Multivariate Linear Regression Model For Application Of Flotation Recovery And Grain Size Uniformity In Dashan Plant And Its Application

Grinding fineness has been considered as a major indicator in the production of mineral processing, while the influence caused by the particle size uniformity of grinding products was easily neglected. There exists the "both ends are more but the middle is less" phenomenon in particle size distribution of grinding products. For many concentrating plants, the non-uniformity of particle size has been gradually become the obstacle in dealing with indexes stability and improvement in mineral processing. How to effectively improve the particle size uniformity of grinding products will always be one of the main subjects for study. Recently, a serious of achievements have been made, such as choosing efficient grinding and classification equipment, improving the grinding or classification technological process, adjusting processing parameters of grinding and classification reasonably and some other measures. The article shows how to adjust the adding ball charge to increase the particle size uniformity of grinding product based on the ore-bearing rock mechanics properties for grinding ore and the related parameters of the mill. The multivariate linear regression model between particle size uniformity and flotation recovery is established by using the method of regression statistical analysis on the basis of a large number of small flotation tests.The loading and adding ball charge for 5.5×8.5m overflow ball mill of Dashan concentrating plant have been studied in laboratory. The comprehensive analysis of grinding indicators shows that the best size and component of loading ball charge is Φ70:Φ60:Φ40:Φ30= 20:30:20:30, and the best size and component of adding ball charge is Φ70:060:Φ40=35:35:30, including the yield of -0.2+0.01mm easy dressing fraction and -0.074mm fraction,-0.074mm mill utilization coefficient and grinding technical efficiency. Under the condition of Φ80mm and Φ50mm steel balls added, the test showsΦ50mm steel ball ratio can be appropriately added to achieve the ideal effect of grinding in view of the limitations of production conditions and actual situation of ball storehouse.Based on laboratory tests, an industrial test of adding ball charge in the grinding was carried out. Results of the industrial tests indicate that the right direction is to increase Φ50mm steel ball proportion in order to improve the grinding fineness and the particle size uniformity of grinding products, and the final amount of add ball charge (Φ80:Φ50=55:45) is determined. After adjusting adding ball charge propely the grinding effect has been improved evidently:firstly, the particle size uniformity of grinding and classification products was effectively improved. The content of-0.2+0.038mm easy dressing fraction and -0.074mm fraction increased by 5.86 and 3.83 points respectively, and the content of +0.2mm coarse fraction reduced to 3.75 points for overflow of cyclone. Secondly, mill productivity and classification efficiency were greatly improved. Examples as mill capacity of per unit time increased by 3.61 t·h-1 compared with the referenced mill in stable period of industrial test, the -0.074mm and -0.2+0.038mm utilization efficiency improved by 7.00% and 11.56% respectively, the classification efficiency improved by 6.00 points compared with the last industrial test. What’s more, the processing indexes were effectively improved, the yield and recovery of copper concentrate increased by 0.51 and 1.24 points. Meanwhile, its ore grade level had been reduced, which can reduce the loss of the metal. In conclusion, the previous prediction of improving copper flotation indicator has been achieved by improving particle size uniformity of grinding products.In this thesis, the copper roughing recovery, run-of-mine grade, copper roughing grade and -0.2+0.038mm dressing fraction’s content of grinding products of Dashan Ore-dressing Plant were multiple linear regression analyzed. And a multiple linear regression model of these were achieved by the least square method using EViews statistics software, and it through various statistical analysis and empirical tests, which is s= 36.28871+33.22137α-1.2798610+0.916851γ, and the really means represented by the multiple linear regression model coincidence of the production status of Dashan Ore-dressing Plant and it has guiding significance for production management of dressing plant.