Slave controller

With this arrangement, the output of one controller is used to adjust the set point of another. Cascade control can give smoother control in situations where direct control of the variable would lead to unstable operation. The slave controller can be used to compensate for any short-term variations in, say, a service stream flow, which would upset the controlled variable the primary (master) controller controlling long-term variations. Typical examples are shown in Figure 5.22c (see p. 235) and 5.23 (see p. 235). 

One of the most useful concepts in advanced control is cascade control. A cascade control structure has two feedback controllers with the output of the primary (or master) controller changing the setpoint of the secondary (or slave) controller. The output of the secondary goes to the valve, as shown in Fig. 8.2n. 

Figu re 9.14 Principles of a cascade controller. The master controller controls the reactor temperature (Tr) the slave controls the cooling system temperature (Tc). 

Cascade control. The actuator of the slave controller (electric heater) maintains the temperature of the jacket water controlled in a value set by the master controller. 

Cascade loops consist of two or more controllers in series and have only a single, independently adjustable set point, that of the primary (master) controller. The main value of having secondary (slave) controllers is that they act as the first line of defense against disturbances, preventing these upsets from entering and upsetting the primary process, because the cascade slave... 

Providing external reset for the cascade master from the slave measurement is always recommended. This guarantees bumpless transfer when the operator switches the loop from slave control to cascade control (Figure 2.45). The internal logic of the master controller algorithm is such that as long as its output signal (m) does not equal its external reset (ER), the value of m is set to be the sum of the ER and the proportional correction (Kc(e)) only. 

In reactor temperature control applications, a slave controlling the jacket outlet temperature is recommended so that the dynamics of the jacket is transferred from the primary to the secondary loop. In temperature-on-temperature cascade systems, such as shown in Figure 2.45, the secondary controller should have little or no integral. 

Selective control configurations also require external feedback to protect them from reset windup. Part (d) has a combination of selective and cascade systems and in such configuration, the external reset (ER) signal (not shown) is taken from the measurement of the slave controller (FRC). 

In heat exchanger applications, cascade loops are configured so that the master detects the process temperature and the slave detects a variable, such as steam pressure, that may upset the process temperature. The cascade loop, responds immediately and corrects for the effect of the upset before it can influence the process temperature. The cascade master adjusts the set point of the slave controller to assist in achieving this. Therefore, the slave must be much faster than the master. A rule of thumb is that the time constant of the primary controller should be ten times that of the secondary, or the period of oscillation of the primary should be three times that of the secondary. One of the quickest (and therefore best) cascade slaves is the simple and inexpensive pressure regulator. 

A pressure regulator is an ideal slave controller in a cascade loop. This is possible due to its speed of response. 

Cascade control significantly reduces the effect of certain types of disturbances by applying two control loops in tandem, i.e., the output of one controller is the setpoint for the other controller. The secondary or slave controller receives its setpoint from the primary or master controller and operates on a much faster cycle time than the primary. As a result, the secondary controller can eliminate certain disturbances before they are able to affect the primary control loop. 

In tuning a cascade control system, the slave controller is tuned first with the master controller in manual. Often only a proportional controller is needed for the slave loop, since offset in that loop can be treated by using proportional plus integral action in the master loop. When the slave controller is transferred to automatic, it can be tuned using the techniques described earlier in this section. Seborg et al. (1988) and Stephanopoulos (1984) provide further analysis of cascade control systems. 

Basically, a cascade control system has a master controller and a slave controller. The slave controller detects and corrects minor changes in the system, allowing the master controller to keep the primary variable constant. 

The tray temperature controller is the secondary (slave) controller. It is set up in exactly the same way as we did in the previous section. It looks at tray temperature (Stage 9) and manipulates reboiler heat input. However, its SP is not fixed. The SP signal is the output signal of the composition controller, which is the primary (master) controller. 

Another variation on the manipulated distillate scheme is to add a cascade slave control loop for the distillate flow rate as shown in Figure 2.1. The temperature control loop then manipulates the setpoint for the slave distillate flow control loop. Similarly, a slave flow control loop can be used for the reflux flow rate and another for the bottoms flow rate. At one point in time, the use of cascade control loops was called advanced control, it was compared to SISO (single input single output) control, because it required the addition of more hardware PID controllers. Modern computer control systems simply require the addition of software code to the program when the flow sensors are present. 

Another variation on the manipulated reflux scheme is to add a cascade slave control loop for the reflux flow rate as shown in Figure 2.1. The temperature control loop then manipulates the setpoint for the slave reflux flow control loop. Similarly a slave flow control loop can be used for the distillate flow rate and another for the bottoms flow rate. 

A cascade control system consisting of a master PID and two slave PI controllers was employed to maintain the polymerization temperature within 0.1 °C of the setpoint value by manipulating the cold and hot water flowrates to the reactor jacket. The master controller monitors the reaction temperature and its output drives the setpoint of the slave controller. The latter monitors the outlet temperature of the jacket fluid and drives the two separate control valves. 

The reason for the non-satisfying control performance of batch process units very often is the slave process that can have a more complex dynamics than the master loop has. As the slave process is determined by the mechanical construction, it is straightforward to design a model-based controller based on a nonlinear tendency model of the slave process. It has been shown that the parameters of the model-based slave controller (namely the parameters of the tendency model) can be easily determined by simple process experiments, and the complexity of the controller is comparable to that of a well furnished PID controller. Real-time control results showed that the proposed controller effectively handles the constraints (no windup) and gives superior control performance. 

In this configuration, the outer portions of IPMC serve as one sector (Sector 1) and are driven by input i(0> while the middle portion serves as a second sector (Sector 2) and is driven by input U2(t). Thus, toe IPMC comprises two controllable sectors. To control the performance of the sectored IPMC, toe feedforward architecture in Fig. 15 can be used. In this diagram, Ci(s) and C2 s) are the controllers and Gi(5) and G2(s) are their respective transfer functions associated with Sector 1 and Sector 2, respectively. Tstructure represents an open loop version of the master-slave control system. The advantage of the open... 

The slave controller, C2(s), serves to counteract the back relaxation of Sector 1 by inducing back relaxation in the reverse direction for Sector 2. With the inputs accordingly weighted, the back relaxation of each sector roughly cancels out. Then, C2(s) is responsible for setting U2(s) in appropriate proportion to l/i(s), such that the IPMC sectors relax the same amount. 

The feed flow is used for adjustment of the throughput, consequently the number of controlled variables exceeds the number of manipulated variables by one. One possibility is to leave one controlled variable free, but it is better to make the level or temperature control a slave controller of a quality controller by means of a cascade control stmeture (see chapter 33). The choice is determined by the variable with the highest impact on quality, either the temperatme or the residence time. 

Quality control can best be achieved by a master controller, which uses control of the most pronounced process condition as a slave controller. The control scheme is shown in Fig. 33.6. 

This is in particnlar the case if the process contains two large time constants, of which one is eontained in the slave loop. This is schematically shown in Fig. 33.7. The slave control loop is shown with an ideal measurement (no measurement dynamics). 

When Ti = 10 minutes and = 9, the slave control loop can be replaced by a first-order transfer function with a time constant of 1 minute. The value of the gain K is usually limited by secondary effects, such as non-linearities and smaller time constants, which also play a role). 

Reduction of disturbances that can be measured inside the slave controller... 

This is only relevant if the slave controller is faster than the master controller. An example of the use in reactor temperature control is shown in Fig. 33.10. 

Disturbances in the steam supply can be suppressed at an early stage by the slave controller. This controller is trimmed by the master controller, which controls the temperature of the reactor. The reactor temperature controller has considerably slower dynamics. Through the use of this cascade control structure disturbances are eliminated. Without this structure, the reactor temperature would first have to change before control action is taken. 

Usually, the outlet temperature is controlled automatically. This is most likely also the highest temperature that the feed will experience. For safely reasons, the control stracture is realized by adjrrsting the fuel supply, often via a slave controller. Figure 33.16 shows a pressure slave eontroller that adjusts the fuel supply to the burners. The advantages of this slave eorrtroller are ... 

---