This study presents the process design and tune-up of robust artificial neural networks (ANN) to be used as virtual sensors for the diagnosis of a three-cylinder Diesel engine operating at various conditions. Particularly, a feed-forward neural network based on radial basis functions (RBF) is employed. The use of different radial basis functions, and their relevant parameters, is investigated in detail, with their effect on the network accuracy. The RBF network is validated using data not included in training, showing good correspondence between measured and reconstructed pressure signal. The accuracy of the predicted pressure signals is analyzed in terms of mean square error and in terms of a number of pressure-derived parameters. Results are promising in terms of performance and accuracy, both for the predicted pressure signals and for the pressure-derived engine parameters that can be used in a closed loop engine control system.

ANN-based virtual sensor for on-line prediction of in-cylinder pressure in a diesel engine

Continillo G;
2014-01-01

Abstract

This study presents the process design and tune-up of robust artificial neural networks (ANN) to be used as virtual sensors for the diagnosis of a three-cylinder Diesel engine operating at various conditions. Particularly, a feed-forward neural network based on radial basis functions (RBF) is employed. The use of different radial basis functions, and their relevant parameters, is investigated in detail, with their effect on the network accuracy. The RBF network is validated using data not included in training, showing good correspondence between measured and reconstructed pressure signal. The accuracy of the predicted pressure signals is analyzed in terms of mean square error and in terms of a number of pressure-derived parameters. Results are promising in terms of performance and accuracy, both for the predicted pressure signals and for the pressure-derived engine parameters that can be used in a closed loop engine control system.
2014
978-0-444-63434-4
Accelerometer; Internal combustion engine; Neural networks
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12070/10016
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 15
  • ???jsp.display-item.citation.isi??? 7
social impact