No abstract

This paper focuses on the detection methods of infrared photography to acquire the central data of GroupNo.6 blast furnace throat gas flow form Baotou Steel Corp. According to the situation that the distribution of blast furnace gas flow center is difficult to reflect certain regularities and class information as well as the problems that the traditional clustering methods failing to eliminate the influences of clustering noises, the density-based algorithm DBSCAN insensitive to noise with the merits of enabling to discover the class clusters of any shapes is adopted for the first time to establish the irregular classification model eliminating the noise data influence to analyze the relationship between the distribution of gas flow center and the gas utilization rate. Concept of "gas flow center deviation" is introduced in this paper. According to the research results, the gas utilization rate and the non-deviation of the gas flow center are positively correlated; the marginal development gas utilization rate is the lowest; the distribution of gas flow center presents the synchronous development type, which means the gas utilization rate is the highest when occupancy rate of gas with non-deviation and minor deviation is close to the sum and above 94%. Methods adopted in this paper could rapidly identify the relationship between the gas utilization rate and the gas distribution of the gas flow center, so as to provide further evidence for the online control of the blast furnace conditions.