Static, Dynamic and Mixed Eccentricity Faults Detection of Synchronous Generators based on Advanced Pattern Recognition Algorithm

Static eccentricity (SE) and dynamic eccentricity (DE) fault are prevalent types of faults in the synchronous generators operating in the hydropower plant. Although there exist several methods for SE and DE fault detection of hydropower generators, the proposed fault-sensitive methods are based on invasive approach such as air-gap magnetic field monitoring. Whereas non-invasive methods based on the stator voltage or current monitoring are not sensitive to low-severity fault. Conclusively, stray magnetic field can be used since it overcomes the aforementioned shortcomings. The proposed detection algorithms for eccentricity fault detection require a baseline data from a brand new generator while it is impossible to have required baseline data for a synchronous generator operating for decades in the power plants. A precise and sensitive algorithm for early detection of SE and DE fault that does not require a baseline data of a brand new generator based on time-frequency analysis of the stray magnetic field is proposed. The proposed algorithm is verified on a 100 kVA laboratory setup and in two power plants with a power rating of a 22 MVA, and a 42 MVA.