基于数字拟合数据方法的金属选矿产品结构预测与优化

Prediction and Optimization of Metal Beneficiation Product Structure on Mathematical Fitting Data Method

在既定的分选工艺条件下,为实现金属选矿产品结构的最优化,提出了一种数学拟合数据方法。根据产品的密度和分配率进行分配曲线的拟合,并采用差分进化算法通过Matlab编程,对各类S曲线数学模型进行最优参数求解,以最小拟合误差的模型作为分配曲线拟合的依据,而后根据E值和I值不变性,通过平移分配曲线得到各分选系统的产品预测模型,同时,将优化问题归结为分选密度最优值求解,构建三分选工艺的产品优化模型,并以潘家冲铅锌矿为例,在最大经济效益目标下,采用差分进化算法获取各个分选环节的最优分选密度和精矿产率及品位,可为生产工艺优化提供支撑。

In order to optimize the product structure under the established separation process conditions,we have proposed a mathematical method of data fitting,according to the product density and distribution ratio calculation,to fit the distribution curve. By using the differential evolution algorithm to made an optimization calculation of S curve mathematical model optimizing in Matlab programming,we took the mathematical model with the least fitting error as the basis for fitting the distribution curve. Based on the invariance of E and I value,the product prediction model of each sorting system was obtained by shifting distribution curve. Meanwhile,the optimization problem was reduced to the optimal solution of the separation density to construct the product optimization model of three separation process. We also took Panjiachong lead-zinc deposit as an example,under the maximum economic benefit target to get the optimal separation density. Yield and Fe-grade of concentrate are used in the optimal sorting magic differential evolution algorithm, which could provide effective support for the optimization of production.