A Novel Method Based on Fuzzy Logic and Data Mining for Synchronous Generator Digital Protection A new internal fault protection scheme for synchronous generators, based on the fuzzy inference system, is proposed. The proposed method is developed and tested by using an oscillogram dataset gathered from fault simulations and experiments conducted on real-life synchronous generators. Data mining is used to support the choice of the most relevant variables to identify synchronous generator internal faults. The selected variables are used as inputs to a fuzzy-logic system designed to identify and detect faults. A part of the oscillogram dataset is used to demonstrate the effectiveness of the proposed protection. The results show that the proposed protection scheme is more sensitive in detecting ground faults in synchronous generators with high-impedance grounding compared to conventional devices. It is also shown that the proposed scheme is able to identify interturns and interpath faults. The proposed method was embedded in a PC104 processor and tested in a real-time environment to show its practical feasibility.