INTELLIGENT SYSTEM FOR AUTOMATIC CONTROL OF SHIP ENGINE WITH BUILT-IN NEURAL NETWORK Russian patent published in 2024 - IPC B63H21/22 G05D101/15 G06F17/00 G06N3/43 

FIELD: automatic control systems; ships and other floating facilities.

SUBSTANCE: invention relates to intelligent systems for automatic control of ship engine based on neural network. System comprises a unit for reducing input signals and data from sensors to standard conditions, a unit for a neural network model of a ship engine and actuators, neural controller unit, regression calculation unit, Kalman filter unit, ship engine and actuator automatic control system unit, a unit of an influence coefficient matrix, a unit for reducing to real conditions of navigation, a unit for comparing influence coefficients, to which values of the influence coefficients are received from the unit of the standard conditions reduction and from the Kalman filter unit, which is connected to the output of the regression calculation unit. Technical result is provided by using fuzzy neural networks.

EFFECT: provision of estimation of expected consumption of fuel and level of emissions of hydrocarbons H_nC_m, carbon dioxide gas CO₂, nitrogen oxides NO_x in exhaust gases under various real conditions of navigation and generation of adequate commands for control of engine and actuating mechanisms for reduction of ecological load on environment.

1 cl, 3 dwg