Research On Fault Diagnosis Of AUV Thruster And Sensor Based On Fuzzy Qualitative Simulation

As an assistant of exploring and developing the ocean, Autonomous underwater vehicles (AUV) will play an important role in the field of marine resource exploitation and other aspects. Because of the complexity and unpredictability of marine environment, once AUV malfunctions, not only it can not complete underwater tasks, but also AVU itself may not be recoverable, resulting huge loss. Therefore, the fault diagnosis of AUW which supports the reliability and security is more and more concerned, becoming one of the hotspots in the field.Using an open shelf AUV experimental platform "Beaver" as research object, according to the object and method of fault diagnosis, we state the basic definition method of fuzzy qualitative analysis of AUV. Then, the relations of variation trend between variables in dynamics model and historical process data is translated into the fuzzy relations of variables. By means of fuzzy qualitative reasoning through fuzzy relations restrain with timing information, we obtained fuzzy qualitative state transition sequence in the normal and fault state respective.In the base of establishing an AUV fuzzy qualitative model, a fault diagnosis system combining decision tree method is given, and method for determining its basic elements as well. Then, a systematic monitoring decision tree with hierarchical characteristic and a fault identification decision tree are established, according to the predicted state of fuzzy qualitative model. Finally, the hierarchical decision tree based fault identification and the method of degree identification is proposed.Comparing the experimental data which is qualitative processed, with the fuzzy qualitative state sequence of the established model, the AUV fuzzy qualitative model is validated. And the feasibility of fuzzy qualitative simulation in application of AUV systems modeling is approved. Furthermore, fuzzy fault diagnosis method is used in obtaining fault identification results, which shows that the method is effective in AUV fault diagnosis.