2016
EMSS 2016 Proceeding
Modelling a biomass boiler using an artificial neural network
Authors: P. Ruiz, G. Sorrosal, C. E. Borges, A. M. Macarulla
Abstract
This paper is focused on the application of Artificial Neural Networks to model a recovery biomass boiler from the Joutseno paper mill (Finland). The cross- validation technique has been used to train the neural model. The validation phase has been carried out with new data, which have never been known or seen during the training procedure. As a result of the validation stage, the model has achieve only 1.77% of MAPE error for the main output variable (net steam), and reaching good performance metrics for the estimation of the main gas emissions. This work will be the basis of a future development using Artificial Neural Networks, in order to control and minimise the impact of the air emissions produced by the industrial plant.
I3M Scientific Sponsors
I3M Industrial Sponsors
I3M Media Sponsors
