미니트램의 차량 자이로 센서 기반 온라인 파라미터 추정을 통한상태 변수 보정 및 고장 신호 검출 전략

Abstract

This paper introduces a new approach to fault diagnosis and isolation (FDI) of a mini-tram that shuttles in a certain region and follows commands from a central control system. In order to diagnose system faults such as errors in steering angle, steering actuator, wheel speed and yaw rate, residuals with respect to system signals are needed. The residuals are obtained from analytical redundancy, which is well known for its dynamic model-based approach in FDI. System parameters are changed by a variety of reasons so that online parameter estimation is necessary to compensate for the error of the system based on the vehicle dynamic model. The analytical redundancy for the residue is mainly divided into three steps: a model based process, a probabilistic process and command model-based process. Finally, the fault signals are isolated by fault logic, which is determined from a combination of the residues in a probabilistic way. In this study, the fault signals are virtually generated by an off-line process with real experimental data. The proposed FDI algorithm is verified with these signals.