| Nowadays, Autonomous Underwater Vehicles (AUVs) are widely used in commercial, scientific, and military for various purposes. Higher request is put forward for the intelligent level of Autonomous Underwater Vehicles, which is an important component of ocean high-tech. Because of complex and unknown environment, it is necessary to embed fault diagnosis paradigms into AUVs to increase the reliability of the vehicles and enable them to execute and finalize complex missions.Underwater vehicles are liable to faults or failures during underwater missions. Thrusters are one of the most common and most important sources of faults. In all but the most trivial cases the existence of a fault may lead to canceling the mission. The implication of small faults could be very expensive and time consuming.This paper introduces a novel thruster fault diagnosis system (FDS) for open-frame underwater vehicles (UV). Basically, the FDS is a control allocator, but this primary function is enhanced with the ability of automatic thruster fault detection and diagnosis. The proposed FDS consists of two subsystems: a underwater vehicle motion state monitor subsystem (MSMS) and a thruster fault detection subsystem (TFDS). The MSMS analyses the residuals of the output of UV's motion model which is constructed with an improved Elman neural network and the real state value, to monitor its motion state. The TFDS uses fault detector units (FDUs), associated with each thruster, to monitor their state. Robust and reliable FDUs are based on RBF neural network and faulty states classifying methods.Fusion fault diagnosis unit (FFDU) colligates information provided by the two subsystems to locate and identify thruster fault. Robust and reliable FFDU is based on D-S evidence theory and fusion fault diagnosis rules. The indicators of faults degree and credibility for result of diagnosis are proposed to improve the reliability of fault diagnosis, and overcome distort of fault diagnosis. |
PDF Full Text Request