摘要
针对目前筛分理论的研究仅局限于单因素考虑的问题,提出应用遗传算法(GA)优化的BP神经网络对数据空间进行全局寻优,且考虑所有因素对筛分结果的综合影响.首先,通过离散单元法的筛分仿真试验,获取实际筛分过程中难以获取的数据.然后,利用GA优化的BP神经网络对平摆复合振动筛的振动参数进行优化,选择5-9-1的BP神经网络结构类型,得到优化后的振动参数组合,即振幅为2 mm,振动频率为26Hz,振动方向角为46°,摆动频率为21Hz,摆角为1°.对优化后的结果进行一次模拟仿真验证,结果表明:验证结果与测试结果相吻合.
Aiming at the problem that the current screening theory was limited to one factor,a BP neural network optimized by genetic algorithm（GA）was proposed for global optimization of data space when the effects of multiple factors on screening results were considered.Firstly,the data difficult to obtain in the actual sieving process was obtained by using the simulation experiment based on discrete element method.Then,the BP neural network with structure of type 5-9-1 optimized by GA was adopted to optimize the vibration parameters of vibrating screen of translation-swing composite motion.The vibration parameters after optimization were as follows：vibration amplitude 2 mm,vibration frequency 26 Hz,vibrating direction angle 46°,swing frequency21 Hz,and swing angle 1°.Finally,the optimized results were verified by simulation experiment.The results show that the simulation experiment results are in good agreement with test results.
作者
沈国浪
童昕
李占福
SHEN Guolang;TONG Xin;LI Zhanfu(College of Mechanical Engineering and Automation,Huaqiao University,Xiamen 361021,China;Fujian Key Laboratory of Digital Equipment,Fujian University of Technology,Fuzhou 350108,China)
出处
《华侨大学学报（自然科学版）》
CAS
北大核心
2018年第4期509-513,共5页
Journal of Huaqiao University(Natural Science)
基金
国家自然科学基金资助项目(51175190)
福建省科技创新平台资助项目(2014H202)
福建省高校自然科学青年基金重点资助项目(JZ160460)
华侨大学研究生科研创新基金项目(1601103005)
关键词
振动筛
振动参数
离散单元法
筛分效率
遗传算法-BP神经网络
vibrating screen
vibration parameter
discrete element method
screening efficiency
genetic algorithm-BP neural network
作者简介
通信作者：童昕（1964-），男，教授，博士，主要从事机电系统动态分析与控制的研究．E-mail：xtong@fjut．edu．cn．
