This content is not included in your SAE MOBILUS subscription, or you are not logged in.
Revolution Control of Generator Diesel Engine by Neural Network Controller
ISSN: 0148-7191, e-ISSN: 2688-3627
Published March 03, 2003 by SAE International in United States
Annotation ability available
Three types of control system for an electric governor of a diesel engine developed for generators were examined experimentally. The quantity of fuel injection of a diesel engine for generators is usually controlled by an electric governor system in these decades, and a PID controller has been installed for the electric governor. Even when the optimal parameters for PID controller are well tuned, it is difficult to keep constant rotating speed of the engine, because the applied load to generators may vary according to its running conditions. In this research, a neural network was applied to regulate the parameters in PI controller for the axial-moving DC motor to control the quantity of fuel injection. Experimental studies showed that the parameter regulation system using neural network presented here had good performance under various running conditions. Further, the neural network control system without PID controller also showed good performance in controlling diesel engines for generators.
CitationTsuchiya, T., Morishita, S., Enomoto, T., Sasaki, H. et al., "Revolution Control of Generator Diesel Engine by Neural Network Controller," SAE Technical Paper 2003-01-0365, 2003, https://doi.org/10.4271/2003-01-0365.
- JSAE Handbook on Automotive Engineering 2 1991 118 120
- Okubo, Y. Introduction of Fuel Injection System Sankai-do Press 1990
- Den Hartog, J.P. Mechanical Vibrations Dover Pub. Inc. 1985 309 317
- Anderson, J.A. Rosenfield, E. Neurocomputing, Foundation of Research MIT Press 1988 675
- Chinmoy PAL Morishita, S “Dynamic System Identification by Neural Network Using Time Series Data” JSME International Journal 38 4 1995 686 692
- Narendra, K.S. Parthasarathy, K. “Adaptive Control of Nonlinear Multivariable System Using Neural Networks” IEEE Trans. on Neural Network 1 1 1990 4 27