Forecast Model For The Thermal Parameter Of Pellet Mine

With the development of upsizing blast furnace and automation, and the furthest demand for lowering the coke ratio, the quality and detection of the material of the blast furnace is gradually valued. One of the reasons, why the quality of the pellet between China and North America is so large, is that the thermal parameter of the pellet mine is not stable, causing disadvantageous to the blast furnace. So, it is necessary for us to do certain research on the thermal parameter of the pellet mine.The thermal parameters of pellet mine include reduction disintegration index, reduction degree, the expansion index and losing index. However, the test of thermal process index is complex, which needs12hours or even more. The process of pellet production has ended before the point is tested, and the test results make no meanings to the production process. Therefore, under the condition of exsisting technological process, materials and production equipment, how to research the thermal parameters of the process in order to obtain the relative transparency of sintering process and to make the whole production running in the optimal state, is the key problem in the production means of pellet.First of all, this paper further analyzes the production process flow of the grate-kiln and emphasizes the stages in dry, pre-heating, sintering, freezing system, collecting the relative data. And then describe the detail of the modeling by mechanism and neutral network. Combining with the characteristic of production process of the pellet mine, establish the model of the thermal parameter of pellet mine using the modeling by mechanism and neutral network. On account of the four parameters for the thermal characteristic, four models are established, for the first three parts, namely, RDI, RI, RSI, is modeled by mechanism due to lots of historical research. For the last part, namely LSI, we can determine its model using the neutral network due to its numerous experimental data. The accuracy of predictable model can be guaranteed when the model is simulated.