Fault-Tolerant Control of Mechanical Systems Using Neural Networks

Fault-Tolerant Control of Mechanical Systems Using Neural Networks

Sunan Huang (National University of Singapore, Singapore), Kok Kiong Tan (National University of Singapore, Singapore) and Tong Heng Lee (National University of Singapore, Singapore)
DOI: 10.4018/978-1-4666-1806-0.ch010
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Due to harsh working environment, control systems may degrade to an unacceptable level, causing more regular fault occurrences. In this case, it is necessary to provide the fault-tolerant control for operating the system continuously. The existing control techniques have given some ways to solve this problem, but if the system behaves in an unanticipated manner, then the control system may need to be modified, so that it handles the modified system. In this chapter, the authors are concerned with how this control system can be done automatically, and when it can be done successfully. They aimed in this work at handling unanticipated failure modes, for which solutions have not been solved completely. The model-based fault-tolerant controller with a self-detecting algorithm is proposed. Here, the radial basis function neural network is used in the controller to estimate the unknown failures. Once the failure is detected, the re-configured control is activated and then maintains the system continously. The fault-tolerant control is illustrated in two cases. It is shown that the proposed method can cope with different failure modes which are unknown a priori. The result indicates that the solution is suitable for a class of mechanical systems whose dynamics are subject to sudden changes resulting from component failures when working in a harsh environment.
Chapter Preview
Top

Introduction

Control systems include many components, such as transducers, sensors, actuators and mechanical parts. These components are required to be operated under some specific conditions. However, due to prolonged operations or harsh operating environment, the properties of these devices may degrade to an unacceptable level, causing more regular fault occurrences. It is therefore necessary to provide the fault-tolerant control which compensates for the fault of the component by substituting a configuration of redundant elements so that the system continues to operate satisfactorily (Patton,Frank and Clark, 1989).

Traditionally, the fault-tolerant control is achieved through the use of hardware redundancy (Sun,Wang,Howe & Jewell,2008). Multiple hardware elements are distributed spatially around the system to provide protection against system failures. The major problems encountered with hardware redundancy are the extra cost and space requirement. To overcome the problems, at least in part, software control approaches based on signal processing technique have been developed and reported during the last two decades. In Hamada, Shin & Sebe (1998), a design method for fault-tolerant multivariable control systems is proposed. In Aberkane, Sauter & Ponsart (2007), a static feedback controller is designed for a class of nonlinear stochastic discrete system subject to random failures. In Deutsch (2008), a fault-tolerant control system for a linear combustion engine is designed subject to system failures. In Rothenhagen & Fuchs (2009), a bank of observers is used to reconfigure the drive and reenter closed-loop control to a doubly fed induction generator. In Kambhampati, Perkgcz,Patton & Ahamed (2007), a reconfigured network control strategy is designed based on the fault isolation information. The advantage of this work is that the fault location information can be incorporated into the reconfigured controller. In recent years, intelligent techniques have grown to an important field in fault-tolerant control and to a certain extent, is being fueled by advancements in computing technology. Examples of intelligent techniques include expert systems (Visinsky,Cavallaro & Walker, 1994), fuzzy systems (Diao,Passino, 2001; Zhang,Huo,&Zhang,2008; Visinsky,Cavallaro & Walker,1995; Wang,Tong & Tong,2007; Tong, Wang & Zhang, 2008), and artificial neural networks (Polycarpou & Helmicki,1995; Polycarpou,2001; Liu,Wu & Zhang,2008; Farrell,Berger & Appleby,1993).

This chapter further expands on these reported works. We propose and develop a set of schemes suitable for fault-tolerant control of mechanical systems. The basic idea of the proposed method is to use the information provided by the neural networks to accommodate faults in order to permit continued operation of the system. First, a model-based fault detection is designed based on a nominal system model. The fault detection embedded into the controller is carried out by comparing the states with their signatures. Secondly, if the fault is detected, the controller is reconfigured by incorporating with neural networks which are used to capture the nonlinear characteristics of unknown faults. Finally, the case studies are conducted based on real-time machine actuation systems, while most of existing results are simulated ones. Unlike the existing neural network fault-tolerant control (Polycarpou and Helmicki,1995; Polycarpou,2001; Liu,Wu, Shi & Zhang,2008; Farrell, Berger & Appleby,1993)., the proposed controller can achieve the automated fault-tolerance using a deadzone operator.

Complete Chapter List

Search this Book:
Reset