二容系统分析.pdfVIP

Part I Reconfiguration Problem This part of the manuscript provides an answer to the question of what control reconfiguration is. In order to derive a general formulation of the problem, only few assumptions are made. Five reconfiguration goals of different strength are defined, which correspond to common control objectives. By including one goal or another, different variants of the problem can be constructed. This makes the approach very flexible and suitable for different applications. The first chapter of this part introduces the 2-Tank System, which will serve as a running example of control reconfiguration. Based upon the discussion of this example, the nonlinear reconfiguration problem is developed. The notion of a “reconfiguration block” is introduced in order to define a general system theoretical framework for reconfiguration. In the last chapter of this part, the reconfiguration problem is applied to a linearised model of the system. This allows for the application of powerful methods from linear time-invariant control theory. 3 Running Example: the 2-Tank System 3.1 Nonlinear Model A simple system will be used as a running example throughout the manuscript. It serves to illustrate typical faults and the resulting reconfiguration approa- ches. The example is a reduced version of the popular Three-Tank Benchmark Problem first described by Heiming and Lunze [1999]. The system contains two tanks connected by a valve and filled by a pump. The goal is to main- tain a constant outflow of the system, which requires a constant level in the right tank. Although the system is simple (see Fig. 3.1), it is sufficient to demonstrate the relevant effects encountered with respect to reconfiguration. Pump u1 Tank 1 Tank 2