基于PSO-BP神经网络的煤矿双层振动筛运动参数研究  

Research on Motion Parameters of Double-Layer Vibrating Screen in Coal Mine Based on PSO-BP Neural Network

作　　者：张晋霞 王秋月 牛福生 ZHANG Jinxia;WANG Qiuyue;NIU Fusheng(College of Mining Engineering,North China University of Science and Technology,Tangshan 063210,Hebei,China)

机构地区：[1]华北理工大学矿业工程学院,河北唐山063210

出　　处：《有色金属(选矿部分)》2025年第3期83-90,共8页Nonferrous Metals(Mineral Processing Section)

基　　金：国家自然科学基金资助项目(52374265);河北省自然科学基金资助项目(E2022209108);中央引导地方科技发展资金项目(236Z4106G)。

摘　　要：为提高振动筛筛分效率,利用BP神经网络的高度非线性映射、自学习、自组织和联想记忆等功能,避免了繁琐的数学理论模型,通过筛分效率与各参数之间的联系,开展振动筛的参数化研究。首先建立振动筛三维模型,设计正交试验,在离散元软件EDEM中导入相关的筛机参数。基于PSO-BP神经网络的双层振动筛运动参数研究是通过观察振动方向角、振幅、频率三个参数对双层振动筛筛分效率的影响,将获取的仿真数据导入到BP神经网络中进行深度学习。采用集成学习的训练集对训练好的模型进行筛面参数分析,得出频率、振幅、振动方向角对筛分效率有较大影响,故以这三种参数组合表征振动筛运动状态,对不同参数组合对应筛分效率进行分析。频率与振幅增加,可以增大筛面的振动强度,加快物料松散速率,还可以优化物料的筛分效率,但振动强度过大时,物料跃迁时间过长,使得物料与筛面接触时间变短,筛分效率下降;随着振动方向角的增加可以加速物料的铺展速率,但角度过大时,会降低物料与筛面的碰撞概率,弱化物料的筛分效率,从而影响筛分效果。最后通过粒子群算法优化BP神经网络得到当振幅为3 mm、振频为20 Hz,振动方向角为30°时筛分效率为82.91%,筛分效果最优,此研究对振动筛参数优化及煤矿筛分过程具有一定的指导意义。To enhance the efficiency of the vibrating screen,a BP neural network is utilized to circumvent the need for complex theoretical mathematical models.This includes highly nonlinear mapping,self-learning,self-organizing,and associative memory,facilitating the investigation of vibrating screen parameterization.The relationship between screening efficiency and parameters has been established.Firstly,the pulsating screen's threedimensional representation is reduced,an orthogonal test is created,and all related panel parameters are put into the discrete element modeling initiative,in which EDEM.The objective of this study is to examine the effects of the vibration's frequency range,magnitude,and location on the monitor effectiveness of a double-layer vibrating screen that utilizes a PSO-BP neural network.The simulation data obtained will be fed into a BP neural network for further analysis and deep learning.The integrated learning training set is then used to analyze the screen surface parameters of the trained model.The screening efficiency is found to be significantly influenced by the vibration direction angle,vibration frequency,and vibration amplitude.Consequently,the three-parameter combinations are used to characterize the operating state of the vibrating screen and analyze the screening efficiency corresponding to different parameter combinations.Increasing the vibration's frequency and amplitude may enhance the viewing surface's disturbance strength.This not only accelerates the material loosening rate and optimizes the material screening efficiency but also improves the material blocking.When the vibration strength is excessive,the material transition time is prolonged,and the contact time between these materials and the picture is reduced,leading to a decrease in the quality of screening.As vibrations occur orientation angle decreases,and the material's spreading rate can be accelerated.However,when the angle is too large,the collision probability between the screening becomes less effective,due to a decreas

关 键 词：双层振动筛 BP神经网络 煤矿振动参数 粒子群算法 

分 类 号：TD452[矿业工程—矿山机电]