The Application And Study Of A Combining Forecasting Method In Forecasting Sintering Return Mines

The return mines of sintering is one of the important factors in sintering process. Output and quality can be raised by adding return mines to mix raw materials of sintering. The blending ratio of return mines is determined by the balance principle of return mines. In order to optimize resource distribution, reduce the production costs and enable the sintering process to be stable and smooth, it is necessary to know the output of return mines. However, the sintering process is a continuous process system with multivariable factors and complex technology. It is hard to calculate the output of return mines directly according to the working condition. At present, the blending ratio of return mines is determined by the level of material tank and the personal operating experience. Since this method is lack of data supporting, the level of material tank fluctuates heavily. In order to offer the basis of blending ratio in data, it is important to predict the actual output of return mines.In this paper, firstly the process of sintering and its characteristic was introduced briefly, and the grey characteristic of return mines of sintering was analyzed.Secondly the author analyzed the methods of grey system theory and tune series analysis, which are extensively used at present. According to the methods' own characteristics and scope of application, on the basis of consulting other research results, taking account for actual problem of controlling the level of material tank, the author presented a kind of forecasting method which is combining the grey theory with time series analysis. This method firstly adopts grey system theory (GM(1,1)) to extract the term of definite trend, then use Auto-regressive Mvoing Average (ARMA) model to forecast the residual, and use the value of residual forecasting to mend the value that forecasted by GM(1,1). The forecasting accuracy and be improved by using this method.Finally the forecasting model was examined by sample data collected on site. The predict result indicates this method has high degree of accuracy and satisfied results which can offer data support to the determination of blending ratio of return mines.