Title :
Computational intelligence and low cost sensors in biomass combustion process
Author :
Pital, Jan ; Mizak, Jozef
Author_Institution :
Dept. of Math., Inf. & Cybern., Tech. Univ. of Kosice, Presov, Slovakia
Abstract :
Artificial intelligence techniques have been used for carbon monoxide and oxygen low cost sensors signal processing in biomass combustion. Considering a large scatter of the measured data two approximation tools using artificial neural networks have been tested for approximation of carbon monoxide emissions dependence on oxygen concentration in the flue gas: AForge. Neuro library and Neural Network Fitting Tool of Matlab. The comparable results of approximation have been obtained by testing of both approximation tools on the off-line measured data.
Keywords :
approximation theory; artificial intelligence; combustion; environmental factors; flue gases; neural nets; production engineering computing; AForge; Neuro library; approximation tool; artificial intelligence technique; artificial neural network; biomass combustion process; carbon monoxide emission; computational intelligence; flue gas; low cost sensor; neural network fitting tool; oxygen concentration; Approximation methods; Biological neural networks; Biomass; Boilers; Combustion; Process control; Sensors; biomass combustion; carbon monoxide emissions; orificial neural networks; oxygen concentration;
Conference_Titel :
Computational Intelligence in Control and Automation (CICA), 2013 IEEE Symposium on
Conference_Location :
Singapore
DOI :
10.1109/CICA.2013.6611681
