Controller with neural network for estimating gas turbine internal cycle parameters

A gas turbine control system for generating control signals to govern operation of a gas turbine, said control system comprising:

• a controller coupled to a plurality of respective operating parameter sensors, said operating parameter sensors being coupled to said gas turbine to provide respective turbine operating condition signals to said controller;
• and
• at least one neural network estimator, said estimator being coupled to receive a plurality of neural net input signals, at least one of said neural net input signals originating from one of said operating parameter sensors;
• said controller further being coupled to a plurality of turbine actuator systems, said actuator systems being coupled to said gas turbine so as to control the operation of said gas turbine in response to respective actuator system control signals generated by said controller;
• said controller comprising a processor for generating said respective actuator system control signals in correspondence with said turbine operating condition signals, said at least one neural network estimator being trained to generate at least one estimated turbine operating condition signal;
• said estimated turbine operating condition signal generated by said neural network estimator representing a turbine internal cycle parameter for which said turbine has no corresponding operating parameter sensor.

A control system for generating control signals to govern operation of a gas turbine, said control system comprising:

• a controller coupled to a plurality of turbine actuator systems, said actuator systems being coupled to said gas turbine so as to control the operation of said gas turbine in response to respective actuator system control signals generated by said controller;
• and
• a plurality of respective operating parameter sensors coupled to said gas turbine and to said controller so as to provide respective turbine operation condition signals to said controller, said parameter sensors comprising respective turbine pressure and temperature sensors;
• said controller comprising a processor for generating said respective actuator system control signals in correspondence with the plurality of turbine operating condition signals, said turbine operating condition signals comprising respective turbine pressure, temperature, and fluid flow signals;
• said processor comprising a neural network estimator trained to generate an estimated turbine combustion reference temperature signal in response to a plurality of neural net input signals, at least one of said neural net input signals originating from one of said operating parameter sensors.

A control system for generating control signals to govern operation of a gas turbine, said control system comprising:

• a controller coupled to a plurality of turbine actuator systems, said actuator systems being coupled to said gas turbine so as to control the operation of said gas turbine in response to respective actuator system control signals generated by said controller;
• and
• a plurality of respective operating parameter sensors coupled to said gas turbine and to said controller so as to provide respective turbine operation condition signals to said controller said parameter sensors comprising respective turbine pressure and temperature sensors;
• said controller comprising a processor for generating said respective actuator system control signals in correspondence with a plurality of turbine operating condition signals, said turbine operating condition signals comprising respective turbine pressure, temperature, and fluid flow signals;
• said processor comprising a neural network estimator trained to generate an estimated turbine combustion reference temperature signal in response to a plurality of neural net input signals, at least one of said neural net input signals originating from one of said operating parameter sensors;
• said neural network estimator comprising a feedforward neural network net having an input neuron layer, a hidden neuron layer, and an output neuron layer;
• said input neuron layer and said hidden neuron layer each comprising a plurality of neurons, and said output neuron layer comprising a single neuron;
• said neural net input signals comprising a compressor discharge pressure signal, a turbine exhaust temperature signal, a compressor discharge temperature signal, and a turbine air flow input signal.