基于WOA-BP算法的自动压滤机脱水指标预测模型研究  

Optimization of Dewatering Process of Automatic Filter Press Based on WOA-BP Algorithm

作　　者：刘惠中[1,2] 闻成钰 曾聪 万小青 王朔 LIU Huizhong;WEN Chengyu;ZENG Cong;WAN Xiaoqing;WANG Shuo(School of Mechanical and Electrical Engineering,Jiangxi University of Science and Technology,Ganzhou 341000,Jiangxi,China;Jiangxi Province Engineering Research Center for Mechanical and Electrical of Mining and Metallurgy,Ganzhou 341000,Jiangxi,China;Ganzhou Nonferrous Metallurgical Machinery Co.,Ltd.,Ganzhou 341000,Jiangxi,China)

机构地区：[1]江西理工大学机电工程学院,江西赣州341000 [2]江西省矿冶机电工程技术研究中心,江西赣州341000 [3]赣州有色冶金机械有限公司,江西赣州341000

出　　处：《有色金属（选矿部分）》2024年第9期72-79,共8页Nonferrous Metals（Mineral Processing Section）

基　　金：江西省“双千计划”引进高层次创新人才项目(jxsq2018101046)。

摘　　要：随着全球工业化的不断发展,矿山的开采规模正在不断扩大,导致矿物资源逐渐贫化,细杂等难选矿物资源越来越多。选矿磨矿粒度越来越细,导致矿物分选后产品的脱水过滤越来越困难。为保证后续运输和冶炼工序对精矿含水率的生产需求,需要使用自动压滤机对精矿进行高效率的脱水处理。在精矿的过滤脱水过程中,影响自动压滤机脱水效率的因素众多。为更好地对脱水过程及生产指标进行控制,基于鲸鱼算法WOA优化的BP神经网络构建了一种WOA-BP神经网络模型,以入料浓度、入料时间、压榨时间、风干时间等4项影响脱水指标的因素为输入因子,以滤饼含水率和单位面积每小时处理量为输出因子,建立了脱水指标的预测模型,并对比分析单一BP神经网络模型和WOA-BP神经网络模型。结论如下:WOA-BP预测模型对滤饼含水率和单位面积每小时处理量的平均绝对误差分别为4.98%、8.83%,均方根误差分别为0.86%、3.43%,与单一的BP神经网络预测模型相比,该预测模型预测误差明显小于单一BP神经网络预测模型,脱水指标的预测结果更接近实测值,具有较高精确度。利用构建的WOA-BP预测模型,可以有效预测压滤机的脱水过滤指标,为后续对脱水过程的控制进行优化奠定了基础。With the continuous development of global industrialization,the mining scale of mines is expanding,resulting in the gradual depletion of mineral resources,more and more fine impurities and difficult-to-process mineral resources.Beneficiation and grinding products are getting finer and finer,resulting in more and more difficult dewatering and filtration of the products after mineral sorting.In order to ensure that the subsequent transportation and smelting process does not affect the water content of the concentrate,the use of an automatic filter press is required to carry out high-efficiency dewatering treatment of the concentrates.In the process of filtration and dewatering concentrate,there are many factors that affect the dewatering efficiency of an automatic filter press.In order to better control the dewatering process and production indexes,a WOA-BP neural network model based on the BP neural network optimized by Whale Algorithm WOA is constructed and a prediction model of dewatering indexes by taking four factors affecting the dewatering indexes,i.e.feed pulp density,feed time,pressing time,and air-drying time,as the input factors,and the water content of the filter cake and the hourly processing capacity of the unit area as the output factors is established.A single BP neural network model and a WOA-BP neural network model were compared and analyzed.The conclusions are as follows:The average absolute error of the WOA-BP prediction model is 4.98%and 8.83%for the water content of the filter cake and the hourly processing capacity of the unit area,and the root-mean-square error is 0.86%and 3.43%respectively.Compared with the single BP neural network prediction model,the prediction error of this prediction model is obviously smaller than that of the single BP neural network prediction model,and the prediction of dewatering indexes is closer to the actual measured value with high accuracy.The prediction results of dehydration indexes are closer to the actual measured values and have high accuracy.The construct

关 键 词：脱水效率 BP神经网络模型 鲸鱼算法 指标预测 

分 类 号：TD462.2[矿业工程—矿山机电]