Intermittent Sensor Faults Detection Based on Fractional-Order Transient Chaotic System

Abstract

Safety and reliability are the most important issues for modern automation systems, fast, precise, and reliable fault detection (FD) is a feasible and easy-implemented method for preventing property loss and casualties. Considering the importance of sensors in modern systems and the characteristics of intermittent faults, the problem of intermittent sensor FD under unknown disturbance and strong noise is the focus of this article. Oriented by the problem, a kind of intermittent sensor faults detection scheme is proposed. The fractional-order transient chaotic system is adopted for the processing of the residual signal and the calculating of the detection threshold. A continuous stirred tank heater (CSTH) model is adopted for the simulation experiment, a simplified F-16 aircraft vertical dynamic model is used for the comparison experiment, and a real-world liquefied natural gas (LNG) filling process is provided for actual experiment. The experiment results show that the proposed intermittent faults detection (IFD) scheme can achieve the purpose of IFD under unknown disturbance and strong noise and performs well. The detectability of the intermittent sensor faults within the proposed IFD scheme is explored.