基于Matlab和粒子群算法的磨矿技术效率预测模型

Grinding Technical Efficiency Prediction Model Based on Matlab and Particle Swarm Optimization

研究了磨矿时间、干矿质量分数和充填率对锡石多金属硫化矿磨矿技术效率的影响.结果表明,在最优的磨矿参数条件下,即磨矿时间为8 min、干矿质量分数为65%、充填率为42%时,锡石和硫化矿二元结构所对应的磨矿技术效率最佳.通过Matlab的广义回归神经网络(GRNN)计算程序建立了一种磨矿技术效率预测模型,利用粒子群算法对模型参数进行优化,并通过试验验证了模型的适用性和可靠性.

The effect of grinding time,dry ore mass fraction and filling rate on the grinding technical efficiency of cassiterite polymetallic sulfide ore were studied. The results showed that the grinding efficiency corresponding to the binary structure of cassiterite and sulfide ore is the best when the grinding time is 8 min,the dry ore mass fraction is 65%,and the filling rate is 42%. A grinding technical efficiency prediction model was established by using the generalized regression neural netw ork( GRNN) program of M atlab. The model parameters were optimized by the particle sw arm optimization. The applicability and reliability of the model were verified by experiments.