Yen Gary G, Ho Liang-Wei
School of Electrical and Computer Engineering, Oklahoma State University, Stillwater, OK 74078, USA.
ISA Trans. 2004 Oct;43(4):549-69. doi: 10.1016/s0019-0578(07)60168-2.
As dynamic systems become increasingly complex, experience rapidly changing environments, and encounter a greater variety of unexpected component failures, solving the control problems of such systems is a grand challenge for control engineers. Traditional control design techniques are not adequate to cope with these systems, which may suffer from unanticipated dynamic failures. In this research work, we investigate the on-line fault tolerant control problem and propose an intelligent on-line control strategy to handle the desired trajectories tracking problem for systems suffering from various unanticipated catastrophic faults. Through theoretical analysis, the sufficient condition of system stability has been derived and two different on-line control laws have been developed. The approach of the proposed intelligent control strategy is to continuously monitor the system performance and identify what the system's current state is by using a fault detection method based upon our best knowledge of the nominal system and nominal controller. Once a fault is detected, the proposed intelligent controller will adjust its control signal to compensate for the unknown system failure dynamics by using an artificial neural network as an on-line estimator to approximate the unexpected and unknown failure dynamics. The first control law is derived directly from the Lyapunov stability theory, while the second control law is derived based upon the discrete-time sliding mode control technique. Both control laws have been implemented in a variety of failure scenarios to validate the proposed intelligent control scheme. The simulation results, including a three-tank benchmark problem, comply with theoretical analysis and demonstrate a significant improvement in trajectory following performance based upon the proposed intelligent control strategy.
随着动态系统变得越来越复杂，经历快速变化的环境，并遭遇更多种类的意外部件故障，解决此类系统的控制问题对控制工程师而言是一项巨大挑战。传统的控制设计技术不足以应对这些可能遭受意外动态故障的系统。在这项研究工作中，我们研究在线容错控制问题，并提出一种智能在线控制策略，以处理遭受各种意外灾难性故障的系统的期望轨迹跟踪问题。通过理论分析，得出了系统稳定性的充分条件，并制定了两种不同的在线控制律。所提出的智能控制策略的方法是持续监测系统性能，并通过基于我们对标称系统和标称控制器的最佳了解的故障检测方法来识别系统的当前状态。一旦检测到故障，所提出的智能控制器将通过使用人工神经网络作为在线估计器来近似意外和未知的故障动态，从而调整其控制信号以补偿未知的系统故障动态。第一种控制律直接从李雅普诺夫稳定性理论推导得出，而第二种控制律基于离散时间滑模控制技术推导得出。两种控制律均已在各种故障场景中实施，以验证所提出的智能控制方案。包括三容水箱基准问题在内的仿真结果与理论分析相符，并表明基于所提出的智能控制策略，轨迹跟踪性能有显著提高。
