Multivariable control systems - an engineering approach

ion is a key feature of control engineering. To realise that the dynamic behaviour of an aircraft can be represented with the same tools, and even equations, as a distillation column provides a platform to conceive control systems in a generic way. But this common framework for representing the dynamics of different processes does not mean that the particular characteristics of any process are included in the generic model. First, remember that a model is always a partial representation of a process. Second, it is worth remembering that the requirements, constraints, operating conditions and many other circumstances may be very different from one system to another. There are two basic approaches to getting the model of a process. On one hand, if the process is physically available and some experiments can be carried out, its dynamic behaviour can be captured and (partially) represented by a model, using “identification” and “parameter estimation” techniques [84]. On the other hand, if the operation of the process is fully understood and the governing principles are known, a first-principle model can be drawn, although some experiments should be carried out to determine some parameters. A model-based control may be developed and, in this case, the main steps in the design will rest upon the mathematical treatment of the global closedloop model [41].