The Analytics Of Blast Furnace Burden Distribution And Its Extensional Research

Blast furnace production is a continuous and complex process.Although the production manager has accumulated a considerable experience, but that the ironmaking process requirements become more and more sophisticated, therefore how to conduct scientific control of the production process has become an important technical issue. This paper mainly focuses on the analytics of the bell-less blast furnace burden distribution process, and establishes a bell-less blast furnace burden distribution model and then develops a soft system to solve the practical problem. It also establishes a model of the blast furnace top center temperature prediction to predict the center temperature. Furthermore, by extending the idea of analytics into hot rolling production process performance management, this paper studies the influencing factors of blockade hot-rolled steel coil, and then develops a hot rolling production process management system. The main contents of this paper include the following components:(1) According to the analytics of the blast furnace burden movement process, this paper respectively establishes the burden flowing trajectory model in the freefall space, burden surface function model, the correction model of burden descent and ore-coke ratio calculation model based on physics knowledge to analyze every aspect of the burden movement. Considering the impact of burden stream width on the material surface, this paper divides the burden stream into multi-strand when it flows out of the chute and calculates surface coordinates based on the actual volume of dumped material, and then calculates the burden surface feature information after the correcting the burden surface decrease.(2) Considering the inaccuracy of blast furnace top temperature detection, this paper gives a Least Square Support Vector Machine model for blast furnace center temperature prediction based on Particle Swarm Optimization (PSO) algorithm. It regress the relationship between the input variable and output variable by using training data, and uses the PSO algorithm to optimize the key parameters in this model and then obtains the final model. Verified by the result of test data, the result indicates that the model can accurately predict trends of the central temperature.(3)Based on blast furnace burden distribution model, the paper develops a blast furnace burden distribution system. The system is capable of calculating the burden distribution data in real-time. By the verification of actual production, the system is able to accurately describe the burden distribution, and realizes the visualization of the burden distribution. In addition, the development of off-line simulation model provides producers for simulating the burden distribution in different regulation, which makes the process of charging burden more effective and scientific.(4) The quantity of blockade hot-rolled steel coil which is the key indicators of hot rolling production process performance system is analyzed. With a large number of historical data, this paper aims to analyze the causes for blockade hot-rolled steel coil by mining and business analytics. It mines and analyzes the cause and key factors by using the decision tree algorithm of SAS/EM software package, and then gives the prediction and prescription. By the regulation of key production parameters, managers reduce the amount of hot rolled coils blockaded. Hot rolling production process management system achieves the hot rolling process data monitoring and management, and helps managers make scientific decision to improve the enterprise's business performance.