Incipient fault detection in nonlinear non-Gaussian noisy environment. (April 2021)

• Record Type: Journal Article
• Authors: Safaeipour, H.; Forouzanfar, M.; Ramezani, A.
• Is Part Of: Measurement. Volume 174(2021)
• Publication Date: 2021-04
• Subjects: Autocorrelation -- Incipient fault detection -- Model-based -- Non-Gaussian -- Residual signal -- Stochastic; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2021.109008
• Languages: English
• ISSNs: 0263-2241

Abstract

Abstract: Incipient fault detection in real-time nonlinear closed-loop systems in the presence of unwanted stochastic terms remains a challenging issue, especially in mixed Gaussian and non-Gaussian environments. This paper is concerned with incipient-fault detection in such systems. To this goal, based on the autocorrelation of the windowed residual signal and the reasonable assumptions in the nonlinear system, an online incipient fault detection with acceptable computational efforts and an adaptive-robust residual scheme is provided. Also, a closed-loop form of the three-tank system (DTS200) has been devised and simulated to demonstrate the effectiveness of the proposed solution. Graphical abstract: Highlights: Incipient fault detection in real-time nonlinear closed-loop systems in the simultaneous presence of Gaussian and non-Gaussian stochastic uncertainties. The proposed method is formulated on simple and low computational efforts, which is appropriate for practical purposes. A combination of adaptive and robust residual evaluation scheme is provided.