An Artificial Neural Network Study on the Formation and Control of Dioxin from Incinerators

• Main Author: 王安良
• Other Authors: 謝樹木
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/82225232982647989539

碩士 === 東海大學 === 化學工程學系 === 90 === This study, utilizing artificial neural networks, correlates the emission of dioxins to the feed compositions and operating conditions of some incinerators. The ultimate concern is to search for the optimum operating conditions so that minimum dioxin emission values are obtained. Three successive stages were carried out to the designed purposes, namely, (1) finding the best neural networks for incinerators under study, (2) completing the learning and testing of the proposed network structure and (3) locating the optimum operating conditions. Both a commercial scale fluidized-bed and a water cooled incinerators are subjected to the simulation using Neuralworks Professional II/Plus developed by Neural Networks. Results indicate that the lowest dioxin emissions(at a value of 6.9 ng-TEQ/Nm3) occur if the temperatures are controlled at 909℃ and 253℃ for the furnace top and boiler out, respectively, for fluidized-bed incinerator; while the minimum dioxin emissions(at the value of 133 ng/Nm3) occur provided that the flue gas temperature is controlled at 210℃ and the furnace top(radiation zone) operated at 800℃ in the case of water cooled incinerator.