METHOD FOR ASSESSING THE TECHNICAL STATE OF A CONSUMER CONTROLLER BASED ON NEURAL NETWORK DIAGNOSIS Russian patent published in 2020 - IPC G06N3/02 G01R31/34 

FIELD: measurement technology.

SUBSTANCE: invention relates to diagnosis of technical electromechanical equipment. Method comprises measuring parameters, calculating signs of rapid diagnosis, identifying faults and calculating deviations, taking into account algorithms of artificial neural network of Kohonen type, based on which are calculated and estimated coefficients of technical state of consumer-regulator by diagnostic subsystems, then based on obtained values of coefficients, technical condition is evaluated and diagnostic mode is selected, wherein coefficients are calculated taking into account changes in electric and vibration parameters over observed period of operation, then deviation of registered parameters is compared against standardized ones using neural network of Kohonen type and possibility of operation of electric installation as consumer-regulator is determined.

EFFECT: high accuracy and quality of assessing the technical state of equipment.

1 cl, 3 dwg, 3 tbl