Improving quality and combustion control in pyrometallurgical processes using multivariate image analysis of flames
|Abstract:||Combustion is used throughout the mineral processing industry to produce steam in boilers, to dry concentrates in rotary dryers, and to apply heat treatments in pyrometallurgical furnaces. Tight combustion control is very important in the latter type of furnace since the combustion conditions directly affect final ore quality. However, achieving tight combustion control is not straightforward since most of the flames encountered in industry are turbulent non-premixed flames, they are affected by several unmeasured disturbances, various flow rates, continuous variation in the mix between fuels since they are often produced by simultaneously burning several types of fuel, some of them coming from other parts of the plant. A novel method is proposed in this study to improve process and product quality control as well as to optimize the combustion conditions based on digital flame color images. Multivariate Image Analysis and Regression is used to extract the flame color characteristics from images to predict the solids discharge temperature of an industrial rotary kiln related to product quality. It is shown that this method yield extremely good 20 minutes, 40 minutes as well as 80 minutes ahead forecasts of the discharge temperature of mineral ore. This should lead to a substantial reduction in product quality variability as well as in fuel consumption.|
|Document Type:||Mémoire de maîtrise|
|Open Access Date:||12 April 2018|
|Collection:||Thèses et mémoires|
All documents in CorpusUL are protected by Copyright Act of Canada.