基于EDEM与BP神经网络的平-摆筛参数化研究

Research on Flat-Pendulum Sieve Parameterization Based on EDEM and BP Neural Network

为直观分析振动参数与料体分层效果复杂映射关系及筛分过程物料松散规律,以便振动筛参数化研究的开展。首先以振动筛简化后三维模型导入EDEM软件,利用正交试验法进行设计,并将试验方案中对应筛机参数分别导入EDEM中。基于平动与摆动双运行模式振动筛,提出以分层沉降比的定义对料体分层效果进行表征,以离散元软件分析振动筛的振幅、振动频率、振动方向角和摆动频率四个参数对分层效果的影响,将获得的仿真数据导入BP神经网络进行深度学习。采用集成学习的训练集对训练好的模型进行筛机参数影响权重的分析,得出振动频率、振动幅度,振动方向角和摆动频率和对分层效果有较大影响,故以这四种参数组合表征振动筛运行状态,对不同参数组合对应筛分效率进行分析。振动频率与振动幅度增加,可以增大筛面的振动强度,不仅加速物料松散速率,优化物料的分层效果,又可以改善物料的堵筛情况,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,分层效果下降;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,弱化物料的分层效果,从而影响筛分效果。当振幅为3.6mm、振频为18.3Hz、振动方向角为39.1°、摆动频率为13.8Hz时,振动筛的分层沉降比为0.77,分层效果最优,此时的分层构型最便于筛分的进行,此研究对平动与摆动双运行模式的振动筛优化设计具有一定指导意义。

In order to intuitively analyze the complex mapping relationship between vibration parameters and material body layering effect and the loose law of materials in the screening process,so as to carry out the research on vibrating screen parameterization.Firstly,the 3D model of the simplified vibrating screen is imported into the EDEM software,the orthogonal test method is used to design the test,and the corresponding sieve machine parameters in the test scheme are imported into EDEM respectively.Based on the translational and swinging dual operation mode vibrating screen,the layering effect of the material body by the definition of layered settlement ratio is proposed to characterize,and the influence of the four parameters of amplitude,vibration frequency,vibration direction angle and swing frequency on the layering effect of the vibrating screen is analyzed by discrete element software,and the simulation data into the BP neural network is imported for deep learning.The training set of ensemble learning was used to analyze the weight of the influence of sieve machine parameters on the trained model,and the vibration frequency,vibration amplitude,vibration direction angle and swing frequency had a great impact on the layering effect,so the four parameter combinations were used to characterize the vibrating screen operating state,and the corresponding screening efficiency of different parameter combinations was analyzed.The increase of vibration frequency and vibration amplitude can increase the vibration intensity of the sieve surface,which not only accelerates the loosening rate of the material,optimizes the layering effect of the material,but also improves the blocking of the material,but when the vibration intensity is too large,the material transition time is too long,so that the contact time between the material and the sieve surface becomes shorter,and the layering effect decreases.With the increase of the vibration direction angle,the spreading rate of the material can be accelerated,but when the angle is too large,it will reduce the collision probability between the material and the sieve surface,weaken the layering effect of the material,and thus affect the screening effect.When the amplitude is 3.6 mm,the vibration frequency is 18.3 Hz,the vibration direction angle is 39.1°,and the swing frequency is 13.8 Hz,the layered sedimentation ratio of the vibrating screen is O.77,and the layering effect is the best,and the layered configuration at this time is the most convenient for screening,which has certain guiding significance for the optimal design of the vibrating screen in translation and swing dual operation modes.