基于人工神经网络系统的选矿指导子模块的构建-----2015年6月份增刊
投稿时间:2015-03-18  修订日期:2015-03-18  点此下载全文
引用本文:
摘要点击次数: 392
全文下载次数: 1
作者单位E-mail
聂轶苗 河北联合大学矿业工程学院 nieym168@163.com 
中文摘要:由于影响选矿流程及选矿指标的因素较多,且因果关系往往呈非线性关系,使传统的数学建模方法难以利用。本文基于矿物工艺矿物学参数实现计算机自动识别和初步计算的基础上,提出了建立基于BP神经网络的选矿指导专家系统来解决这一问题,对选矿子模块进行了构建。重点研究了5个重要环节:子模块的基本结构、基本感知单元和多层感知器的构建、输入与输出的合理转变以及神经网络的功能构架。通过这几个方面的探讨和研究,最终建立了实用性很强的基于人工神经网络系统的选矿指导子模块并将其预测指标与实际选厂指标进行了对比,结果表明,该遗传神经网络具有较高的预测精度。此子模块可用来预测选矿流程或选矿试验的基础数据及产品指标,为优化选矿工艺流程奠定基础。
中文关键词:人工神经网络  选矿指导子模块  构建
 
Fig.1 Schematic diagram ofSsub-modulesSdressing guidance
Abstract:It was difficult to precast the quality of ore separation concentration because there were many factors to affect the flow sheet and processing targets. Based on the automatic identification and basic calculation of minerals using computer, ore separation expert system on artificial neural network was proposed in this paper and ore separation sub-modules were established. Five important links were researched, which was the main structure, basic sensing unit, multilayer perceptron, input-output convertor and the neural network function architecture. Through these researches, ore separation sub-modules were established with strong feasibility. Comparison the predict index to the actual value, it was found that the neural network was successful with high precision. These study and establishment of ore separation sub-modules based on artificial neural network provided basic data for the optimization of ore dressing process.
keywords:Artificial Neural Network  Ore Separation sub-modules  Establishment
查看全文  查看/发表评论  下载PDF阅读器
设备展厅
您是第5294953位访问者  备案许可证编号:京ICP备13036454号
版权所有:《中国矿业》杂志社有限公司
地  址: 北京市朝阳区安定门外小关东里10号院东小楼 邮编:100029 电话:010-68332570 88374940 E-mail: zgkyzzs@163.com