摘要
针对磨矿过程中磨矿粒度实现在线实时测量难度较大,仅能通过事后化验,具有较大的滞后性的问题,引入一种线性生成机制(LGMS)、混沌搜索、粒子群优化算法(PSO)和变邻域搜索算法(VNS)修正果蝇算法(IFOA),然后利用IFOA良好的搜索全局最优解的能力自适应地调整BP网络的权值和阈值建立磨矿粒度在线软测量模型。最后以某公司样本数据为例进行仿真验证,结果表明其鲁棒性和测量精度明显提高,且网络具有较强的收敛性能。
Aiming at the problem that it was difficult to achieve real?time online measurement of grinding size,but only meas?ured it by off?line test,which has greater hysteresis,introduced a linear generation mechanism of candidate solution(LGMS),cha?otic search,particle swarm optimization algorithm(PSO)and variable neighborhood search algorithm(VNS)to modify fruit fly al?gorithm,then by using capability of searching the global optimal solution of modified FOA,the weight and threshold of BP neural network were adjusted adaptively,and established a soft?sensing model of grinding size.Finally,taking the sample data of a com?pany as an example,the results show that the robustness and measurement accuracy are improved obviously,and the network has strong convergence performance.
作者
杨刚
王建民
YANG Gang;WANG Jian-min(College of Electrical Engineering,North China University of Science and Technology,Tangshan 063000,China)
出处
《仪表技术与传感器》
CSCD
北大核心
2018年第8期122-126,共5页
Instrument Technique and Sensor
关键词
磨矿粒度
软测量
果蝇算法
混沌搜索
LGMS
粒子群算法
变邻域搜索算法
grinding size
soft measurement
fruit fly algorithm
chaotic search
LGMS
particle swarm optimization algorithm
varia.ble neighborhood search algorithm
