Publication: Sliding mode observer based fault detection system
Loading...
Date
2023-07
Authors
Ng, Chun Poh
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Fault detection is crucial in the manufacturing industry because it allows for the early diagnosis of malfunctioning systems before they impact overall processes.
This importance is heightened in networked systems, such as those with multi-robot manipulators working together on tasks. Because of the interconnectedness of subsystems, the deployment of an estimating system capable of measuring the health of each subsystem is required. To solve this, our research will first investigate a linear observer and then simulate it in the presence of noisy encoder feedback. This project focuses on studying a linear observer initially and simulating its performance under a scenario with noisy encoder feedback. The study then expands to develop a robust nonlinear observer using sliding mode theory. The state-space model representing the dynamic behaviour of the DC motor under examination is transformed into a nominal canonical form before designing the robust nonlinear observer. The fault introduced in the encoder sensor feedback takes the form of bounded white Gaussian noise. Simulations demonstrate the robust nonlinear observer's ability to successfully reconstruct the corrupted sensor feedback, resulting in accurate convergence to the true
position value. The experimentation solidifies the effectiveness of the proposed approach for fault detection and compensation in interconnected systems. Throughout the research, High Order Sliding mode observer demonstrates greater functionality and resilience in reconstructing plant values, even in the presence of external noise.