As recently as , process control was hardware limited. Since then primary measuring elements, controllers, computing devices, control stations, and control valves have improved greatly in reliability, sensitivity, and speed of response. Consequently these characteristics less and less frequently pose limitations to the single-loop designer. Further, the quality of control achieved by single-loop systems is not greatly affected by type of hardware; it makes little difference, for a given algorithm, whether one uses analog pneumatic or electronic gear, microprocessor controls, or a digital computer.
Multivariable Control To avoid the limitations of single-loop design and to provide a more flexible and sophisticated process operating logic than can be implemented by human operators, we use an approach we call multivariable control. Many definitions of this term may be found in the literature, but most of them are expressed in mathematical terms rather than in terms of process hctions. For our purposes we define a multivariable control svstem as one that has the built-in intelligence to look simultaneously at two or more process variables and to choose, in a given situation, the best of several preprogrammed strategies algorithms for manipulating one or more control valves or other final control elements.
For example, the steam valve for a distillation column reboiler, depending on circumstances, may respond to controllers for: Steam flow rate Column AP Column pressure Base temperature Column feed rate Column base level Column bottom-product rate The seven variables listed may also exert control on five or six other valves. To provide automatic control of this sort, we make extensive use of variable configuration controls that are usually implemented by overrides. If, for instance, base composition is normally controlled by steam flow that can be taken over or overriden by high column AP, this is a variable configuration.
Multivariable control may involve both variable configuration and variable structure controls. Hardware permits us to automate this kind of control with a speed, precision, and reliabilinr that are completely beyond the capabilities of human operators. It is common to think of process control fimctions stacked one above another in a pyramid or hierarchic arrangement as with traditional business or military organization structures, particularly when computers are involved. Multivariable control structures, however, with their extensive lateral and diagonal crossovers, are functionally more like the modern matrix concept of management.
By now it is probably apparent that we are striving for control system designs whose performance and design parameters are specified in advance of plant startup. In practice we furnish calibration data for controller parameters and computational devices for the majority of control loops prior to startup.
We calculate these from simulations or simple linear models. For microprocessor computer controls, we calculate scaling parameters for computation blocks either in software or hardware. Our design procedures are accurate enough that only a modest amount of empirical controller tuning is required at startup. Stability, Speed of Response, and Interactions Most existing literature on automatic control is concerned with the stability and speed of response of single loops. The traditional objectives of feedback control system design are:. To get the fastest possible response to set point changes.
To compensate for or to attenuate disturbances as much and as quickly as possible. These must be accomplished with a reasonable degree of closedloop stability. In process control the objectives are often quite different. The objectives of averaging level control, for example, clearly are different from those just mentioned. A typical chemical plant or refinery has hundreds of single loops with many interactions among them.
It is usually far more important to design for a dynamic balance among these loops with a minimum of interaction than to strive for maximum speed of response. Further, it is usually undesirable to make rapid changes in manipulated variables since these may upset the process. For example, distillation column reflux flow and boilup should not be changed too rapidly since these might cause transient flooding or weeping in part of the column.
Our preferred philosophy of controller tuning is discussed in a book and a paper. There are five simple methods by which to make a system noninteracting:. Design material-balance control loops to be at least a factor of 10 slower than related composition control loops. Similarly, in cascade systems, make the secondary or slave loop at least a factor of 10 faster than the master loop. Avoid designs that are intrinsically interacting, such as pressure control and flow control at the same point in a pipeline.
One of the two controllers must be detuned. Select process designs that eliminate or minimize interaction. Use override circuits. Although not specifically intended for this purpose, override circuits, by permitting only one controller at a time to control a given valve, eliminate interactions. Use interaction compensators decouplers. If two control loops, such as top and bottom temperature controls on a distillation column, interact, we can eliminate the interactions by installing two compensators.
One compensates for the action of the top temperature controller on the bottom loop while the other compensates for the action of the bottom temperature control loop on the top one. This is discussed in Chapter These may be analog pneumatic, analog electronic, or microprocessor based. In the last case, the station may be physically distinct, like an analog station, or may be represented on a CRT display as a faceplate.
Each provides an indication of the process variable flows, level, temperature, etc. There is also a manualautomatic switch, which some vendors label hand-automatic. In the manual mode, the feedback controller is disconnected and there is a knob that enables the operator remotely to set the valve position. This may or may not be subject to restrictions imposed bv feedforward compensation, overrides, and so on, depending on the design philosophy for a particular project. If cascade control is involved, as, for example, liquid level control cascaded to flow control, the secondary station not only has manual-automatic switching, but also another fUnction-remote-locaI.
In the remote position, the secondary controller set point comes from the output of the primary controller. In the cloca position, valve position may be set manually or the controller set point may be set by the operator clocal-auto. Although cascade functions are sometimes combined into one station or CRT faceplate for space and money-saving reasons, we recommend dual stations.
Most single-station designs with which we are familiar are very inflexible and complicated; they do not permit ready implementation of feedforward, overrides, and so forth. Better measurements, especially of compositions. Better control valves. With regard to these, more progress has been made in the design of valve bodies and trim than in the design of actuators. Valve positioners should always be used it is assumed in this book that they will be.
As far as CCR hardware is concerned, we have a decided preference for microprocessor controls. They are technically more versatile and are less expensive some versions than analog. As of late , many are featuring satisfactory antireset windup and override capabilities. In adchion, they provide more advanced logic capability, dead-time simulation, and adaptive tuning. Some of the last named achieve self-tuning via stochastic techniques or by pattern recognition.
Others have gain scheduling, where reset time and proportional gain are functions of some process variable or the controller error signal. Microprocessor controls usually have a sampling time of a fraction of a second. Although slightly slower than analog controls, their performance can generdy be approximated by analog control algorithms. Other advantages include freedom from drift, and the fact that they can be calibrated more precisely, can be reconfigured or restructured without wiring changes, have a larger range of tuning parameters, and contain more control algorithms.
For most projects today, it is possible to find worthwhile applications for a supervisory digital computer with a good data historian, regardless of the type of basic controls selected pneumatic, electronic, or microprocessor. Digital readouts for important variables are worthwhile because they permit seeing their magnitudes with sensitivity approaching that of the original analog measurements.
Most typical analog measurements have a sensitivity ranging from one part per to one part per 10, Most analog readout devices, however, are limited to 0. For maximum advantage a supervisory computer should be programmed to have the control algorithms discussed in Chapter These are position rather than velocity algorithms. It is our opinion that using such a computer to imitate unenhanced two- and three-mode analog controls is poor practice.
Some worthwhile applications for computers will be discussed later. Computer consoles were originally provided in the CCR for the convenience of the operators. Sometimes a consolidated console is also provided for production supervision. Engineers consoles, perhaps at another location, facilitate technical studies. Separate consoles for maintenance personnel the computer can be a powerful maintenance tool are highly desirable. For both there are substantial, coercive stem forces when the valves are in flowing streams. Positioners compensate for this and maintain the valves inherent flow characteristic expressed as a function of controller output signal.
It should be noted, however, that some users prefer not to use positioners. After careful discussion with the process engineer and column designer, and after careful review of the overall process flow sheet, prepare a simplified flow sheet that defines the control concepts: a. Select the overall material-balance control scheme first, preferably l l proceeding from final product inventory to raw material inventory.
A individual equipment-piece material-balance controls must be consistent with this scheme. Select composition control schemes.
Add feedforward and interaction compensators as required. Add miscellaneous temperature and pressure controls. Prepare material-balance and composition-control signal flow diagrams. Determine holdup volumes required for smooth material-balance control and for liquid-level override controls at each end of the column. With holdups determined calculate column-composition transfer functions.
Select measurement spans and calculate control valve sizes. Calculate feedforward and interaction compensators. Calculate a l l other overrides. Calculate feedback-controller gain and reset settings and control-loop natural frequencies. Check feed-tank material balance and mixing time constants for adequacy. Use simulation for some columns, particularly those in critical service or with a new untested control system or process configuration.
Simulation of the column and its control system will be usefd in confirming control concepts and controller tuning parameters. It may also save startup time. Hardware vendors may now be selected and the measurement and control equipment may be ordered. Part I1 of this book deals with the qualitative and heuristic aspects of steps 1 and 3.
Quantitative information for the other steps is presented in Part This is due to the trend toward increasinglv tight column design. Column diameter and tray spacing are now kept to smaller values, with the. The combined effect of these design policies is to make columns much touchier and harder to control. Experience indicates that typical incremental instrument investment over that required for unenhanced feedback controls is percent large projects tend toward the lower figure.
This incremental investment not only provides better normal control, but it also helps to avoid inadvertent shutdowns. It is, therefore, our opinion that these controls should be used to some extent on almost every column. Suppose, however, that the customer insists on minimum application of feedback controls with no feedfonvard compensation or overrides. What column design philosophy should be followed? Having had considerable adverse experience with columns with primitive controls, particularly sidestream drawoff columns, and columns with heat recovery schemes, we suggest the following:.
Design for normal operation at 60 percent of the rate for flooding. Provide five extra trays or increase the number of trays by 10 percent, whichever is larger. Provide increased tray spacing. Provide larger condensers. If water cooled, use tempered water. Provide larger reboilers lower heat flux. Control column Al by boilup, or flow control steam. Provide percent reserve capacity in heat-recovery schemes or avoid them altogether.
Avoid sidestream drawoff designs. Provide surge tanks between columns with at least minutes of holdup each. Frequently encountered problems include unstable or ineffective controls, off-specification product or products, and f l d g or dumping. In addition, it is fairly common practice to use excessive boilup and reflux to make sure of meeting or exceeding product specifications. This not only wastes energy; it also reduces column capacity. To save this amount of steam would probably be only a modest accomplishment for most columns.
If composition control of each product stream is desired and this is usually. Most commonly temperature in the upper or lower section of the column or both is used in lieu of true composition measurements. Frequently composition control is attempted at only one end of the column, and sometimes at neither end. Another shortcoming frequently observed is the use of fixed flow controls for steam, reflux, or product drawoff.
Any such unaided flow control should be regarded with suspicion. With rare exceptions flow-control set points must be changed to accomplish either composition control or material balance control. In view of the preceding comments about problem areas and likely opportunities for improvement of composition control and reduction in energy consumption, the following guidelines are suggested:. Make sure that column material-balance controls are properly designed and tuned, and that hardware, especially level and flow transmitters and control valves, is in good working condition.
If PI level controllers areused, follow the tuning procedures of Chapter 16; auto overrides or nonlinear controllers should be used. It is usually desirable to cascade level control to flow control, in which case flow measurement should be linear. Provide averaging level control of column feed. Column feed rate should also be held between maximum and minimum limits. Provide a linear flow measurement for steam flow control.
Also provide temperature and pressure compensation. If condenser controls are a problem, review the control schemes in Chapter 3. At this point with flows established, smoothed, and in some cases limited, it will probably be possible to see some improvement in composition control, at least for part of the time. For further improvement provide steam-to-feed ratio control, internal reflux-to-feed or distillate-to-feed ratio control, high A P override on steam to protect against flooding, and a minimum steam flow ltmiter to protect against dumping.
If composition is measured and is cascade controlled via reflux or boilup, or both, ratio controls should be replaced by impulse feedforward compensation. For pressurized or vacuum columns, make sure an adequate scheme is provided for maintaining an inert balance. If temperature by itself is not an adequate measure of composition, consider one of the schemes in Chapter 10 for using temperature, pressure, and, in some cases, flow measurements to deduce compositions.
Throughout the remainder of this book, we will frequently illustrate control schemes with pneumatic components. This is done for convenience. ALL pneumatic controllers, as far as we know, can achieve antireset windup by the same external reset feedback method. This appears to be the most universally usell method we use it extensively , and has been adopted by one manufacturer of analog electronic instruments and by several vendors of microprocessor-based distributed controls. Other vendors of electronic analog and digital controls feature a wide variety of techniques; some of these work fairly well, while some are quite inflexible.
A brief discussion is presented in Chapter Units used in this book are those commonly employed in chemical engineering: pounds, feet, degrees Celsius, mols, and so on. To facilitate the calculation of control engineers time constants, we have mostly used time units of minutes or seconds [e.
For projects that make partial use of metric or SI units, we found it convenient to convert them to the above units. Generally speaking, we have found no advantage in writing equations with SI units. Instead, in programs for computers or programmable calculators, we mostly write the equations in the older units and add subroutines for going back and forth to metric or SI units. The most common type of controller used in the chemical and petroleum industries was once called proportional plus automatic reset, later shortened to proportional reset.
Today it is more common to use PI, which stands for proportional-integral. We will also use reset time, usually in minutes, rather than its older reciprocal, repeats per minute. For drawings in which we are trying to present a perspective of control concepts configurations , we use very simplified symbols. In drawings where we are trying to illustrate concepts of structure overrides, feedfonvard, etc. In pneumatics it is c o m o n to refer to most signal-conditioning devices other than controllers as relays. Included are adders, subtractors, multipliers, and so on.
For electronic analog and digital controls, it is more common to use terms such as signal scalers and mult lien. A table of nomenclature and symbols will be found at the end of this book. For anyone seriously interested in distillation control, two books are highly recommended. The first is an easy-to-read, nontheoretic as far as control is concerned work by F. The treatment of energy conservation alone is worth the price of the book. The second book is by Rademaker, Rijnsdorp, and Maarleveld. It also contains an extensive bibliography. For basic reference books on distillation, we have made much use of those Treybal,l6 and Hengstebeck by Van Winkle13 and King.
For basic books on control, we recommend two written by the one by Harriott, and one bv Murri Nonchemical engineers with no background in distillation may find an introductonr text bi7 Nisenfeld and Seemann useful. Maarleveld, DynamicJ and Control 1. Buckley, P. Harbert, W.
Van Winkle, M. New York, Ref 29 10 : King, C. McGraw-Hdl, New York, Ref 35 11 : Holland, C. Modelling o f Separation Processes, D. Lamb, 5. Luyben, W. Hill, New York, Uitti, K. Trevbal, R. Hengstebeck, R. Symposium, Jan. Harriott, P. Koppel, L. Shinskey, F. McGraw-Hill, New York, Rademaker, O.
Rijnsdorp, and Douglas, J. Gould, L. McAvoy, T. Addson-Wesley, Reading, Mass. Stephanopolous, G. McGraw-Hd, New York, Nisenfeld, E. Seemann, Murrill, P. Triangle Park, N. The control engineer must have a basic understanding of any process before an effective control system can be developed.
Most readers with chemical engineering backgrounds will be familiar with this material and can skip some sections of this chapter. Fundamental Objectives Distillation columns are very widely used in the chemical and petroleum industries to separate chemical components into more or less pure product streams.
This separation is based on differences in volatilities tendencies to vaporize among various chemical components. For example, a mixture of methanol and water can be separated by distillation because methanol is more volatile or boils at a lower temperature than water. In a distillation column, the more volatile, or lighter, components are removed from the top of the column, and the less volatile, or heavier, components are removed from the lower part of the column.
Nomenclature Figure 2. At this point we will consider only a simple single-feed, two-product column separating a binary two-component mixture. Feed rate is F mols per minute. Feed composition is zF mol fraction of the more volatile component. The column trays are numbered fi-om the base upward, with feed introduced on the NE tray. The total number of trays in the column is NT.
Products removed fi-om the top and bottom of the column are called distillate or top product, and bottoms or bottom product, respectively, with flow Heat is transferred into the process in the reboiler n7pically a tube-andshell heat exchanger to vaporize some of the liquid from the base of the column. The vapor coming from the top of the column is liquified in another tubeand-shell heat exchanger called a condenser. Heat is transferred out of the condenser at a rate g. Liquid from the condenser drops into the reflux drum. Distillate product is removed from this drum.
This liquid reflux and the vapor boilup in the base of the column are necessanr to achieve the separation or fractionation of chemical components. Overall View f r o m a Control Perspective One can stand back and look at a distillation column with its associated reboiler, condenser, and reflux drum as a black box process.
Feed, heat, and reflux are inputs into this box see Figure 2. Outputs from the box are the two product streams D and B with compositions x,, and x ,. The usual situation with a distillation column is that the feed rate and feed composition must be. Heat input q R and external reflux Lacan be adjusted to achieve the desired control objectives.
These are called manipulated variables. Distillate and bottom product rates can also be manipulated, so there are four variables that can be adjusted: La, q R , D, and B. They are not independent, however. If, for example, D is controlled, then B is dependent. Fundamental Manipulated Variables Two of the four manipulated variables listed above must be used to maintain liquid inventories in the reflux drum and in the column base.
Therefore, we are left with just two manipulated variables that can be used to control compositions in the column. No matter what manipulated variables are chosen to control what controlled variables, there are basically two fundamental manipulated variables that affect compositions. These are feed split and fractionation. Feed split means the fraction of the feed removed as either distillate or bottom product. The steady-state effectiveness of both the Qrect and indirect schemes is identical.
In either case feed split has a very strong effect on product composition. A slight change in feed split can change product compositions very drastically, particularly when product purities are high.
Fractionation also affects product composition. Fractionation means the degree of separation. It varies with the number of trays in the column, the energy input to the reboiler, and the intrinsic difficulty of separating the components. For a fixed column operating at a fixed pressure with given chemical components, heat input to the reboiler is the only variable that can be used. Heat input can be used directly; alternatively reflux can be adjusted, if this is more convenient, since reflux and heat input are tied together through overall energy and mass balances.
We will discuss the pros and cons of various choices of control schemes in more detail in later chapters. These trays promote mass transfer of light components into the vapor flowing up the column and of heavy components into the liquid flowing down the column. Vapor-liquid contacting is achieved by a variety of devices.
The most widely used trays in recent years have been sieve trays and valve trays because of their simplicity and low cost. Sieve trays are simple flat plates with a large number of small holes. Vapor flows up through the holes, preventing the liquid from falling through. See Figure 2. Valve trays are built with a cap that fits over the hole in the tray and that can move up and down, providing more or less effective hole areas as vapor flow rate changes.
This fairly complex process of flow of vapor up the column and of liquid across each tray and down the column is called tray hydraulics. It is important in control system design because it imposes very important constraints on the range of permissible liquid and vapor flow rates. If liquid cannot flow dowh the column, or if vapor-liquid contacting is poor, the separating ability of the column drops drastically. Vapor flows from one tray up through the tray above it because the pressure is lower on the upper tray. Thus there is an increme in pressure from the top of the column to its base.
Liquid must flow against this positive pressure gradient. It is able to do so because the liquid phase is denser than the vapor phase. A liquid level is built up in the downcomer to a height sufficient to overcome the difference in static pressure between the tray onto which the liquid is flowing and the trav from which it is coming. This pressure difference depends on the vapor pressure drop through the tray which varies with vapor velocity, number and size of holes, vapor density, etc. This is usually due to excessive boilup vapor rate but sometimes may be caused by excessive reflux.
The control system must keep the column from flooding. Therefore, there are maximum vapor and liquid rates. On the other end of the scale, if vapor rates are reduced too much, the vapor pressure drop through the openings in the tray will be too small to keep the liquid from weeping or dumping down through the ho1es. The same thing occurs if liquid rates are so low as they often are in vacuum columns that it becomes difficult to hold enough liquid on the tray to get good vapor-liquid contacting.
These hydraulic constraints can be handled in control system design by using maximum and minimum flow limiters on heat input and reflux. A measurement of column pressure drop can also be used to prevent flooding. These concentration differences are analyzed and quantified using basic thermodynamic principles covering phase equilibrium. Vapor-liquid equilibrium VLE data and analysis are vital components of distillation design and operation. Vapor Pressure The liquid phase of any pure chemical component, speciesj , exerts a certain pressure at a given temperature. This pressure is called the pure component vapor pressure P,.
It is a physical property of each component. Vapor-pressure data are obtained by laboratory experiments where both liquid and vapor phases of a pure component are held in a container see Figure 2. Pressure is measured at various temperatures. The temperature at which the pure component exerts a pressure of one atmosphere is called its normal boiling point. Light components have low normal boiling points and heavy components have high normal boiling points. When the data are plotted on linear coordinates see Figure 2.
Therefore, vapor-pressure data are usually plotted using coordinates of log pressure versus reciprocal of absolute temperature as illustrated in Figure 2. Note that the constantsA, and BI must be determined for each pure component. They can be easily calculated by knowing two vapor-pressure points P, at T1 and P2 at T 2. When working with a distillation column, T1and T2 are usually selected to be near the temperatures at the top and at the bottom of the column.
Then temperature and pressure are measured. Samples of the vapor phase and liquid phase are taken and analyzed. Liquid compositions are usually expressed as mol fiaaion of light component and the symbol x is used. Vapor composition is expressed as mol fraction light component, using the symboly. Mixture composition is then changed and the procedure repeated. These results are conveniently presented in graphic form using several types of phase diagrams.
If the data are taken with pressure held constant isobaric , it is convenient to plot two curves on the same paper: temperature versus x and temperature versus y. Figure 2. To determine the composition of liquid and vapor phases in equdibrium with each other at a given temperature, and at the pressure under which the data were obtained, one merely draws a horizontal line at the given temperature and reads off x and y values.
P-xy Diagrams Data taken at constant temperature isothermal are plotted as two curves: pressure versus x and pressure versus y, as illustrated in Figure 2. Either isothermal or isobaric data can be represented by simply plotting liquid composition x versus vapor composition y. This x versus y curve see Figure 2. This type of diagram is the most widely used on distillation. Both T-xyand P-xy are u s e l l in illustrating the concepts of bubble point, dew point, superheated vapor, and subcooled liquid.
Consider the P-xy diagrams sketched in Figure 2. Suppose we have a mixture with composition z, and we hold it at the same temperature T for which the diagram was drawn. If we impose a very high pressure on the mixture, we will be above the x versus P curve saturated-liquid line and no vapor will be present.
There will be only. At this high pressure, the liquid is called ccsubcooled. If we now begin to drop the pressure, we will move down the vertical line drawn through composition 2, shown in Figure 2. When pressure reaches the point labeled PBp,vapor will begin to appear. This therefore is c d e d the bubble-point pressure of this mixture of composition z and at temperature 7. The composition of this first bubble can be read off the y versus P curve by moving across horizontally at PBp. As pressure is reduced further, more and more vapor is formed. Finally, at a pressure Popd the liquid has vaporized.
This is called the dew-point pressure of this mixture of composition z and at temperature T. The y versus P line is called the saturated-vapor line. At pressures below Pop only a single phase exists, superheated vapor. The same concepts can be visualized using constant-pressure T-xy diagrams Figure 2. The mixture is superheated vapor at temperatures above the dew point TDp and subcooled liquid at temperatures below the bubble point TBp. Note that we can talk about either bubble-point temperature or bubblepoint pressure, depending on which variable is fixed isothermal or isobaric situations.
The same is true for dew-point temperature and dew-point pressure. These calculations are called bubble-point, dew-point, and flash calculations. Thermodynamic Basis The second law of thermodynamics tells us that the chemical potential of each component must be equal in both liquid and vapor phases at phase equilibrium. Koppel, L. Shinskey, F. Rademaker, O. Rijnsdorp, and Douglas, J. Control 2vols. Gould, L. McAvoy, T. Stephanopolous, G. Ray, W. Cliffs, N. Nisenfeld, E.
Seemann, Murrill, P. The control engineer must have a basic understanding of any process before an effective control system can be developed. Most readers with chemical engineering backgrounds will be familiar with this material and can skip some sections of this chapter. Fundamental Objectives Distillation columns are very widely used in the chemical and petroleum industries to separate chemical components into more or less pure product streams.
For example, a mixture of methanol and water can be separated by distillation because methanol is more volatile or boils at a lower temperature than water. In a distillation column, the more volatile, or lighter, components are removed from the top of the column, and the less volatile, or heavier, components are removed from the lower part of the column. Nomenclature Figure 2. At this point we will consider only a simple single-feed, two-product column separating a binary two-component mixture.
Feed rate is F mols per minute. Feed composition is zF mol fraction of the more volatile component. The column trays are numbered fi-om the base upward, with feed introduced on the NE tray. The total number of trays in the column is NT. FlGURE 2. The vapor coming from the top of the column is liquified in another tube- and-shell heat exchanger called a condenser. Heat is transferred out of the condenser at a rate g. Liquid from the condenser drops into the reflux drum.
Distillate product is removed from this drum. Feed, heat, and reflux are inputs into this box see Figure 2. Outputs from the box are the two product streams D and B with compositions x,, and x,. The usual situation with a distillation column is that the feed rate and feed composition must be. Heat input q R and external reflux Lacan be adjusted to achieve the desired control objectives. These are called manipulated variables. Distillate and bottom product rates can also be manipulated, so there are four variables that can be adjusted: La, q R , D, and B.
They are not independent, however. If, for example, D is controlled, then B is dependent. Fundamental Manipulated Variables Two of the four manipulated variables listed above must be used to maintain liquid inventories in the reflux drum and in the column base. Therefore, we are left with just two manipulated variables that can be used to control compositions in the column. The steady-state effectiveness of both the Qrect and indirect schemes is identical. In either case feed split has a very strong effect on product composition.
A slight change in feed split can change product compositions very drastically, particularly when product purities are high. Fractionation also affects product composition. Fractionation means the degree of separation. It varies with the number of trays in the column, the energy input to the reboiler, and the intrinsic difficulty of separating the com- ponents. For a fixed column operating at a fixed pressure with given chemical components, heat input to the reboiler is the only variable that can be used.
Heat input can be used directly; alternatively reflux can be adjusted, if this is more convenient, since reflux and heat input are tied together through overall energy and mass balances. We will discuss the pros and cons of various choices of control schemes in more detail in later chapters. These trays promote mass transfer of light components into the vapor flowing up the column and of heavy components into the liquid flowing down the column.
Vapor-liquid contacting is achieved by a variety of devices. The most widely used trays in recent years have been sieve trays and valve trays because of their simplicity and low cost. Sieve trays are simple flat plates with a large number of small holes. Vapor flows up through the holes, preventing the liquid from falling through. See Figure 2. Valve trays are built with a cap that fits over the hole in the tray and that can move up and down, providing more or less effective hole areas as vapor flow rate changes. If liquid cannot flow dowh the column, or if vapor-liquid contacting is poor, the separating ability of the column drops drastically.
Vapor flows from one tray up through the tray above it because the pressure is lower on the upper tray. Thus there is an increme in pressure from the top of the column to its base. Liquid must flow against this positive pressure gradient. It is able to do so because the liquid phase is denser than the vapor phase. A liquid level is built up in the downcomer to a height sufficient to overcome the difference in static pressure between the tray onto which the liquid is flowing and the trav from which it is coming.
This pressure difference depends on the vapor pressure drop through the tray which varies with vapor velocity, number and size of holes, vapor density, etc. This is usually due to excessive boilup vapor rate but sometimes may be caused by excessive reflux. The control system must keep the column from flooding.
Therefore, there are maximum vapor and liquid rates. On the other end of the scale, if vapor rates are reduced too much, the vapor pressure drop through the openings in the tray will be too small to keep the liquid from weeping or dumping down through the ho1es. The same thing occurs if liquid rates are so low as they often are in vacuum columns that it becomes difficult to hold enough liquid on the tray to get good vapor-liquid contacting. These hydraulic constraints can be handled in control system design by using maximum and minimum flow limiters on heat input and reflux.
A mea- surement of column pressure drop can also be used to prevent flooding. These concentration differences are analyzed and quantified using basic thermodynamic principles covering phase equilibrium. Vapor-liquid equilibrium VLE data and analysis are vital components of distillation design and operation. Vapor Pressure The liquid phase of any pure chemical component, speciesj , exerts a certain pressure at a given temperature.
It is a physical property of each component. Vapor-pressure data are obtained by laboratory experiments where both liquid and vapor phases of a pure component are held in a container see Figure 2. Pressure is measured at various temperatures. When the data are plotted on linear coordinates see Figure 2. Note that the constantsA, and BI must be determined for each pure component.
They can be easily calculated by knowing two vapor-pressure points P,at T1 and P2 at T 2. When working with a distillation column, T1and T2 are usually selected to be near the temperatures at the top and at the bottom of the column. Then temperature and pressure are measured. Samples of the vapor phase and liquid phase are taken and analyzed. Liquid compositions are usually expressed as mol fiaaion of light component and the symbol x is used.
Vapor composition is expressed as mol fraction light component, using the symboly. Mixture composition is then changed and the procedure repeated. S Temperature vs. T-xy Diagrams If the data are taken with pressure held constant isobaric , it is convenient to plot two curves on the same paper: temperature versus x and temperature versus y. Figure 2. To determine the composition of liquid and vapor phases in equdibrium with each other at a given temperature, and at the pressure under which the data were obtained, one merely draws a horizontal line at the given temperature and reads off x and y values.
P-xy Diagrams Data taken at constant temperature isothermal are plotted as two curves: pressure versus x and pressure versus y, as illustrated in Figure 2. This x versus y curve see Figure 2. Both T-xyand P-xy are u s e l l in illustrating the concepts of bubble point, dew point, superheated vapor, and subcooled liquid.
Consider the P-xy diagrams sketched in Figure 2. Suppose we have a mixture with composition z, and we hold it at the same temperature T for which the diagram was drawn.
If we impose a very high pressure on the mixture, we will be above the x versus P curve saturated-liquid line and no vapor will be present. There will be only. At this high pressure, the liquid is called ccsubcooled. When pressure reaches the point labeled PBp,vapor will begin to appear. The composition of this first bubble can be read off the y versus P curve by moving across horizontally at PBp. As pressure is reduced further, more and more vapor is formed. Finally, at a pressure Popd the liquid has vaporized. This is called the dew-point pressure of this mixture of composition z and at temperature T.
The y versus P line is called the saturated-vapor line. The same concepts can be visualized using constant-pressure T-xy diagrams Figure 2. The mixture is superheated vapor at temperatures above the dew point TDpand subcooled liquid at temperatures below the bubble point TBp. Note that we can talk about either bubble-point temperature or bubble- point pressure, depending on which variable is fixed isothermal or isobaric situations.
The same is true for dew-point temperature and dew-point pressure. These calculations are called bubble-point, dew-point, and flash calculations. Thermodynamic Basis A. The second law of thermodynamics tells us that the chemical potential of each component must be equal in both liquid and vapor phases at phase equilibrium. If the components are chemically quite s d a r , there is little attraction or repulsion of neighboring molecules of different types. Nonideality will be discussed in more detail in Section 2. We will assume ideal VLE behavior for the rest of this section for purposes of simplicity.
The K value K, of thejth component is defined as the ratio of vapor composition y t to liquid composition x,. In addition we must be given either the pressure or the temperature of the svstem. Bubble-Point Temperuture Cakulutwn. This is by far the most common type of calculation encountered in distillation work because column pressure is usually known. The calculation procedure is iterative: 1. Guess a temperature T.
Calculate vapor pressures of all components at T. Check to see if is sufficiently close to PT. If not, reguess T and go back to step 2. When convergence has been achieved, calculate vapor compositions. Notice the enriching of lighter component that occurs in the vapor in the above example. Benzene, the lightest component, has a higher concentration in the vapor than in the liquid. This illustrates precisely why a distillation column can be used to separate chemical components. The vapor rising in the column gets richer and richer in light components at each stage.
The liquid moving down the column gets richer and richer in heavy components. Bubble-Point Pressure Calculation. In this case temperature T and liquid- phase composition are known. Total system pressure is easily calculated with no iteration involved from:. Vapor pressures Pj are known since temperature is given. Dew-Point Calculations C. Dew-Point Temperature PGiven. Calculate pressure directly fiom equation 2. Isothermal Flash Calculations These calculations combine vapor-liquid equilibrium relationships with total mass and component balances. Both the temperature and the pressure in the drum are given.
Variables that are unknown are liquid and vapor compositions and liquid and vapor flow rates. Therefore, another iterative, trial- and-error solution is required. The lek-hand side of equation 2. Before starting any flash calculation, it is vital that one checks to see that the pressure and temperature gven are such that the feed mixture is in the two-phase region. That is, the system pressure must be between the bubble-point pressure and the dew-point pressure for a m i x t u r e with a composition equal to the feed composition and at the given temperature T.
Relative Volatility Relative volatility is a very convenient measure of the ease or difficulty of separation in distillation. The volatility of component j relative to component k is defined as:. Values of ajkclose to 1 imply that the separation will be very difficult, requiring a large number of trays and high energy consumption. For binary systems relative volatility of light to heavy component is simply called a:. Rearrangement of equation 2. The larger the relative volatility a, the fatter is the equilibrium curve. In other words, relative volatility is. This is true for many components over a limited temperature range, particularly when the components are chemically similar.
Distillation columns are frequently designed assuming constant relative volatility because it greatly simplifies the vapor-liquid equilibrium calculations. Relative volatilities usually decrease somewhat with increasing temperature in most systems. For multicomponent systems, applying the basic definition [equation 2.
Given the liquid compositions and relative volatilities, calculate the vapor compositions: -5 ffj xJ ffl YJ 0. Nonideality In most distillation systems, the predominant nonideality occurs in the liquid phase because of molecular interactions. Equation 2. When chemically dissimilar components are mixed together for example, oil molecules and water molecules , there can be repulsion or attraction between dissimilar molecules.
If the molecules repel each other, they exert a higher partial pressure than if they were ideal. If the molecules attract each other, they exert a lower partial pressure than if they were ideal. Activity coefficients are less than unity negative deviations. Activity coefficients are usually calculated from experimental data. Empirical equations Van Laar, Margules, Wilson, etc. Azeotropes occur in a number of nonideal systems. There are several types of azeotropes. Figures 2. Negative deviations attraction can give a higher temperature boiling mixture than the boiling point of the heavier component.
Positive deviations repulsion can give a lower temperature boiling mixture than the boiling point of the light component. A modest amount of repulsion can lead to the formation of a minimum boiling azeotrope Figure 2. If the repulsion is very strong, the system may break into two-liquid phases with different compositions in each liquid phase. A detailed discussion of azeotropes is beyond the scope of this brief intro- duction. The control engineer should be aware that the existence of azeotropes imposes restrictions on the operation and performance of a distillation column.
These techniques are very useful in gaining an appreciation of the effects of various design and operating parameters. System Figure 2. Liquid and vapor rates below the feed in the stripping section are called Ls and V,. These are assumed to be constant. In many systems this is a pretty good assumption. A total condenser is used to produce liquid reflux and distillate product.
The composition of this liquid pool is the same as the bottom product composition. Thermosyphon, kettle, internal, and forced-circulation reboilers are all usually partial reboilers. The parameter q will be used to describe the thermal condition of the feed. The ratio of the internal reflux flow rate L R to the distillate flow rate D is called the reflux ratio R.
Equations Overall Balances A. Mass and component balances can be written around the entire column system. Note that distillate and bottom product rates can be calculated from equations 2. Stripping Section A light component balance around the nth tray in the stripping section yields see Figure 2. This is called the operating-line equation. The slope of the line is the ratio of liquid to vapor flow rates in the stripping section. This straight line can be plotted on an x y diagram see Figure 2.
Substituting into equation 2. From a total mass balance around the system in Figure 2. Rectifying Section A similar component balance around the upper part of the column above the nth tray in the rectifjmg section see Figure 2. It intersects the 45" line at X, and has a slope equal to the ratio of the liquid- to-vapor rates in the recufylng section. Stepping off Trays Tray-to-tray calculations involve the solution of vapor-liquid equilibrium relationships and component balances.
Our VLE relationship gives us y, if we know x E. This is the stripping operating line [equation 2. We know y,. We could plug into equation 2. Alternatively, we can solve graphically for x1 simply by moving horizontally on a straight line through y, until the operating line is intersected. Then this. This stepping procedure is continued up through the stripping section until the intersection of the operating lines is passed see Figure 2.
This determines the number of trays required in the stripping section. Then the rectifying operating line is used, and the stepping is continued until x, is reached. The number of trays in the rectifying section can be determined in this manner. Thus we can design a column Le. As we will show in the next section, the operating line slopes are both known if the feed thermal condition and reflux ratio have been specified. Feed Thermal Condition The feed to a distillation column can be liquid or vapor, or both, depending on the temperature, pressure, and composition of the feed.
To quanti the thermal condition of the feed, the parameter q is defined as the fraction of the feed that is liquid. If the feed is a vapor-liquid mixture, q is a fraction. Values of q greater than 1 indicate subcooled liquid feed. Values of q less than 0 indicate superheated vapor feed. Now let us look at the intersection points x f , yi of the stripping and rectifying operating lines. Use of equations 2. It is called the q line and intersects the 45"line at 2,. Thus the intersection of the operating lines must lie on the q line, which can be easily drawn given zFand q.
Thus the rectiqing operating line can be drawn if xD and R are specified. Then the smpping operating line can be drawn if xB and q are specified. It is a straight line joining xBon the 45"line with the intersection of the rectlfylng operating line and the q line. Design Problem We are now ready to summarize the graphical design technique for deter- mining the number of trays required to achieve desired product purities, given a reflux ratio. As we will show later, the lower the reflux ratio specified, the more trays are required.
This will be discussed fUrther under "Limiting Conditions. Feed: zF,F, q 2. Desired product purities: xD, XB 3. VLE curve in x-y coordinates 4. Calculate: 1. Total number of trays: NT 2. Feed tray location: NF Procedure: 1. Draw VLE curve. Draw 45" line. Locate xE,xD, and ZF on 45" line. Calculate B and D fi-om overall balances [equations 2. Calculate liquid and vapor flow rates in re-g and stripping sections. The first step corresponds to the partial reboiler. The next step is tray 1, the next is tray 2, and so on. Thus the feed tray NF has been determined.
Switch to the rectifying operating line and continue stepping. When the xD point is crossed, this is the total number of trays NT. This last step will not go through the xD point exactly, implying a noninteger number of trays. Actual industrial columns seldom achieve this ideal situation, so an efficiency factor must be used to determine the number of actual trays installed in the column which must be an integer number. Typical efficiencies run from 40 to 90 percent, depending on the system. Rating Problems The graphical McCabe-Thiele methods studied in the previous sections for the design of distillation columns are also widely used to analyze the operation of an existing column.
In this case the total number of trays in the column NT is fixed. The feed tray may also be fixed, or, if there are multiple feed points available on the column, it may be varied. There are a variety of possible rating problems. Basically one guesses a solution and sees if the stepping procedure produces exactly the same number of trays in each section as has been specified see Figure 2. Notice that in both of these problems, two variables must be specified to define the system completely. This magic number of two occurs again and again in distillation see Section 4. Mathematically the two degrees of freedom are the result of subtracting all the constraining equations describing the system mass, component, and energy balances; VLE equilibrium relationships; and specified variables from the total number of system variables.
Limiting Conditions McCabe-Thiele diagrams are useful for getting a clear picnrre of some of the limiting conditions on the separation that can be achieved in a distillation column. Minimum Reflux Ratio The minimum reflux ratio for specified product purities and feed conditions occurs when an infinite number of trays are required to make the separation. It occurs when the operating lines just intersect on the VLE curve. The actual reflux ratio used must be higher than the minimum. Increasing reflux ratio requires fewer trays less capital cost but increases energy costs.
Economic optimization studies have led to the commonly used heuristic rule of thumb that the optimum actual reflux ratio is 1. In some unusual VLE systems, the pinch between the VLE curve and an operating line can occur at some point other than the feed point. Minimum Number of Trays The minimum number of trays to make a specified separation is found when an infinitely large reflux ratio is used.
This situation actually takes place in a column when it is operated under "total reflux" conditions. No feed is introduced and no products are withdrawn, but heat is added in the reboiler and all the overhead vapor is condensed and returned to the column as liquid reflux. The L and H subscripts refer to light and heavy components.
Design Case A. The minimum reflux ratio is increased somewhat, but this increase is only very gradual since the upper end of the rect ing operating h e merely approaches closer and closer to the 1, 1 point, which changes the operating line slope only slightly at high purities. Increasing Relative Volatility The number of trays is reduced. Minimum reflux ratio is also reduced. Relative volatility has a strong effect on the cost of separation. Therefore, columns are usually designed to operate at the lowest economical pressure, since lower pressure means lower temperatures and higher relative volatilities in most systems.
The lowest economical pressure is usually the pressure that provides a temperature in the overhead condenser that is high enough to permit cooling water or air to be used for heat removal. Going to lower pressures would require refiigeration, which is very expensive. Increasing Feed q Reflux ratio is reduced as the feed is made colder as q increases but energy input to the reboiler increases. Therefore, higher or lower feed compositions should require less energy. If two or more feed streams with different compositions are to be separated in the column, they should not be mixed and fed in at a single feed point.
Instead they should be fed on separate feed trays at locations where tray com- positions approximate feed compositions. Increasing product purities increases reflux ratio. Increasingfeed q reduces reflux ratio condenser load but increases heat input. Changingfeed mposition changes reflux ratio and energy input but not in the same way for all columns. The effect depends on product purities and relative volatilities see reference 9. Redzuzngpressure usually reduces reflux ratio and energy consumption if product purities are kept constant.
McGraw-Hill, New York, Billet, R. Treybal, R. Kirschbaum, E. Robinson, C. Smith, B. Overhead System 3 Arrangements.
Design of Distillation Column Control Systems by P. Buckley, J. Shunta, W. Luyben - upahucypidig.ml
The design of a satisfactory distillation control system involves far more than theory or mathematics. The engineer must have some idea of what constitutes effective equipment configurations and arrangements, as well as an appreciation of equipment performance limitations, and must be able to recogme when undesirable side effects are apt to interfere with an otherwise good control system.
Typical equipment, control schemes, problems, and solutions are discussed in this section. The supporting mathematics and theory are covered in Part The column overhead system is generally more complicated than either the feed system or the bottoms system. It usually must condense most of the vapor flow fiom the top tray, remove inerts, provide reflux flow back to the column, maintain column pressure in the right range, and satisfy part of the column material balance requirements. Condensate is generally subcooled at least slightly, partly to minimize the likelihood of flashing and cavitation in valves and pumps, partly to control the amount of inerts in the system, and partly to control product losses through the vent.
If subcooling is required for a pressure or vacuum column, the preferred method is to have the condensate-temperature controller manipulate the vent flow in some way. This arrangement avoids the instabilities and other control difKculties that often characterize condensate-temperaturecontrol systems based on manipulation of the condenser cooling water. In the case of vacuum columns, there may be problems associated with the control of the vacuum jets, and column turndown is usually limited, compared with that of atmospheric or pressurized columns.
Material balance control on the condensate may be accomplished in several ways :. If a smooth flow to the next step in the process is needed, a reflux drum, with averaging level control of distillate, should be employed. As an alternative, vapor may be taken off on flow control cascaded fiom top composition control while column pressure is controlled by heat input. Although it is less exact, this term is widely used in the petroleum industry. Horizontal shell-and-tube condenser with liquid coolant in the tubes and vapor on the shell side Figure 3.
This is probably the most popular type in petroleum refineries. In addition, at startup time, column inerts are usually vented more easily ix. The design itlustrated in Figure 3. Some designs bring the vapor in at one end and vent uncondensables at the other. Sometimes condensate is taken out through two drawoffs instead of one.
The cooling water valve is normally at the exchanger exit to make sure the tubes are fdled at all times. Since the exit water is hot, the valve may need anticavitation trim. This type is popular in the chemical industry because it m i n i m i z e s condenser cost when highly corrosive process materials must be handled. With a longer condensing path, it is also better suited to applications in which it is desired to absorb the maximum amount of low boilers in the condensate.
This condenser commonly has at its lower end a vapor-liquid disengaging pot, which also serves as a condensate receiver. Because all vapors must pass through the tubes, the speed of venting inerts at startup time is limited. Here we have a number of different designs. Reflux may return internally via an overflow weir, or externally through a gravity flow line with a control valve. Here condensate is recirculated through. This type of condenser is most commonly used in vacuum service because of its low pressure drop. At startup time total reflux may be achieved by using the reflux valve to control the level in the condensate receiver.
For those columns that must be protected from atmospheric oxygen or moisture, a vent system such as that shown in Figure 3. Note that inerts usually should be added a er the condenser, to minimize product losses. Sometimes, however, it is necessary to add inerts ahead of the condenser, for pressure control. Figure 3.
A potential and frequent source of trouble with both arrangements is the control of condensate temperature via cooling water. As shown by a study by B. This compounds stability problems; we need an increasing controller gain and decreasing reset time as total heat load increases. Further, subcooling heat load must be a reasonable fraction of total heat load-say 5 percent-or the system will lack adequate sensitivity. Finally, many cooling water valves do not: have adequate turndown; they are wide open in summer and almost closed in midwinter.
A small and a large valve in parallel should often be used. The suggested approaches to a v o i h g these ddKculties are as follows: 1. Select the number of degrees of subcooling so that the sensible heat load will be at least 5 percent of the total heat load. This will have a secondary advantage of reducing the probability of cavitation in control valves and pumps. If water-header pressure fluctuations are a problem, use a cascade tem- perature water-flow control system.
- Genetically Engineered Food: Methods and Detection.
- Design of Distillation Column Control Systems - Knovel?
If summer-winter heat-load variations are sufKciently severe, use dual, split-range water valves. The smaller valve should open first and will provide adequate winter cooling. The ratio of cooling water rate for the maximum summer heat load to that for the minimum winter load is often two to three times as great as process turndown.
See also discussion in Chapter 11, Section 6. For the horizontal condenser, the temperature detector preferably should be located in the liquid line just beneath the condenser for maximum speed of response. For the vemcal condenser, the temperature detector should be located in a trough at the lower end of a drip collector just below the tube bundle and above the reflux drum. At least one, or preferably both, of bottoms off take b and distillate d should be controlled by either a symptom of imbalance or composition.
The composition link should only be considered if a sloppy split is adequate. This generally provides poor control due to excessive dead time and process delay before changes in flow can affect the level. All other flows in the column should also be adjusted in the same ratio to maintain separation conditions. This form of control should be avoided if possible. An example of how to design a feedback control system for a particular arrangement is given in 5.
An exception to this general rule may be when, for example, reflux drum level is regulating reflux and it is important to increase reflux to match any increase in reboil rate. The reasons for setting up such a form of control is discussed later. In all other cases the level control system should be tuned to maximize the use of surge capacity. This means that a proportional only controller should be used tuned so that the capacity never empties or overfills. Sometimes proportional error 2 control is suggested for such applications so that little control action takes place at the mid measurement point but more drastic action takes place as the measurement reaches its limits.
In the experience of the author properly tuned proportional only control is normally acceptable. It is simpler and is better understood. There are no critical flows, so, from Table 2 first make a choice of composition regulation. This is fairly evidently reboil h and the other controls can be built up for the reasons given in Table 3. In these circumstances it is normal to flow control reflux. This is discussed later. We have not considered the type of condensation system provided.
If, for example, it were a water cooled condenser it is quite possible that no regulation would be available because the necessary valves are likely to be very large and cooling water is relatively inexpensive.
- Direct and Large-Eddy Simulation I: Selected papers from the First ERCOFTAC Workshop on Direct and Large-Eddy Simulation.
- Compact Hierarchical Bipolar Transistor Modeling With Hicum (International Series on Advances in Solid State Electronics and Technology).
- G Protein-Coupled Receptor Screening Assays: Methods and Protocols.
How should pressure be controlled? In this circumstance there might be a temptation to control pressure by regulating reflux flow. But then there would be no independent way of setting the energy input to the column and, if the pressure and temperatures are mismatched as they always will be , the reboil and reflux would ramp up or down until a balance is found. This might be maximum or minimum energy input.
In these circumstances the correct action is probably to let the pressure float to achieve the lowest possible value at any time. The problems of temperature control with the variable pressure are covered in 5. The reflux should be flow controlled or ratioed to feed.
Then the reflux flow and its temperature sets the heat input to the column to balance the reflux flow to maintain steady compositions. Carrying out a similar calculation to previously, it can be shown that the ratio of reboil to bottoms off-take is In this circumstance, it is very difficult to maintain a bottoms mass balance by regulating bottoms off-take; reboil has a much greater effect.
In cases like this it is not uncommon to regulate the bottoms level by adjusting reboil and to use the composition measurement, particularly if it is close to the bottom of the column, to regulate the bottoms off-take. Similar arguments apply to the control of distillate and reflux from a temperature control near the top of the column and the reflux drum level. Consider the following table. In this case the bottoms mass balance is the critical parameter to be measured so consider it first.
TABLE 4 5. Overhead condensers may be either total condensers where the condensable part of overhead vapor should ideally be completely condensed leaving only the inerts to be purged from a suitable point in the condensation system. Alternatively, they may be partial condensers where only a fraction of the condensable components are condensed, the remainder passing out with the inerts as a vapor top product.
Condensation control methods usually vary one or more of the first four of these quantities. They are considered in turn. Typical condensation arrangements are shown in Figures 9 a to 9 c. In Figures 9 b and 9 c the vapor top off-take may not be present if condensation is total but there is always likely to be some means of manually venting inerts which have built up in the system. If it is required to reduce the rate of condensation, the control valve closes and the fraction of non-condensables in the condenser will then build up.
The particular system shown can only be used with columns operating above atmospheric pressure but alternatives using the same principle can be used when operating at below atmospheric pressure. This is considered later. Strictly speaking, adjustment of the vapor purge valve also varies the overall heat transfer coefficient as well as the partial pressure of the condensables since the vapor side heat transfer coefficient is dependent on the fraction of the non-condensables in the vapor. At high flow rates the average temperature of the coolant is reduced so the rate of condensation is increased.
Sometimes boiling refrigerant is used as the cooling medium, absorbing the latent heat of condensation as its own latent heat of vaporization. This can be used for condensation control see 5. Manipulation of the coolant flow is rarely used to adjust heat transfer in water cooled condensers on distillation columns. Response of the condensing rate to water-flow variations is non-linear and slow.
The speed of response also varies with coolant flow which can cause loop stability problems. Low velocity and also leads to high fouling and metal corrosion. For more details see Ref. In this case only the part of the surface which is not submerged is available for condensation of vapor and the liquid will leave the condenser sub-cooled.
The basic condensation control systems differ in dynamic performance. For example, the speed of operation of the system shown in Figure 9 a will depend entirely upon the proportion of inerts present in the overhead vapor from the column. If the proportion is very small then, even if the vent valve is closed, it could take a long time to build up the sufficient proportion of inerts in the condensation system to affect heat transfer.
Sometimes a split range control system is provided where nitrogen can be added to the system to speed up the response. The speed of operation of the system shown in Figure 9 b depends very much on the thermal capacity of the tube wall separating the hot and cold sides of the condenser and on the residence time of the cooling medium when this is a cold liquid. The system in Figure 9 c has slow dynamics.
Movement of the control valve affects the rate of change of the flooding depth in the condenser and this in turn affects the rate of change of condensation and hence pressure 5. The temperature of the condensing vapor tends to be constant if the tubes are not flooded with liquid and the following equation applies: The rate of heat transfer is limited by the air film and so the 0.
As a consequence, heat flow should be reasonably linear with air flow. Controlling air flow is another matter. Variable speed fans are rarely used because of the high cost of drive units. Most air coolers use multiple fans which can be energized in stages but this gives only incremental control.
Some fans are equipped with variable pitch blades or adjustable louvers but they have a tendency not to work satisfactorily. A further problem with air condensers is the effect of ambient conditions. Rain tends to convert the dry condenser into a wet condenser which can have a marked affect on the temperature of the condensate and hence the internal reflux in the column. In critical conditions this may require the provision of internal reflux control computation.
Adjustment of cooling in the air condenser can be effected by: a flooding with condensate, or, b bypassing. Most air cooled condensers are horizontal and hence flooding does not give smooth control. Sometimes the condenser is mounted at an angle and this helps but since the tubes are of large capacity and can contain a substantial amount of liquid, response can be slow.
Bypassing hot vapors around the condenser is shown in Figure Column or reflux drum pressure can be controlled by adjustment of the bypass valve. The sizing of the bypass valve is difficult but critical to the effective operation of the condensation system. In particular, it is necessary to decide what proportion of the vapor shall be bypassed under different operational circumstances, throughput and ambient conditions. For more information on air cooled heat exchangers see Ref.
There is then no reflux drum and the condensed liquid reflux falls directly from the condenser onto the top plate through some distributor system. Uncondensed vapor can be taken off directly as top product or an internal weir may be fitted, allowing the distillate to be withdrawn. This limits the options for control. Ideally, all condensate should be trapped and withdrawn for metering and control.
If this is not done the system behaves as if the reflux flow were regulated via a very high gain reflux drum level controller. Large and sudden changes in top off-take can cause corresponding changes in reflux and, in extreme situations, reflux flow may be lost for a period. In principle, this balances the rate of condensation to the overhead vapor flow by altering the proportion of inerts air present in the condensing vapor. Pressure control is rapid but varies with atmospheric pressure. When a column is one of a number in a distillation train and a partial condenser is used, then the vapor top product may form the vapor feed to the next column in the train.
If some method of pressure control is used on the second column, the first may be maintained at very nearly the same pressure as the second simply by ensuring that the line carrying the vapor offers little resistance to flow. Referring to the Table 1, this effectively means that the pressure is related to the vent rate v although physically there is no controller or control valve.
The purge line is often connected to the suction side of some form of vacuum pump or steam ejector. With this arrangement it is possible to use any of the basic condensation control systems but, since the proportion of inerts in the overhead vapor may well be high, a modification of the system shown in Figure 9 a is normally used. Three common arrangements are shown in Figures 11 a to 11 c. The temperature of the boiling refrigerant depends on the pressure maintained in the vapor space above it which may very easily be regulated by manipulating a control valve in the refrigerant exit line.
The refrigerant is often contained in the shell of a condenser as shown in Figure 12 and condensation of the process vapor takes place in the tubes. A constant refrigerant level is maintained by admitting a controlled flow of liquid refrigerant to the shell. The vapor condensation rate can be varied to a certain extent by altering the refrigerant level but the temperature of refrigerant boiling is usually used as the principal method of condensation control.
If the vapor velocity is low, as will be the case with a total condenser, it may not be easy to keep the tubes well drained. In this case, better heat transfer may be obtained by condensing the vapor in the shell with the refrigerant vaporizing in the tubes. A system of this type is shown in Figure The refrigerant is held in equilibrium with its vapor in a separate accumulator to which a control flow of refrigerant liquid is added as again, the temperature of this liquid is determined by the pressure in the vapor space above it.
The reflux drum is then sited at the same level as the condenser or even slightly above it. Liquid formed by condensation cannot fall into the reflux drum under gravity.
Design of Distillation Column Control Systems - p. Buckley
In practice, to avoid the need to pump liquid into the reflux drum, a pressure differential is maintained between condenser and reflux drum. The condenser works at almost the same pressure as the top of the column and the rate of condensation is varied by flooding part of the cooled surface with condensed liquid.
Its vapor pressure, which is the reflux drum pressure, is therefore lower than the pressure in the condenser. This pressure difference drives the condensed liquid into the reflux drum against the hydrostatic head due to the difference in level and the frictional losses in the piping. Pressure control is effected as shown in Figure 15 by bypassing some vapor from the column to the reflux drum. This alters the EP between condenser and drum and so alters the flow of condensate and thus the level in the condenser.
This alters the condensation rate and so the column pressure. If the bypass valve is adjusted to control the column pressure there is an inverse response which can cause instability. Controlling the reflux drum pressure by adjusting the bypass appears to be more satisfactory.
The different physical construction or heat sources can affect control. In either case the heating medium may be steam or liquid within the tubes. Natural circulation or "thermosyphon" reboilers may be mounted either horizontally or vertically. Vertical thermosyphon reboilers are used primarily in the chemical industry and are typically steam heated see Figure 17 a. Horizontal thermosyphon reboilers are more common in petroleum refining or similar operations in the chemical industry.
They are usually heated with circulating oil, see Figure 17 b. Forced circulation is used with vacuum distillation or when the heat input is obtained from oil or gas furnaces. In some low temperature applications a refrigerant is used. Superheated steam is a relatively poor heat transfer medium, as are most gases. Saturated steam is an excellent medium because of good heat transfer and a high latent heat of vaporization.
The boil-up rate may be controlled either by steam flow or shell pressure. The steam flow measurement is usually made upstream of the control valve where the pressure is normally constant. The steam rate to the reboiler may be controlled either by a valve in the steam line or in the condensate off-take. With the control valve on the inlet to the reboiler, the saturation pressure in the shell varies with heat load. Since the heat is being transferred between a condensing and a boiling fluid, neither changes temperature greatly in the process.
A steam trap or similar condensate seal is necessary to drain condensate without releasing steam.