Artificial Neural Networks in Sensors Signals Processing within Marine Diesel Engine Process

摘要

Application of new technologies on ship systems, especially software ones: neural networks, expert systems, fuzzy logic, genetic algorithms, pattern recognition, information fusion and so on, gives new signal processing, diagnostic and control possibilities and complete supervision and control of ship as fully integrated system - realisation of the idea of intelligent ship. This work presents a part of research results [1] which deals with artificial neural networks (ANN) possibilities in the area of sensors signals processing within control and diagnosis of marine diesel engine process. Two characteristic examples have been illustrated: 1. engine speed signal estimation (faulty speed sensor) by means of sensor information fusion; 2. engine telegraph position change detection (demand value) and speed command processing according to new engine telegraph position. The most appropriate network structures have been chosen and tested and some of simulation results have been given.