor temperatures. The sophistication of the continuous control configurations to achieve this in a batch facility rivals their counterparts in continuous plants, especially when coupled with objectives such as minimizing the consumption of a utility such as refrigerated glycol.

Discrete measurements and discrete logic receive little attention in most engineering programs. Consequently, most practicing engineers acquire this knowledge on-the-job in an as-needed manner. One tends to learn what to do, but not always the reasons for doing it that way. Starting with the basics, this book devotes an entire chapter to discrete measurements and discrete logic, the intended audience being process engineers who usually define what the process requires but do not actually implement the logic.

The last part of this book is devoted to recipes, sequence logic, and the like. Especially for highly flexible manufacturing facilities for products such as specialty chemicals, analyzing the requirements of the process and developing a logical organization or structure for meeting these requirements is absolutely essential. But in the end, this is applying well-known concepts (at least within the computer industry) such as structured programming to the logic required to automate a batch manufacturing facility.

This book goes beyond explaining what we should do, but why it is so important to do it in a certain way. Batch automation involves many compromises. Common examples are the compromises made on load cell installations within a process facility. While made for good reasons (especially those pertaining personnel safety), the accuracy of the weight measurement suffers. With regard to batch logic, the main objective is to separate equipment-specific logic from product-specific logic. But process issues often become obstacles to achieving this ideal, necessitating compromises. This book explains the consequences of such compromises.

The basic building block for batch automation seems to be the batch phase or something comparable by another name. Explain this concept to someone with a good understanding of software development. Their reaction: “Obviously, but what's the big deal?” Although normally activated in a manner other than static linking, a batch phase is really the counterpart to a function or subroutine, and these have been around for a long time.

Many requirements of batch automation derive from concepts that have been well known in computer circles for years. Automating a batch production facility involves managing multiple simultaneously active sequences, which entails coordination of two (or more) sequences at certain times, allocating shared resources, avoiding deadlocks, and the like. Multitasking operating systems faced these issues years before they arose in the context of batch automation.

Presentations on batch automation often place great emphasis on various standards. The process industries are favorably disposed to standards, and much effort has been expended developing standards for batch automation. I have not been involved in any way with the standards efforts, but I personally know many of the people who have played key roles. But one side effect of being in computing is that one becomes cynical about certain endeavors that are cherished in certain quarters. For me, standards fall into this category.

Like many people my age, I got into computing the hard way, working with punched cards and learning FORTRAN from IBM's programming manual in 1964. Sort of a rite of initiation, I suppose. But once hooked, computing is so dynamic and fascinating that giving it up is unthinkable. Even after 50 years, I have no regrets.

In the computer industry, change is a way of life. Computer people not only live with change; they expect it! A few years ago, I was working with a company that was using equations for the thermodynamic properties of steam that were 30 years old. The reaction from the computer geeks was enlightening. They wanted to know how anyone could be using the same thing for 30 years. I observed that the thermodynamic properties of steam today are the same as 30 years ago. Being technical people, they understood this and were somewhat embarrassed; they were so accustomed to everything changing that they had given it no thought. But in some respects, they were correct. The thermodynamic properties of steam do not change, but the equations that approximate these properties hopefully improve with time.

From the early days, standards existed in the computer industry, but to what effect? One of the more interesting standards applied to magnetic tapes. The standard essentially stated that the magnetic tape had to be 0.5 in. in width (I have no idea if it was ever updated to metric units). The standard assures that the tape will physically fit into the drive, but presumably, one mounts the tape with the desire to transfer information to or from it. The standard is silent on the latter aspect. Writing tapes in “IBM-compatible format” enabled them to be transferred from one manufacturer's computer system to another.

The computer industry is suspicious of anything with the potential of standing in the way of change. The desire is for the marketplace to determine winners and losers. Consequently, standards do not lead; they follow, often doing little more than crowning the winner. Attempts to drive technology through standards prove futile. Over the years, the computer industry has become very adept at developing standards that fall short of what the name implies.

This book is the fifth I have written on various aspects of process automation. Maybe the sixth should be one on discrete measurements and discrete logic, but written from a process perspective instead of the customary ladder logic and programmable logic controller (PLC) programming perspective. I did a lot of writing while a member of the LSU chemical engineering and computer science faculties from 1966 through 1979. One of my regrets is that I did not continue to write after leaving academia. Only in the last few years did I again take up writing, and it has been a real pleasure.

Fortunately, my wife Charlotte indulges me in such endeavors. If anything, she is too indulging, but I appreciate it more every year.

Cecil L. Smith

Houston, TX

January 2014

1Introduction

Within the span of a half-century, the state of automation in batch facilities has progressed from total manual operation to quite high degrees of automation, even approaching fully automated (or “look ma, no hands”).

Prior to 1970, the instrumentation and controls in a batch facility consisted of the following:

Measurement devices.

  sight glasses for level, pressure gauges using Bourdon tubes, manometers for differential pressure, temperature gauges using bimetallic strips, and so on.

Final control elements.

  hand valves, start/stop buttons for motors, and so on.

Controller.

  process operator and possibly a couple of simple loops for temperature, pressure, and so on.

Printed documents called standard operating procedures, working instructions, or similar were provided to the operator for each batch of product to be manufactured. All data were collected manually, often recorded on these printed documents.

Within the industry, batch was commonly disparaged, and not only from those in continuous segments such as oil refining. Those in batch commonly made derogatory comments about themselves such as “We make a batch of product and then figure out who we can sell it to”: This attests to the high batch-to-batch variability. But as most customers were using these products in processes with similar degrees of automation, the situation was tolerated. “The only reason we have valves is that stopcocks are not manufactured in the size we require”: This reflects the influence of chemists in the decision processes within batch facilities in the chemical industry. Often the plant equipment was merely a larger-size version of the equipment used in the laboratory. The author has visited a batch facility that used wooden vessels for reactors, and not that many years ago!

However, one must be careful with such disparaging remarks. The production volumes might be small, but the profit margins are high. Energy costs are usually a relatively minor part of the production costs. The energy crisis in the 1970s devastated the commodity segment of the chemical industry but had only a minor effect on the batch segment. And corporate management took notice.

Realistically, automation of batch processes only became feasible with the advent of digital control technology. Prior to 1970, the control system suppliers largely ignored batch facilities, basically selling them products developed for other markets. This has changed dramatically. All control system suppliers now offer products with features developed specifically for automating batch processes.

Whether intentional or not, the marketing hype could be understood as “just install our equipment and all will be well.” Unfortunately, one can spend a million dollars to automate a hunk of junk (a few data points are already available; no more are needed). Install a turbocharger on a Model-T and what do you have? A Model-T. Spending money on controls must go hand in hand with modernizing the production equipment.

This book takes a balanced approach to automating batch production facilities that includes the following topics:

Process measurements.

  The basis of every automation system is quality data from the process measurements.

Chapter 2

focuses on the issues that commonly arise in batch facilities.

Control issues.

  All batch processes operate over a range of conditions, which imposes requirements on the controls to do likewise.

Chapter 3

examines such issues.

Discrete devices.

  In most batch facilities, the number of discrete values greatly exceeds the number of analog (actually digital) values.

Chapter 4

examines the various aspects of discrete measurements and final control elements.

Material transfers.

  The material transfer systems are generally capable of either transferring one of several raw materials to a destination or transferring a given raw material to one of several destinations.

Chapter 5

examines various equipment configurations for material transfers.

Structured logic.

  Structured programming and other concepts from computer science are applicable to batch processes, the key being to separate as much as possible equipment-specific logic from product-specific logic.

Chapter 6

presents these concepts.

Batch or process unit.

  Properly defining batch or process units is crucial in any batch automation endeavor.

Chapter 7

discusses these issues.

Sequence logic.

  Batch processing involves conducting the appropriate sequence of activities in the process equipment required to manufacture the desired product.

Chapter 8

examines various approaches to implementing sequence logic.

Batches and recipes.

  In a batch facility in which a variety of products are manufactured in the same production equipment, the product recipe must be the controlling document for manufacturing each product batch.

Chapter 9

examines these and related topics.

1.1.  Categories of Processes

Processes are generally classified as follows:

Continuous.

  Product flows out of the process on a continuous basis. The conditions within the production equipment are ideally constant, but in practice, minor adjustments to operating conditions are occasionally appropriate.

Batch.

  Product is manufactured in discrete quantities called batches. The conditions within the production equipment change as the product progresses through the manufacturing process.

Semi-batch.

  Part of the process is continuous, but other parts are batch. Why not call these semi-continuous?

1.1.1.  Continuous Processes

Continuous processes are normally appropriate for manufacturing a large quantity of a given product, such products being referred to as commodity products. The production facility must manufacture the product in the most efficient manner possible, meaning the lowest manufacturing cost.

But there is a downside. Flexibility is sacrificed to gain efficiency. The facility can do what it is designed to do very, very well. Maintaining the efficiency of the production process is essential—commodity markets are competitive, and even those with deep pockets cannot manufacture a commodity product at a loss for very long. But changes in the nature of the raw materials, the specifications for the final product, and so on, that were not anticipated during the process design can lead to significant and expensive changes to the production facility.

1.1.2.  Batch Processes

Batch processes are most commonly used to manufacture specialty products such as adhesives, cosmetics, and so on. The lower production volumes do not permit a continuous process to be designed and constructed to manufacture a given product. Instead, several “related” products are manufactured by a given suite of production equipment. In this context, “related” means that the manufacture of these products has something in common, one possibility being that the major raw materials are the same for all of these products.

A few commodity processes are batch, two examples being the polyvinyl chloride (PVC) manufacturing processes and the older style of pulp digesters. But to improve production efficiency, efforts are normally expended to develop continuous processes for commodity products. For example, the Kamyr digester is a continuous digester that has replaced many of the older batch pulp digesters.

Using ethylene as an example of a commodity product and adhesives as an example of a specialty product, there is another important distinction. The ethylene being manufactured today is largely the same as the ethylene manufactured 50 years ago. Perhaps the purity of the ethylene produced today is higher, but basically ethylene is ethylene.

However, the adhesive being manufactured today is very different from the adhesive manufactured 50 years ago. Such products are constantly being tailored to their application, with changes to make it set faster, provide a stronger bond, and so on. Sometimes, tailored products are developed to address a subset of the applications, such as providing a better bond between two specific materials. The number of different adhesives available today is much greater than 50 years ago.

The flexibility of the production facility for a specialty product is more important than its production efficiency. Specialty products tend to evolve, thanks to a development group whose mission is to do exactly that. Many specialty products must perform in a certain manner—adhesives must bond two materials. If an adhesive can be modified to shorten the time required to set, customers view this in a very favorable light, preferring this product over that from competitors and possibly willing to pay a higher price for the product (a shorter setup time makes the customer's production process more efficient). Development groups are tasked with coming up with product enhancements, producing sufficient quantities of the enhanced product for customers to test, and then moving the enhanced product to the production facilities. In many specialty production facilities, predicting what products will be manufactured a year from now is difficult, but rarely is it the same as being manufactured today.

1.1.3.  Semi-Batch Processes

A semi-batch production facility uses batch equipment for part of the production process but continuous equipment for the remainder. A common structure for a process is reactions followed by separations. The following approach is possible (and the opposite arrangement is also possible):

Use batch reactors for reactions.

Use continuous distillation for separations.

The batch reactors discharge into a surge vessel that provides the feed to the continuous stills. On a long-term basis, the throughputs for the two parts of the process must be equal, either by adjusting the frequency of reaction batches, by adjusting the feed rate to the stills, or some combination of the two. On a short-term basis, the throughputs can be different, subject to the constraints imposed by the capacity of the surge vessel.

Many continuous processes utilize batch equipment for certain functions. Oil refineries are normally viewed as the extreme of a continuous process. However, batch operations are commonly used in the regeneration of a catalyst. A production facility with a small degree of batch processing within an otherwise continuous process is not normally deemed to be semi-batch.

1.2.  The Industry

Flexible batch production facilities are commonly encountered in the specialty chemicals, nonseasonal food processing (coffee, ice cream, etc.), agricultural chemicals (pesticides and herbicides), and similar subsets of the process industries. The companies range from large multinationals to small companies that manufacture products targeted to customers with very specific needs.

1.2.1.  Intellectual Property

Go to a paper company and open a discussion on batch pulp digesters. They are generally receptive to having such discussions. Most will admit that what they are doing in the pulp digesters is essentially the same as what their competitors are doing. In some cases, they even visit each other's facilities. They do not view any of their competitive edge being derived from what they are doing in the batch pulp digesters.

With few exceptions, the situation in the specialty batch industry is at the opposite extreme. In some of these facilities, only authorized company employees are allowed to enter. Others will be admitted only when there is a specific need for them to be there, and even then, they are escorted in, they do their job, and they are escorted out. They are only told what they need to know to do their job and are not allowed to just “look around.”

There is a valid reason for such approaches. Patent protection has a downside. To be awarded a patent, the filing must include sufficient information so that one “skilled in the art” can reproduce whatever is being patented. To obtain a patent, you have to divulge considerable information relating to your technology. For some companies, this is unacceptable—divulging such information tells your competitors some things that you do not want them to know. This information is a trade secret, which is intellectual property subject to protection. How do you keep a secret? Don't tell anybody! In any trade secret litigation, one must show that extraordinary steps are taken to prevent others from learning about the technology.

This complicates preparing presentations and publications relating to such industries. Anything relating to the process technology has to be expunged. Everyone in the industry likes to tell “war stories.” One has to be extremely careful with these, or one can unintentionally divulge information. Even details that seem innocuous to control specialists and the like can be very informative to someone “skilled in the art” and employed by a competitor.

Consequences arise in books like this. Concrete process examples must be avoided. The next section uses the kitchen in one's home as an example of a batch facility. Why? Recipes are widely available in cookbooks. No specialty batch company will allow its recipes to be used as the basis for discussion. Even product names are best avoided, instead using made-up names such as “hexamethylchickenwire.” Besides, most of the actual chemical names are so specific to a certain product line that the typical reader would have to look them up anyway. Even worse, a Google search will likely yield little, if anything at all.

1.2.2.  Manual Operations

In older facilities, operators provided the process controls. A flow control loop consisted of a vessel with a sight glass, an operator with a clock, and a hand valve. The operator was provided a charge table stating that the vessel level should be x at the start, y after 30 minutes, z after 60 minutes, and so on. The operator adjusted the hand valve so that the flow rate of material out of the vessel resulted in the desired vessel levels. And if you believed what was manually recorded on the production logs, these operators were good! Similar technology was applied to all parts of the batch manufacturing process.

Why did such an approach work? The manufacturing process relied on forgiving chemistry. As a consequence, many of the products could be manufactured in a bathtub outfitted with a good mixer. Just get about the right amount of materials in the tub, agitate it thoroughly, avoid large temperature excursions, and so on, and the resulting product could be sold to someone.

Rarely were products manufactured to order. The plant produced a batch of product, QC performed their analysis to determine the characteristics of the product, and then sales directed this product to the customers that preferred these characteristics. This approach had a downside—periodically, the warehouse had to be purged of product that nobody wanted to buy.

The early attitude was not to automate these processes, but to convert them to continuous processes. But constructing a process to manufacture small amounts of a specific product made no sense economically, especially considering that some of the products had a short life cycle (replaced by a slightly different product with characteristics preferred by the customers). But as increasing energy prices eroded the profitability of many continuous processes, batch facilities became major sources of corporate profits.

1.2.3.  Driving Force for Change

With the “if it ain't broke don't fix it” philosophy that prevails in the industry, something is required to drive changes. For batch processes, this something came in the form of an evolution from forgiving chemistry to demanding chemistry. The new products coming out of the development groups were targeted to customers with specific needs, which meant that the products must have specific characteristics.

Tighter tolerances on product characteristics necessitated tighter tolerances during the manufacturing process. Almost overnight, forgiving chemistry was replaced by demanding chemistry. At about the same time, statistical quality control, just-in-time production operations, and the like were recognized as the path forward within specialty batch facilities. Issues such as the following had to be addressed within the production facilities:

Tolerances on the total amount of each raw material charged became tighter. Flow measurements consisting of a sight glass and an operator with a stopwatch could not deliver the required performance. Sight glasses were replaced by load cells. Coriolis meters quickly became the preferred flow measurement, in spite of their higher cost.

Temperature control became essential. The requirements evolved from “just avoid large temperature excursions” to “control within a few degrees” to “control within 0.5°C.” Even “control within 0.1°C” is likely to arise, if not already. Meeting such requirements necessitates that changes such as the following be made:

Improved temperature measurement (an automation issue).

More sophisticated temperature control configurations (an automation issue).

More responsive heat transfer arrangements (a plant equipment issue).

Enhancements to the controls must be accompanied by enhancements to the plant equipment.

1.2.4.  Product Specifications

Those from the continuous industries often find those in the batch industries extremely reluctant to change. Many times this is merely resistance to change, but sometimes there are valid reasons that relate to the nature of the product specifications.

Previously, ethylene was cited as a product produced by a continuous process. One specification is the purity of the product. The purity can be analytically determined using a chromatograph. Occasionally, arguments arise as to whose chromatograph is correct, but once these are resolved, the ethylene either meets the required purity or it does not.

Adhesives were previously cited as products produced by a batch process. Product specifications are developed that state the characteristics that an adhesive must possess to be suitable for a certain application. But in the end, the product must perform properly in the application. If an adhesive is to bond two specific materials, it could conceivably meet all product specifications but not provide the expected bond. This does not arise very often, but one must always remember that the product does not just have to meet specifications; it has to perform in the promised manner. This complicates making changes in a batch facility. Even what seems to be a minor change could potentially have side effects on how the product performs.

The situation is far more serious for a product with a guaranteed life, an example being that the bond provided by an adhesive must remain effective for 10 years. If these bonds deteriorate after 8 years, customers will be more than unhappy—they will start litigating. Liabilities can be huge, potentially risking the financial viability of the company manufacturing the adhesive. In this context, the reluctance to make even simple changes in a batch process are understandable. That even minor changes in how a product is manufactured have the potential to place the company at risk is a very sobering thought.

1.2.5.  Automation Technology for Batch

One could get the impression that the control system suppliers were out in front of the curve on this one. Not really. When the need for automation arose within the specialty batch industries, the available equipment for process controls fell into two categories:

Distributed control system (

DCS

).

  Although developed by the process instrument companies, oil refining and other large continuous processes were the target market. Not much for batch automation, but in defense of the suppliers, the specialty batch industries were not spending money on automation. This aspect has changed dramatically.

Programmable logic controller (

PLC

).

  Developed primarily for the automotive industry, this technology was suitable for manufacturing home appliances, toys, electronic equipment, and so on. PLCs offered robust I/O, could do sequence logic if programmed in relay ladder logic, were very cost-effective, and developed a reputation for high reliability. However, the PLC manufacturers were slow to shed the automotive industry mentality—very understandable for a company whose major customer is General Motors.

More on this in a later section of this chapter.

The know-how for automating a specialty batch facility resided in the chemical companies, not in the control system suppliers. In 1993, the German organization Normenarbeitsgemeinschaft für Meß- und Regeltechnik in der Chemischen Industrie (NAMUR; name translates to “Standards committee for measurement and control in the chemical industry”) prepared a document [1] in German whose title translates to “Requirements to be met by systems for recipe-based operations.” With input from the major German chemical companies, the document laid out the requirements for control systems suitable for automating specialty batch processes. That such a document was deemed necessary as late as 1993 speaks volumes.

When one tours different specialty batch facilities, one sees a lot of the same equipment. The production companies purchase their glass-lined reactors from the same vessel fabricators. They purchase their chemical pumps from the same pump manufacturers. They purchase their measurement devices and controls from the same suppliers. The difference is not what equipment is being installed, but what is being done within that equipment.

Here, one must distinguish between equipment technology and process technology. With few exceptions, the equipment technology varies little from one company to the next. But what one company is doing within that equipment is far different from what another company is doing. The latter is proprietary, and companies go to great lengths to assure that what is being done within the equipment will not be discussed.

1.2.6.  Safety and Process Interlocks

The use of the term “interlock” varies. Some would restrict the term to safety systems and have proposed terms such as “process actions” for similar logic implemented within the process controls. An alternative is use “interlock” to designate the logic required to mitigate the consequences of some event, but to specifically distinguish between the following:

Safety interlock.

  Interlocks pertaining to personnel protection and to prevent major damage to the process equipment are considered to be highly critical. In most cases, these are relatively simple, but they must be implemented by experienced personnel using the greatest of caution. The design must be based on a detailed analysis of the hazards that could arise. Thereafter, all proposed changes must be thoroughly reviewed. Modifications to the implementation must confirm that all functions (both those that have changed and those that have not changed) perform as specified.

Process interlock.

  The purpose of process interlocks is to prevent “making a mess” (such as overflowing a vessel), to prevent minor equipment damage that can be quickly repaired (running a pump dry), and the like. These are normally implemented within the equipment that provides the process control functions. Process interlocks range from simple to extremely complex. Manual overrides are sometimes provided within the logic, but if not, knowledgeable personnel (such as shift supervisors) can usually disable a process interlock. Should production be interrupted for hours due to the failure of a simple limit switch that provides an input to a process interlock, plant management will not be amused. But on the other hand, discipline must be imposed so that the problem that required the override be promptly corrected so the override can be removed.

Herein the terms will be used in this fashion.

An interlock is an implementation of a logic expression whose result is used to force a field device to a specified state, regardless of the desires of the process controls or the process operators. Classifying as a safety interlock or a process interlock has no effect on the discrete logic expression. However, it determines how the interlock will be implemented:

Safety interlocks.

  Critical interlocks are usually relatively simple, but they must be implemented in highly robust equipment. The options are the following:

Hardwired circuits.

Programmable electronic equipment dedicated to this function. This typically translates to a PLC. Only discrete I/O is normally required, so a low-end model can be used. This is also consistent with keeping safety equipment as simple as possible.

Process interlocks.

  These are normally implemented within the process controls, either a DCS or a high-end PLC.

1.2.7.  Safe State

The safe state is the state of the final control element on loss of power. To force one or more final control elements to their safe states, a common approach is to remove the power; that is, either disconnect the power supply to electrical equipment or shut off the air supply to pneumatic equipment.

Pneumatic measurement and control equipment has largely disappeared, but with one very important exception. Even today, the actuator for most control valves is pneumatic. The pneumatic diaphragm actuator possesses a characteristic that is very desirable in process applications. On loss of either the control signal or supply air (the power), the actuator for either a control valve or a block valve can be configured in either of the following ways:

Fail-closed.

  On loss of power, the control valve is fully closed (the actuator is “air-to-open”). The “safe state” for such valves is the closed state.

Fail-open.

  On loss of power, the control valve is fully open (the actuator is “air-to-close”). The “safe state” for such valves is the open state.

To date, replicating this characteristic in an economical manner has proven elusive for electric actuators for positioning valves. It is possible for solenoid actuators, but even so, many block valves are equipped with pneumatic actuators.

What determines the safe state? From a purely control perspective, there is no basis for favoring fail-closed versus fail-open. The appropriate state for the safe state is determined as part of the analysis of process hazards. On total loss of power, the actuator for the final control element must be configured so that its failure state is the same as the safe state. Once the safe state has been selected, the control logic can be formulated to perform properly.

The logic for safety interlocks must be implemented in such a manner that the final control element is permitted to leave the safe state only if the appropriate conditions are met. If the conditions are not met, the final control element is forced to the safe state, usually by blocking power. But if the conditions are met, the final control element does not necessarily exit the safe state. This only occurs on the following conditions:

Either the process controls or the process operators command the device to exit the safe state.

Power is applied to the actuator, which means the interlock is not being imposed.

With this approach, safety interlocks and process interlocks differ as follows:

Safety interlocks.

  Being imposed external to the process controls by disconnecting power from the actuator, the safe state must correspond to the state resulting when all power is removed.

Process interlocks.

  Being implemented within the process controls, a process interlock imposes a specific state on a field device by forcing the control output to the appropriate value. As the actuator remains powered, a process interlock can impose a state on a field device other than its safe state. The preference is that a process interlock imposes the safe state, but an occasional exception is usually tolerated.

1.2.8.  Safety Issues Pertaining to the Product

The previous discussion on interlocks and safe states is from the perspective of equipment safety. In this regard, the issues in batch facilities are largely the same as in continuous facilities.

But in batch processes, another issue arises that is rarely encountered in continuous processes. In the chemical industry, most batch processes utilize reactive chemicals. A variety of situations can arise, examples of which are the following:

The reaction proceeds as expected if the order of addition is to add chemical A, then add chemical B, and finally add chemical C.

If the order of addition is first A, then C, and finally B, a rapid reaction ensues that overheats and overpressures the reaction vessel.

Interlocks of the type used for equipment protection are far less effective for issues such as this, if even applicable at all.

Assume the reactor has dedicated equipment for feeding each of the chemicals, with individual block valves for each feed. Following the approach for interlocks to address equipment issues, provision could be made to supply power to the block valve for feeding chemical C only when feeding C is acceptable. But who or what turns on the power to the valve? The possibilities include the following:

Process controls.

  Presumably the process controls would only feed chemical C at the appropriate times. If the process controls cannot be trusted to feed chemical C only at the appropriate times, why trust the process controls to turn on power to the Feed C block valve only at the appropriate times?

Operations personnel.

  Often this means the shift supervisor or equivalent. The “switch” or its equivalent that supplies power to the feed C block valve could be completely independent of the process controls. The obvious issue is the potential for human error. But in the end, is this a “we could not come up with another way to do it, so we left it to the operators” type of solution?

Another issue that will be discussed in a subsequent chapter pertains to charging the wrong chemical. The intent is to charge B, but D is charged instead, the result being to first add A, then add D, and finally add C. What happens when D is mixed with A, that is, do A and D react? What happens when C is added to the mixture of A and D? Answering these questions raises another question: what is D? There is no way to answer this question. Regardless of what it is, chemical D should not be present in the reacting medium.

Such situations are the results of mistakes. Possibly the truck delivering D was unloaded into the storage tank for B (which means a mixture of B and D was actually added). In the specialty chemicals facilities, the number of chemicals is so large that material is often charged directly from drums. This raises the possibility that one or more drums contained D instead of B. The obvious answer is “we must not allow such things to happen.” But exactly how are such situations avoided? Interlocks are not the answer.

And to further drive home how serious this situation can be, sizing relief valves in the chemical industry involves two-phase flow calculations for which the nature of the materials flowing through the relief valve must be known. Adding the wrong chemical has the potential of initiating a runaway reaction that overpressures the vessel. But due to the following, providing adequate protection is not assured:

There is no way to know what chemical will be mistakenly charged.

There is no way to know what chemicals are flowing through the relief valve.

There is no way to be absolutely certain that the relief valve is adequately sized for the situation resulting from the mistake.

This can cause people to lose sleep at night.

1.3.  The Ultimate Batch Process: The Kitchen in Your Home

One measure of a batch process is its flexibility in the range of products that can be manufactured within the facility. By this measure, the kitchen in your home must rank very, very high.

1.3.1.  Recipe from a Cookbook

The recipe is the controlling document in a kitchen. Table 1.1 is a recipe for “Mom's Pecan Pie.” The terminology is different, but product chemists generate recipes for products with a surprisingly similar structure. There are two main sections:

List of ingredients.

  Chemists use alternate terms such as

formula

(used herein) or

raw

materials

. In either case, the objective is the same: convey what is required to make a batch of the product. Often the chemists will state the formula on the basis of a standard batch size, such as 1000 kg, which can then be scaled to the size of the production equipment.

Directions.

  Again, chemists use different terms, such as

processing instructions

or

procedure

(used herein). As in the pecan pie recipe in

Table 1.1

, chemists often itemize the processing instructions as step 1, step 2, and so on. The objective is to state precisely what one must do with the raw materials to manufacture a batch of product.

Table 1.1.  Example of a Recipe

Mom's Pecan Pie

List of Ingredients

4 eggs

200 gm (1 cup) sugar

240 mL (1 cup) light Karo

4 gm (½ tbsp) flour

1½ gm (¼ tsp) salt

5 mL (1 tsp) vanilla

60 mL (¼ cup) butter

220 gm (2 cups) pecans

Directions

Beat eggs until foamy.

Add sugar, Karo, flour, salt, and vanilla.

Beat well.

Stir in melted butter and pecans.

Pour into pie crust.

Bake 60 minutes at 175°C (350°F).

One of the steps in the recipe in Table 1.1 states, “Mix until foamy.” Recipes dating from 50 years ago often contained similar statements. Vessels were often open or outfitted with manholes, permitting operators to visually assess certain aspects pertaining to the batch. Virtually all vessels are now closed, but some nozzles are equipped with sight glasses and perhaps a tank light. However, the trend is to replace qualitative instructions with specific instructions such as “Mix at medium speed for 10 minutes,” the objective being to eliminate as much manual intervention as possible.

1.3.2.  Home Kitchen versus Commercial Bakery

One observation about recipes for baked goods is in order. The recipe in Table 1.1 is for a home kitchen. Commercial bakeries are very different. Generally, the starting point is a large mixer that combines ingredients for a large number of cakes or pies. These are then formed into individual products and placed on a conveyor belt that passes through the oven. In effect, the result is a semi-batch plant—batch for mixing and continuous for baking. More efficient for making pies but far less flexible than the home kitchen. Baking the Thanksgiving turkey using the equipment in a commercial bakery would not turn out well.

1.4.  Categories of Batch Processes

Segmenting the product recipe into a formula and a procedure permits batch processes to be classified as per Table 1.2. The capabilities required of the control system vary greatly with the nature of the batch process, with cyclical batch being the simplest to control and flexible batch the most demanding.

Table 1.2.  Categories of Batch Processes

1.4.1.  Cyclical Batch

One approach to catalyst regeneration is to provide two catalytic reactors in parallel. While one is in service, the catalyst is being regenerated in the other. Cyclic catalytic reformer units in oil refineries use this approach. While in service the process stream flows through the reactor in one direction. With time, carbon buildup on the catalyst decreases its effective surface area. The catalyst is regenerated by essentially burning the carbon buildup by blowing hot air in the opposite direction.

This is a batch process where a batch consists of one cycle of catalyst regeneration. Logic is required to switch between the two reactors, start the hot air flowing to the one being regenerated, and so on. However, this logic is exactly the same from one batch to the next. The objective is to burn sufficient carbon off of the catalyst to restore its effectiveness, but without unnecessarily degrading the catalyst. A key operating parameter is the temperature of the hot air used to regenerate the catalyst. Usually this is the set point to a temperature controller. A formula in the sense of the list of ingredients in a recipe is not normally used.

For a process to be categorized as cyclical batch, Table 1.2 states the following requirements:

Formula.

  If a formula is used at all, no changes in operating parameters are normally implemented from batch to batch.

Procedure.

  Identical logic is used to manufacture every batch.

1.4.2.  Multigrade Batch

The older style of batch digesters in the pulp and paper industry is batch. Batch digesters are essentially very large pressure cookers. Each batch or “cook” involves the following:

Charging the digester with wood, water, and chemicals

Heating the digester to the desired temperature

Maintaining the temperature for the specified time

Releasing the pressure (or “blowing” the digester) into the recovery system

Discharging the contents.

These steps do not vary from one cook to the next. However, selected operating parameters can be changed between batches. One approach is to present a data entry screen to the process operator at the start of each cook. This data entry screen contains parameters such as cook temperature, cook time, chemical-to-wood ratios, and so on, that can be adjusted for purposes such as the following:

The water content of the wood is not measured but is not constant. Based on the kappa number (an assessment of the degree of cooking) from previous cooks, the operator adjusts the chemical-to-wood ratios in the formula.

The production rate can be increased by increasing the cook temperature and shortening the cook time (cook temperature and cook time are related by a parameter known as the H-factor).

Prior to each cook, the operator is given the opportunity to change the parameters for the next cook. For most cooks, the operator uses the same parameters as for the previous cook. But even so, no cook is initiated until the operator confirms the parameters to be used for the cook.

The PVC process is another example of a multigrade batch process. There are distinct grades of PVC. A set of operating parameters (including quantities such as raw material amounts) are defined for each. The logic is identical (or at least largely the same) for each grade of PVC; only the parameters must be adjusted.

For a process to be categorized as multigrade, Table 1.2 states the following requirements:

Formula.

  The process differs from batch to batch. For each batch, a set of parameters must be specified. Even though they are often the same from one batch to another, the operator must confirm their values prior to the start of each batch.

Procedure.

  The logic for this batch is identical to the logic for the previous batch. In this context, “identical” is relaxed to permit minor variations. One approach is to include a logical variable within the parameter set that specifies if a certain action is to be performed. This works well if the possible variations are limited but is cumbersome if the number of possible variations is large.

Some early batch control systems used the term “recipe” to designate what is really only the “formula” in the terminology used today. This worked well for the multigrade category of batch processes but not for the more complex flexible batch category. Today, the term “recipe” is used to encompass both the formula and the procedure.

1.4.3.  Flexible Batch

This type of batch facility is commonly encountered in specialty chemicals, with emulsion reactors being a good example. Certainly, it is possible for two successive batches to be for the same product. However, it is also possible that the norm is successive batches of very different products.

This is also complicated by the structure of the facility. For latex paint, three distinct functions are required to produce a batch:

Prepare the emulsion feed for the batch.

  The proper materials are charged to the emulsion feed vessel, mixed properly to form the emulsion, and then cooled (removing heat from the emulsion feed vessel is easier than removing heat from the reactor).

Carry out the polymerization reaction.

  Normally, some materials are charged to the reactor, the emulsion is fed (possibly with one or more other feeds) over a period of time (often determined by the rate that heat can be removed from the reactor), and then the reaction is driven to completion.

Perform post-processing.

  An example is to filter the product.

The nature of the equipment is such that batches overlap. For example, when the emulsion feed to one batch is completed, preparing the emulsion feed for the next batch can be started. Similar possibilities exist between the reactor and the post-processing. As successive batches could be for different products, a recipe (formula and procedure) must be associated with each batch. This is quite analogous to what can occur in a home kitchen—while baking a pecan pie in the oven, the cook can be mixing the ingredients for a cherry pie.

Due to the limitations of the control equipment, some early automation efforts for such processes attempted to develop a formula with logical variables to govern what actions to perform for each batch. The resulting complexity was enormous. For example, suppose materials A, B, and C are required in most batches. The formula certainly includes entries for the amount of each. In some batches, the procedure is to add A, then add B, and finally add C. But in other batches, the procedure is to add A, then add C, and finally add B. As the number of raw materials increases, the possible variations increase rapidly.