Pregledni rad

PIDNN FOR MARINE DIESEL MAIN ENGINE SPEED CONTROL

Petar Matić orcid.org/0000-0002-1799-5257 ; Pomorski fakultet Sveučilišta u Splitu
Nikola Račić orcid.org/0000-0002-9089-089X ; Pomorski fakultet Sveučilišta u Splitu
Danko Kezić ; Pomorski fakultet Sveučilišta u Splitu

Sažetak

This work deals with Marine Diesel Main Engine speed controller which uses ANN to optimize PID controller
parameters thus obtaining better transient characteristics and better control properties. To validate this claim, PID
Neural Network (PIDNN) controller model was designed using Matlab/Simulink, and implemented in numerical model
of the ship propulsion diesel engine [5]. PIDNN controller uses one of the most common ANN structures (multi-layer
perceptron) and training algorithm (back-propagation), also briefly described in this work. The effectiveness of the
PIDNN controller is shown through simulation and experiment.

Ključne riječi

Artificial Neural Networks (ANN); PID Neural Network (PIDNN); back-propagation; speed control; Marine Diesel Main Engine; Matlab/Simulink

Hrčak ID:

44039

URI

https://hrcak.srce.hr/44039

Datum izdavanja:

7.12.2009.

Podaci na drugim jezicima: hrvatski

Posjeta: 2.630 *