engleski

Utilizing a neural network modelling approach in the fixed bed biomass gasifier

Development of renewable energy resources recorded significant growth in the past several years due to climate changes caused by the combustion of fossil fuels. Conversion of biomass into the syngas suitable for further exploitation is recognized as a valuable energy resource because of the wide distribution and availability of raw materials. Biomass gasification is the thermochemical process of partial combustion in the oxygen-free environment with a final product of hydrogen- enriched gas. Development of mathematical models for describing physical and chemical behaviour of gasification process significantly replaces expensive experimental investigations with an aim to optimize the process. Artificial neural network (ANN) models are presented as a novel modelling method of biomass gasification using an unphysical approach based on experimental input and output data. In this work the possibility of using the dynamics Nonlinear autoregressive model with exogenous inputs (NARX) on fixed bed gasifier was analysed. Results showed that NARX has possibility of predicting the syngas composition from gasification process, although further improvement of the model is mandatory.

gasification ; syngas composition ; NARX ; artificial neural networks

nije evidentirano