Feedforward compensation

PID controller the classical PID regulator (5.1) with a feedforward compensation of the desired reactor temperature (i.e., ypdes is added to the control input). 

Remark. Notice that the gain scheduling is comparable to feedforward compensation. There is no feedback to compensate for incorrect adaptation. 

From previous experimental data we know how the optimal fuel/air ratio changes with air temperature for maximum efficiency. Therefore, to maintain the ratio continuously at its optimal value despite any changes in the air temperature, we can use a programmed adaptive control system. Such a system is shown in Figure 22.3b. It measures the temperature of the air (auxiliary measurement) and adjusts the value of the fuel/air ratio. Notice again that the ratio adjustment mechanism is like feedforward compensation. 

What is gain scheduling control, and why can you use it in chemical process control It was claimed in Example 22.1 that it resembles feedforward compensation. Explain why. What are its advantages and disadvantages ... 

In multistage machines, the antisurge scheme on the first stage will temporarily starve the next stage(s) of gas - possibly causing them to approach surge. Feedforward compensation may be required to avoid significant disturbances. 

The A s represent departures from average operating conditions. The constants Kfi Kfi may be cdculated approximately by the column designer. Luyben has shown that it is necessary to be quite careful in designing feedforward compensation for feed composition changes, particularly when the column is not making a sharp separation. 

It should be noted, too, that feedforward from feed composition may not be needed if the feed comes from a process step with discharge composition control. Feedforward compensation for other process variables, such as bottom product or distillate demand flow, is discussed in Chapter 6. 

For control loops represented in the CCR (central control room), it is normal practice (as mentioned earlier) to furnish control stations. 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, temperamre, etc.), the desired value (set point), and the valve loading signal (controller output, really). There is also a manual-automatic 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 by feedforward compensation, overrides, and so on, depending on the design philosophy for a particular project. 

Experience on many projects shows that even small, simple columns benefit from a modest application of feedforward compensation and overrides. This is due to the trend toward increasingly tight coliunn design. The number of trays is held down because smaller allowances are made ft>r uncertainty in tray efficiency. Column diameter and tray spacing are now kept to smaller values, with the... 

Suppose, however, that the customer insists on minimum application of feedback controls with no feedforward compensation or overrides. V at column design philosophy should be followed Ha Tng had considerable adverse experience with columns with primitive controls, particularly sidestream drawoff columns, and columns with heat recovery schemes, we suggest the following ... 

If composition is measured and is cascade controlled via reflux or boilup, or both, ratio controls should be replaced by impulse feedforward compensation (see Chapter 12) if feed flow turndown is greater than 2 1. 

Ls indicated in Chapter 1, it is most convenient, when starting the design of controls for a new or modernized plant, first to lay out all of the material-balance controls. These are mostly liquid level controls. It was also indicated that feedforward compensation could be used to supplement feedback composition controls to achieve more constant compositions. In the absence of feedback composition controls—usually because adequate composition measurements are lacking—feedforward compensation is almost mandatory. 

We will begin with combinations of level control and feedforward compensation for applications where material-balance control is in the direction opposite to flow. Then we will consider schemes in which material-balance control is in the direction of flow. Unfavorable schemes—those that are hard to design or to make work—will be pointed out their use should be avoided unless no suitable option is available. 

Since the bottom-product flow is the demand flow, steam and reflux are ratioed to it. For all ratio loops, appropriate dynamic feedforward compensators should be provided. 

In calculating reflux drum level controller settings, whether proportional only or PI, one should take care to account for reflux subcooling (see Chapter 16, Section 4). Overhead composition may be controlled by trimming the distillate/bottom-produrt ratio with perhaps a feedforward compensator connected into the overhead level control loop. Base composition may be controlled by trimming the steam/bottom-product ratio control. 

If turndown of more than 2 1 is anticipated, and if composition control is used, impulse feedforward compensation (see Chapter 12) fr>r feed flow will do a better job than ratio controls. 

In addition to the recovery of the latent heat of vapor streams, in many cases it is practical to recover part of the sensible heat in the column bottom product and steam condensate by exchange with column feed. Such schemes have been used in the chemical and petroleum industries for years. Since feed flow is typically set by level controllers or flow-ratio controllers, its flow rate will not be constant. The feed enthalpy or temperature, therefore, is apt to be variable. This may make column-composition control difficult unless one employs either feedforward compensation or a trim heater with control for constant temperature or enthalpy. (See Chapters 5 and 11.)... 

A convenient method for providing feedforward compensation that does not interfere with either norm reset or antireset windup involves the use of an impulse relay and a summer as shown in Figure 9.8. In pneumatics these fimctions are sometimes combined into a single device. The impulse function... 

Single loop with feedforward compensation, derivative. Pi controller, overrides, and predictor... 

Following is a brief review of the mathematics of a PI controller with external reset feedback and impulse feedforward compensation. The basic PI controller equations are ... 

Considering next the system of Figure 12.2 with feedforward compensation, let us write the equation for the summer output ... 

Cascade control probably has been used many times when feedforward compensation would provide superior control. For stability reasons the secondary Icxjp in a cascade system should be five to ten times faster than the primary Icxjp and should have dead-beat (nonresonant) tuning. But we increasingly cascade to flow controls that are usually so much faster than the primary loops that there are no stability problems. 

Use of the Smith and other predictors recently has been shown by Seborg et al. to provide substantially better feedback control of distillation-column terminal compositions than do PI and PID controllers. Further work needs to be done to compare their performance with that achieved by PI or PID controllers with feedforward compensation. 

Although not usually pointed out in the literature, it is also true, as Gould has indicated, that the engineer should be careful not to design a feedback system whose closed-loop natural frequency is within the frequency range of disturbances. Feedforward compensation often may be used instead of feedback to get improved control without problems with either stability or resonance peaks. 

Once a distillation column or train for a continuous chemical process has been designed and built, the three best opportunities for profitable operation via process control are usually (1) maximum-capacity operation if the plant is production limited, (2) minimum-cost operation if the plant is market limited, and (3) increased annual availability (sometimes called utility). The last of these is greatly aided by constraint controls, commonly called overrides, to keep the plant operating safely and smoothly when it mi t otherwise be shut down by interlodu or operator decision. Maximum-capacity or minimum-cost operation is facilitated by the use of on-line models, usually implemented by a digital computer. Fxmctionally these will be, in most cases, sophisticated feedforward compensators, or constraint controls. With these two limited exceptions, optimization is not considered in this book. Since the on-line models to be employed require calibration, some degree of on-line identification is highly desirable. 

Our use of feedforward compensation for distillation columns curroitiy is based almost entirely on feed-rate changes. Thus we have reflux-to-feed, steam-to-feed, and sometimes other ratio-control systems. But there are at least three other variables that can be important feed composition, feed enthalpy or q, and column pressure. 

The three great advantages accruing fi om the use of feedforward compensation are ... 

Column gains for feedforward compensation. For example, if feed composition changes, what changes in boilup and reflux do we need to hold terminal compositions constant Steady-state accuracy is required. 

Here the objective is to find required feedforward compensator gain to hold terminal composition constant as various external factors uy. 

For any of the control systems discussed—Figure 19.2, 19.3, or 19.4— assume that top composition and bottom composition are held constant. We wish to find the changes in Lk, D, B, and V, required to hold compositions constant in the face of a feed composition change, Azp. This information could be used to design feedforward compensators to minimize transioit changes in terminal compositions. The variables chosen for feedforward compensation will depend on which feedback control scheme is used—Figure 19.2,19.3, or 19.4. 

In Chapter 12 we proposed a particular control-loop stmeture that incorporates overrides and antireset windup. It also accommodates feedforward compensation and advanced control techniques without interfering with either normal reset or antireset windup. We also suggested that decouplers could be designed to compensate for interactions in e same way that feedforward compensators are designed. The technique here leads to stable, noninteracting control, but not necessarily to optimum control. Modem control theory, with its more sophisticated approaches to multivariable control, sometimes reqtiires some interaction for optimality. 

Feedforward compensators may be added for almost any disturbance, as shown by Figure 20.7. Physically each may be implemented, as discussed in Chapter 12, with an impulse function and a summer connected inside of the composition feedback controller reset circuit. 

Individual feedforward compensator functions may be determined in a very simple feshion. Let us, for example, look at feedforward compensation for feed-rate changes, F. If the compensators do a perfect job, then yr s) = 0 and Xb s) = 0, and there will be no contribution from the feedback controllers. Then, to make yj s) = 0, we can see by inspection of Figure 20.7 that ... 

Some of the practical problems in feeding forward from feed compensation have been discussed by Luyben. Luyben has also discussed the problems caused by inverse response in the design of feedforward compensators as well as feedback controllers for composition. 

---