The application discloses a method and
system for diagnosing a turn-to-turn
short circuit of a double-water internally-cooled generator rotor based on an intelligent resistance method.The method comprises the following steps: collecting rotor
voltage, rotor current, rotor total water
inlet flow rate, rotor total water
inlet temperature and rotor total water outlet temperature sample parameters, and calculating a rotor resistance value; taking the rotor total water
inlet flow rate, the rotor total water
inlet temperature and the rotor total water outlet temperature as input quantities, taking the rotor resistance as an output quantity, submitting the sample to a
machine learning model training, and obtaining a trained model; obtaining real-time values of the rotor
voltage, the rotor current, the rotor total water
inlet flow rate, the rotor total water
inlet temperature and the rotor total water outlet temperature, and calculating a real-time value of the rotor resistance value; taking the real-time values of the rotor total water inlet flow rate, the rotor total water inlet temperature and the rotor total water outlet temperature as input quantities, inputting the trained model, and obtaining a rotor resistance fitting value through prediction; calculating a deviation value between the real-time value and the fitting value of the rotor resistance, determining whether the rotor has a turn-to-turn
short circuit according to the deviation value, and marking a turn-to-turn
short circuit severity.The method has low implementation difficulty, and the diagnosis result is accurate and reliable.