Study On Fault Diagnosis Of Ship Electric Propulsion System Based On Multi-Source Information Fusion Technology

The history of electric propulsion system used on ship has been nearly100years. In1980s, following the rapid development of electronic technology especially at power supply system, and microelectronics and information technology, ship electric propulsion system has achieved more widely application. Because of its virtue of high mobility, good maneuverability, low vibration and noise, podded electric propulsion has shown a wide application prospect. This paper utilizes electric propulsion system as the object, studies the fault diagnosis method based on multi-source information fusion technology.The reliability of the electric propulsion system affects the safety of the ship production and operation. Timely diagnose and effectively repair is the main contents of the fault diagnosis research. According to the expert experience, fault database is established. Fault diagnosis system using the data of system operation state and the experience of fault database concludes fault cause and fault expected. So the reliability and integrity of data has the important influence to get the correct conclusion. The external environment that electric propulsion system works at is very complex, so the operation state of electric propulsion system is hard to describe completely by single parameter or sensor. Compared to single source information, multi-source information fusion can get more complete, more reliable and more accurate conclusions. Using information fusion technology can effectively eliminate the interference caused by uncertainty of single information source, improve the scientific of conclusion.D-S evidence theory is an early information processing method used in fault diagnosis. It is the expansion of the classical probability theory, a non-exactly inference algorithms, suitable for great deal of uncertainty information processing and analysis. Its core idea is speculative to judge the causes of the incident through a series of events evidence. Statistics is the theoretical foundation of D-S evidence theory. D-S evidence theory can handle the uncertainty caused by randomness and fuzziness. It can rely on the accumulation of evidence, make the range of assumptions narrow and obtain the judgment results through the accumulation of evidence. In data processing, D-S evidence theory does not need a priori probability.This paper first widely studies the basis method of multi-source information fusion technology. The fault of electric propulsion system is mainly concentrated in the propulsion motor section of podded electric propulsion. According to the fault types and fault characteristics of electric propulsion system, this paper uses D-S evidence theory to study the basic method of fault diagnosis and technical feasibility. Trying to figure out a useful fault diagnosis method to electric propulsion system based on multi-source information fusion technology. The work this paper has done will lay the foundation for the study of electric propulsion system fault diagnosis and predictions based on multi-source information fusion technology.