Skip to main content
main-content

Über dieses Buch

The series Advances in Industrial Control aims to report and encourage technology transfer in control engineering. The rapid development of control technology has an impact on all areas of the control discipline. New theory, new controllers, actuators, sensors, new industrial processes, computer methods, new applications, new philosophies . . . , new challenges. Much of this development work resides in industrial reports, feasibility study papers and the reports of advanced collaborative projects. The series otTers an opportunity for researchers to present an extended exposition of such new work in all aspects of industrial control for wider and rapid dissemination. The time for nonlinear control to enter routine application seems to be approaching. Nonlinear control has had a long gestation period but much ofthe past has been concerned with methods that involve formal nonlinear functional model representations. It seems more likely that the breakthough will come through the use of other more flexible and amenable nonlinear system modelling tools. This Advances in Industrial Control monograph by Guoping Liu gives an excellent introduction to the type of new nonlinear system modelling methods currently being developed and used. Neural networks appear prominent in these new modelling directions. The monograph presents a systematic development of this exciting subject. It opens with a useful tutorial introductory chapter on the various tools to be used. In subsequent chapters Doctor Liu leads the reader through identification, and then onto nonlinear control using nonlinear system neural network representations.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Neural Networks

Abstract
The field of neural networks has its roots in neurobiology. The structure and functionality of neural networks has been motivated by the architecture of the human brain. Following the complex neural architecture, a neural network consists of layers of simple processing units coupled by weighted interconnections. With the development of computer technology, significant progress in neural network research has been made. A number of neural networks have been proposed in recent years.
G. P. Liu

Chapter 2. Sequential Nonlinear Identification

Abstract
The identification of nonlinear systems using neural networks has become a widely studied research area in recent years. System identification mainly consists of two steps: the first is to choose an appropriate identification model and the second is to adjust the parameters of the model according to some adaptive laws so that the response of the model to an input signal can approximate the response of the real system to the same input. Since neural networks have good approximation capabilities and inherent adaptivity features, they provide a powerful tool for identification of systems with unknown nonlinearities (Antsaklis, 1990; Miller et al. 1990).
G. P. Liu

Chapter 3. Recursive Nonlinear Identification

Abstract
The system identification procedure mainly consists of model structure selection and parameter estimation. The former is concerned with selecting which class of mathematical operator is to be used as a model. The latter is concerned with an estimation algorithm and usually requires input output data from the process, a class of models to be identified and a suitable identification criterion. A number of techniques have been developed in recent years for model selection and parameter estimation of nonlinear systems. Forward and backward regression algorithms were analysed in Leontaritis and Billings (1987). Stepwise regression was used in Billings and Voon (1986) and a class of orthogonal estimators were discussed in Korenberg et al. (1988). Algorithms with the objective of saving memory and allowing fast computation have been proposed in Chen and Wigger (1995). Methods to determine the a priori structural identifiability of a model have also been studied (Ljung and Glad, 1994). A survey of existing techniques of nonlinear system identification prior to the 1980s is given in Billings (1980), a survey of the structure detection of input output nonlinear systems is given in Haber and Unbehauen (1990) and a survey of nonlinear black-box modelling in system identification can be found in Sjoberg et al. (1995).
G. P. Liu

Chapter 4. Multiobjective Nonlinear Identification

Abstract
The identification of nonlinear systems can be posed as a nonlinear functional approximation problem. From the Weierstrass Theorem (Powell, 1981) and the Kolmogorov theorem (Sprecher, 1965) in approximation theory, it is shown that the polynomial and many other approximation schemes can approximate a continuous function arbitrarily well. In recent years, a number of nonlinear system identification approaches, particularly identification using neural networks, based on the universal approximation theorem (Cybenko, 1989), are applications of a similar mathematical approach.
G. P. Liu

Chapter 5. Wavelet Based Nonlinear Identification

Abstract
The approximation of general continuous functions by nonlinear networks has been widely applied to system modelling and identification. Such approximation methods are particularly useful in the black-box identification of nonlinear systems where very little a priori knowledge is available. For example, neural networks have been established as a general approximation tool for fitting nonlinear models from input output data on the basis of the universal approximation property of such networks. There has also been considerable recent interest in identification of general nonlinear systems based on radial basis networks (Poggio and Girosi, 1990a,b), fuzzy sets and rules (Zadeh, 1994), neural-fuzzy networks (Brown and Harris, 1994; Wang et al., 1995) and hining hyperplanes (Breiman, 1993).
G. P. Liu

Chapter 6. Nonlinear Adaptive Neural Control

Abstract
Neural networks are capable of learning and reconstructing complex nonlinear mappings and have been widely studied by control researchers in the design of control systems. A large number of control structures have been proposed, including supervised control (Werbos, 1990), direct inverse control (Miller et al., 1990), model reference control (Narendra and Parthasarathy, 1990), internal model control (Hunt and Sbararo, 1991), predictive control (Hunt et al., 1992; Willis et al., 1992), gain scheduling (Guez et al., 1988), optimal decision control (Fu, 1970), adaptive linear control (Chi et al., 1990), reinforcement learning control (Anderson, 1989; Barto, 1990), indirect adaptive control (Narendra and Parthasarathy, 1990; Liu et al., 1999a) and direct adaptive control (Polycarpou and Ioannou, 1991; Sanner and Slotine, 1992; Karakasoglu et al., 1993; Sadegh, 1993; Lee and Tan, 1993). The principal types of neural networks used for control problems are the multilayer perceptron neural networks with sigmoidal units (Psaltis et al., 1988; Miller et al., 1990; Narendra and Parthasarathy, 1990) and the radial basis function neural networks (Powell, 1987; Niranjan and Fallside, 1990; Poggio and Girosi, 1990a).
G. P. Liu

Chapter 7. Nonlinear Predictive Neural Control

Abstract
Predictive control is now widely used by industry and a large number of implementation algorithms, including generalised predictive control (Clarke et al., 1987), dynamic matrix control (Cutler and Ramaker, 1980), extended prediction self-adaptive control (Keyser and Cauwenberghe, 1985), predictive function control (Richalet et al., 1987), extended horizon adaptive control (Ydstie, 1984) and unified predictive control (Soeterboek et al., 1990), have appeared in the literature. Most predictive control algorithms are based on a linear model of the process. However, industrial processes usually contain complex nonlinearities and a linear model may be acceptable only when the process is operating around an equilibrium point. If the process is highly nonlinear, a nonlinear model will be necessary to describe the behaviour of the process.
G. P. Liu

Chapter 8. Variable Structure Neural Control

Abstract
Variable structure control with sliding modes was first proposed in the early 1950s (Utkin, 1964; Ernelyanov, 1967; Itkis, 1976) and has subsequently been used in the design of a wide spectrum of system types including linear and nonlinear systems, large-scale and infinite-dimensional systems, and stochastic systems. It has also been applied to a wide variety of engineering systems. The most distinguished feature of variable structure control based on sliding modes is the ability to improve the robustness of systems which are subject to uncertainty. If, however, the uncertainty exceeds the values allowed for the design, the sliding mode cannot be attained and this results in an undesirable response (Utkin, 1964). In the continuous-time case this problem was solved by combining variable structure and adaptive control (Slotine and Li, 1991), but this requires that all the system variables are available and can be measured. This case has also been discussed for linear discrete systems using input output plant models (Furuta, 1990, 1993; Hung et al., 1993; Pan and Furuta, 1995) and for nonlinear discrete systems where the input output model is unknown (Liu et al., 1997b. 1999b).
G. P. Liu

Chapter 9. Neural Control Application to Combustion Processes

Abstract
Combustion processes exist in many applications related to power generation, heating and propulsion; for example, steam and gas turbines, domestic and industrial burners, and jet and ramjet engines. The characteristics of these processes include not only several interacting physical phenomena but also a wide variety of dynamic behaviour. In terms of their impact on system performance, pressure oscillations are of most significance. In some applications, pressure oscillations are undesirable since they result in excessive vibration, causing high levels of acoustic noise and in extreme cases mechanical failure. In the frequency domain, the pressure is characterised by dominant peaks at discrete frequencies which correspond to the acoustic modes of the combustion chamber.
G. P. Liu

Backmatter

Weitere Informationen