Industrial Brewery Modelling by Using Artificial Neural Network

Fermentation is a complex phenomenon well studied which still provides challenges to brewers. In this study, artificial neural network, precisely multi layer perceptron and recurrent one were utilised for modelling either static (yeast quantity to add to wort for fermentation) or dynamic (fermentation process) phenomena. In both cases, the simulated responses are very close to the observed ones with residual biases inferior to 4.5%. Thus, ANN models present good predictive ability confirming the suitability of ANN for industrial process modelling.

[1]  F. Bruckert,et al.  Kinetics of cell spreading. , 2005, Physical review letters.