Interaction between control loops

Interaction can be between two or more processes or between actions produced by different control loops applied to a single process. The former has already been discussed in Section 1.53. Some degree of interaction between control loops will nearly always occur in a multiple-input/multiple-output (MIMO) system. For example, consider the distillation process described in Section 7.3 (Fig. 7.9). Suppose it is desired to control simultaneously the compositions of both the overheads product stream (by manipulating the reflux flowrate) and the bottoms product stream (by regulating the steam flowrate to the reboiler). A typical arrangement is shown in Fig. 7.73. 

In general, we can write, from Fig. 7.74, for any pair of interacting control loops  

FRC flow recorder controller TRC temperature recorder controller LIC level indicator controller PIC pressure indicator controller 

Simultaneous control of the compositions of the overhead and bottom product streams of a distillation process 

Hence the closed-loop transfer function %/% can be determined for R2 held constant ( 2 = 0) and 2/ 2 when R, is held constant (, = 0). 

---

Interaction between control loops is discussed in Section 7.15. 

Fig. 7.74. Block diagram illustrating interaction between control loops on distillation column shown in Fig. 7.73...

Estimating tha Degree of Interaction Between Control Loops... 

It should be noted that interactions between control loops is not just limited to interacting units but will also occur within single units. A typical example is an exothermic reactor where a change in the control loop which controls the level in the reactor will have an effect on the amount of material in the reactor. This in turn will affect the heat removal requirements and, therefore, the cooling water control loop. Many strategies for reducing loop interactions, and for selecting control loops so as to minimise interactions, can be found in most standard control textbooks.1-11... 

If /Ijj is close to 1, then there is almost no interaction between control loops. The system is almost decoupled. Note that interactions are not necessarily bad, because interacting loops may help each other in rejecting disturbances. 

When considering alternative eontrol schemes, one should pay attention to interaction between control loops. Interaction means a mutual influence between the eontrol loops, onesided interaction is bothersome but does not affect the stabihty of the eontrol system. An undesirable effect of interaction can be that the dynamic behavior of one control loop is strongly affected by another eontrol loop. An extreme case of undesirable interaction is a situation in which a sign change takes effect, negative feedback changes to positive feedback and consequently the control loop becomes unstable. 

In this chapter, control of a fluid catalytic cracker will be discussed. A simulator, based on an available model as discussed in the literature, has been developed, which allows the reader to simulate various control schemes and to see the effect of interaction between control loops. Also, the selection of control loops is discussed for this case study. In this particular case study, there are more controlled variables than manipulated variables, which creates the need to develop cascade control structures. Some problems that could occur are highlighted. 

---