Two degrees of freedom controller

Centralized control can be also designed based on disturbance rejection or robustness requirements. In this case, the controller is not a static linear feedback law, as (45), but a dynamic feedback controller is obtained. Additionally, two degree of freedom controllers allow for a better control behavior in tracking and regulation. All these alternatives are beyond the scope of this introductory local control design treatment and are the subject of specialized references (see, for instance, [19]). 

Some DCS include an algorithm described as the two degrees of freedom controller. This has the form... 

The denominators of YID in (16-24) and Y/Ygp in (16-22) are the same, but the numerator terms are quite different in form. Figure 16.11 shows disturbance responses for Example 16.2 (0 = 2) for PI controllers with and without the Smith predictor. By using the two degree-of-freedom controllers discussed in Chapter 12, it is possible to improve the response for disturbances and avoid this undesirable behavior. In fact, In-gimundarson and Hagglund (2002) have shown that for step disturbances and a FOPTD process, the performance of a properly tuned PID controller is comparable to or better than a PI controller with time-delay compensation. 

The IMC block diagram in Fig. 12.5 can be expanded to include a block A in the feedback path as well as a disturbance transfer function G. The block A can be used to predict the effect of the disturbance on the error signal to the controller, and it can also provide time-delay compensation. This two-degree-of-freedom controller (see Chapter 12) is known as an analytical predictor (Doss and Moore, 1982 Wellons and Edgar, 1987). 

---

See also in sourсe #XX -- [ ]

---