High Temperature Distribution Measurement of the Blast Furnace Raceway Through Imaging Techniques and Optimization Algorithms

The high temperature distribution of the blast furnace raceway plays an important role in continuous and steady production. There are a number of techniques which have been developed for flame temperature measurements, but most of them have considerable errors and instability. To obtain more accurate results, an efficient optimization algorithm is necessary. This paper presents a novel measurement approach based on the two-color method and an optimization algorithm using images of raceway flame captured by the RGB CCD (Charged-couple Device) camera to calculate the projection temperature distri- bution. Furthermore, both the back-propagation neural network and the principal component analysis are applied to calibrate the correlation between blackbody temperature and the gray-level intensity. Experimental results obtained from the blast furnace raceway demonstrate that the proposed optimization approach performs well in the measurement offlame temperature compared with the measured temperature.