This chapter presents the main ideas of the book to deal with constraints in feed back control systems. The proposed method aims at preserving the simplicity of conventional anti-windup (AW) algorithms, but at the same time it is supported by a solid theoretical background and exhibits distinctive robustness features. Among them, the method does not require an exact model of the plant (it only needs knowledge of the model structure), it can be employed to address constraints in non-linear processes and it can transparently deal with a wide variety of non linearities apart from amplitude saturation, such as rate limiter, asymmetry, dead zone, and time dependency. For the sake of simplicity, only single-input single output (SISO) systems are considered in this chapter.

This chapter presents multivariable dynamical system and improvement in the decoupling of multiple-input multiple-output (MIMO) systems in the presence of different types of constraints, that is, to reduce interactions among the loops as much as possible when the plant under control is subjected to either physical, structural or dynamic constraints. First, we recall some basic concepts of MIMO systems analysis. Then, the affine parameterisation of all the controllers which internally stabilise the feedback system is introduced, together with the closely related Internal Model Control (IMC). Based on this control strategy, some design procedures are described in order to achieve the dynamic decoupling of stable minimum phase (MP) systems, non-minimum phase (NMP) systems and unstable systems. Finally, the effects and limitations of diagonal (full) dynamic decoupling in systems with right-half plane (RHP) zeros are briefly discussed.

The methods for reducing or cancelling crossed interactions studied so far in multivariable systems were based on centralised multiple-input multiple-output (MIMO) controllers, which is also usual in multivariable control literature. How ever, despite the performance advantages of centralised controllers, the great majority of industrial process control applications still rely on decentralised or multiloop control structures. Because of their structural constraints, decentralised controllers are not able to suppress by their own interactions of the plant, which are only taken into account at the controller tuning phase. This is not a trivial problem to be solved. In fact, even when supervisory control tools like model predictive control are employed, the coupling among the loops has to be addressed at the lower level (generally decentralised PI/PID control) because of the long sampling time of the supervisory modes. Therefore, the coupling reduction under decentralised structures is a topic of great interest when considering practical control issues. In this chapter, we shall analyse and address this problem. First, some basic concepts related with this control topology are presented. In particular, the relative gain array (RGA) is introduced as a simple interaction measure, whereas the potential effects of crossed interactions are illustrated through a simple example. Then, the sliding mode reference conditioning (SMRC) technique is exploited here in order to limit the amplitude of decentralised control interactions.

This chapter describes the undesired effects caused by manual-automatic or con troller switching in process control. Manual-automatic commutations are often used in the start-up schemes of the process industry, whereas switching among different controllers is frequently used for the control of non-linear systems at given operating points. Both types of switching at the plant input can be attributed to structural constraints, limiting either the controller type (linear controllers are typically preferred) or the switching policy (industry applications seldom include sophisticated approaches, like linear parameter-varying). The performance degradation caused by these mode switches is first tackled by means of sliding mode reference conditioning (SMRC) ideas in single-input single output (SISO) systems. Then, multivariable sliding mode (SM) concepts are introduced and the SMRC bumpless algorithm is extended to deal with multiple input multiple-output (MIMO) processes, so as to avoid inconsistencies between the off-line controller outputs and the plant inputs. As a consequence, jumps at the plant inputs are prevented and undesired transients on controlled variables are significantly reduced.