Luenberger Observer-Based Speed Sensor Fault Detection: real time implementation to DC Motors

Abstract

Fault Tolerant Control Systems (FTCS) have emerged as a critical area of study for enhancing the safety, reliability, and efficiency of modern control systems. The FTCS technique might be active or passive control in general. In this paper, the active control branch's Fault Detection and Diagnosis (FDD) is used to detect faults in DC motor speed sensors. FDD methodologies can be divided into two types based on the process and the type of data available: model-based methods and data-based methods. The proposed method here investigates the use of the Luenberger observer technique, which is part of the model-based approach. The selected method was implemented and experimentally evaluated. This observer is dependent on the residual signal, which serves as a fault indicator in the overall system and represents the difference between the measured and estimated speed signals from the plant. Due to the increasing demand for these motors, particularly in electro-mechanical applications such as robotics, elevators, and electric-driven railways, a DC motor was chosen as a benchmark to test the proposed method. The output speed of the motor was subjected to four sensor faults: sensor fault, abrupt fault, intermittent fault, and incipient fault. The effectiveness of the suggested approach is demonstrated using MATLAB simulations, and the results show that faults are detected as anticipated with a high-performing response. Therefore, the proposed method was also implemented experimentally in real time and the obtained results showed a close match with those from simulation, thus proving the accuracy and reliability of the proposed methodology for fault detection in the DC motor speed sensor.