基于矿石中硅铝铁的品位对安徽省凤阳县灵山-木屐山矿区的矿石小体重进行预测  

Prediction of Small Body Weight of Ores in Lingshan Muchongshan Mining Area,Fengyang County,Anhui Province Based on the Grade of Silicon,Aluminum,and Iron in Ores

作　　者：马彤宇[1] MA Tong-yu(Geological Laboratory of the 313 Geological Team of Anhui Provincial Geological and Mineral Exploration Bureau,Lu'an 237010,China)

机构地区：[1]安徽省地质矿产勘查局313地质队地质实验室,安徽六安237010

出　　处：《世界有色金属》2024年第7期217-219,共3页World Nonferrous Metals

摘　　要：矿石小体重在资源储量估算过程中是一个重要参数,它的准确与否,直接影响矿床的经济评价和矿山储量估算结果。为减小矿石小体重的误差,在矿业技术及数字化时代的技术革新背景下,本文使用Python语言构建矿石中硅铝铁的品位与矿石小体重之间的BP神经网络预测模型,实现了对该矿区矿石小体重的预测,当均方误差损失值为0.0022时,此模型预测结果有着96.55%的准确率。通过对测量值和预测值的对比,筛选结果偏差大的样品进行重新测量,校验样品的小体重,以确保每个样品小体重值的可靠性和准确性。The small weight of ore is an important parameter in the process of resource reserve estimation,and its accuracy directly affects the economic evaluation of mineral deposits and the results of mine reserve estimation.In order to reduce the error of small weight of ore,under the background of technological innovation in mining technology and the digital era,this paper uses Python language to construct a BP neural network prediction model between the grade of silicon aluminum iron in ore and the small weight of ore,and realizes the prediction of small weight of ore in the mining area.When the mean square error loss value is 0.0022,the prediction result of this model has an accuracy of 96.55%.By comparing the measured values and predicted values,select samples with large deviations for re measurement,verify the small body weight of the samples,and ensure the reliability and accuracy of the small body weight values of each sample.

关 键 词：BP神经网络 矿石小体重 矿产储量 

分 类 号：P575[天文地球—矿物学]