This paper is focused on a relatively novel eco-efficient degreasing technique, namely Fluidized Bed Degreasing (FBD), based on a fluidised bed of hard particles. An experimental campaign was aimed to investigate the relationship between FBD operational parameters and degreasing effectiveness. Consistent trends of residual oil according to FBD process parameters were found and both a related power dissipation analytical model and a neural network were developed and verified by comparison with experiments. The Multi-Layer Perceptron (MLP) neural network, trained with Back-Propagation (BP) algorithm, gave the best performance. Finally, Genetic Algorithms (GAs) were used to improve the predicting capability of the neural network solution. In detail, an experimental plan was performed to check the generalisation capability of the neural network model with GA.
Modelling of Fluidized Bed Degreasing (FBD) Process by ANNs / Barletta, M; Gisario, Annamaria; Guarino, S; Tagliaferri, V; Umbrello, D; Filice, L.. - 2:43924(2007), pp. 294-309. (Intervento presentato al convegno 10th CIRP tenutosi a Reggio Calabria; Italy nel 27-28 August 2007).
Modelling of Fluidized Bed Degreasing (FBD) Process by ANNs
GISARIO, ANNAMARIA;
2007
Abstract
This paper is focused on a relatively novel eco-efficient degreasing technique, namely Fluidized Bed Degreasing (FBD), based on a fluidised bed of hard particles. An experimental campaign was aimed to investigate the relationship between FBD operational parameters and degreasing effectiveness. Consistent trends of residual oil according to FBD process parameters were found and both a related power dissipation analytical model and a neural network were developed and verified by comparison with experiments. The Multi-Layer Perceptron (MLP) neural network, trained with Back-Propagation (BP) algorithm, gave the best performance. Finally, Genetic Algorithms (GAs) were used to improve the predicting capability of the neural network solution. In detail, an experimental plan was performed to check the generalisation capability of the neural network model with GA.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.