磨矿破裂过程的蒙特卡洛仿真方法研究  

Research on Monte Carlo Simulation of Grinding Breakage Process

作　　者：卢绍文[1] 

机构地区：[1]东北大学流程工业综合自动化国家重点实验室,辽宁沈阳110819

出　　处：《东北大学学报（自然科学版）》2014年第6期770-773,808,共5页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(61240012);中央高校基本科研业务费专项资金资助项目(N120408003)

摘　　要：研究了磨矿破裂过程的蒙特卡洛仿真方法.蒙特卡洛仿真的误差反比于其计算区域内颗粒数量,导致该方法存在计算精度和计算代价难以协调的矛盾.主要研究目标是在不降低精度的条件下提高仿真速度.首先给出了事件驱动的磨矿过程蒙特卡洛仿真的主要步骤.通过提出一种新的包络函数改进仿真抽样效率.设计了批次磨矿的仿真实验.仿真结果表明:所提出的改进算法能够显著降低无效抽样,提高仿真速度,而且保持仿真精度符合要求.Monte Carlo simulation for breakage process during grinding was studied. The error of Monte Carlo simulation was in reverse proportional to the number of particles in the computation box, which caused the conflicting between the computation accuracy and cost. The main target of this work was to accelerate simulation while keeping its accuracy. First, main steps of the eventdriven Monte Carlo simulation were listed. Then, an improved sampling method was proposed to tackle the problem of the poor performance of traditional rejection sampling. Simulation experiments of batch grinding were designed. The results showed that the proposed algorithm could obviously reduce invalid sampling, and increase the simulation speed up to 5 times while keeping good simulation accuracy.

关 键 词：磨矿过程 破裂过程 蒙特卡洛仿真 舍取法 仿真加速 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]