Ali-Akbar Ahmadi - epartment of Electrical and Computer Engineering, Faculty of Engineering, Kharazmi University, Tehran, Iran;

In this paper, actuator fault detection for networked dc motor system in presence of unknown load torque is considered, while network-induced delays and data packet dropouts are taken into account. The network-induced delays and data packet dropouts are modeled as a bounded unknown time-varying delay in the control input. Then, using the eigenstructure assignment technique and sliding mode observer theory, a novel actuator fault detection observer (FDO) is designed which compensates the effect of the unknown time-varying delay on the residual and simultaneously, decouples the residual vector from the unknown load torque. There is not any restrictive assumption on type of the actuator fault and load torque in the design process, which makes the proposed approach applicable for the real industrial applications. Finally, numerical simulations are carried out to show the feasibility and applicability of the proposed approach.

Networked dc motor system, Fault detection, Actuator fault, Time-delay, Data packet dropout.