冲天炉熔炼过程粒群优化  

Particle Swarm Optimization of Cupola Melting Operation

作　　者：夏伯才[1] 王永强[1] 董杰[2] 

机构地区：[1]中国工程物理研究院工学院,四川绵阳621900 [2]中国工程物理研究院物资部,四川绵阳621900

出　　处：《教学与科技》2005年第3期1-8,共8页Teaching and Science Technology

基　　金：本文受中国工程物理研究院科学技术基金（编号：20000329：20040660）资助

摘　　要：粒群优化是一种非梯度随机优化算法，其思想源于动物群体（如群落）社会动力学行为的最近邻速度匹配和根据距离加速等基本规则。本文综合自适应模糊推理的建模功能和神经网络的学习能力，直接从实验数据中提取推理规则，建立了冲天炉熔炼过程模型。模型具有较高的预测精度和泛化能力，利用它可以帮助操作者认识熔炼规律。同时，将自适应模糊推理模型与粒群优化算法耦合，在预定熔化率和炉温的模糊限制条件下，得到了最高热效率时的送风强度和焦耗。此法可推广应用到其它工艺过程的建模与优化上。Particle swarm optimization is a stochastic gradient free optimization algorithm. The ideas that underlie it are inspired from the social dynamics of flocking organisms, such as swarms, which are governed by fundamental rules like nearest-neighbor velocity matching and acceleration by distance. By the coupled use of adaptive fuzzy inference modeling and artificial neural network learning ability, a set of operating rules, which could help us better understanding the basic principles of cupola melting operation, have been generated directly from the experimental net diagram data. The developed fuzzy inference systems could map the relationships between operating parameters accurately. A new type of diagram is proposed to simplify the determination of optimal blast of and carbon rates by the cupola supervisor. By integrating in the particle swarm optimization procedure, the fuzzy inference systems are used to determine the optimal blast and carbon rate for the lowest energy consumption. The methodology present here could be used for other data based process modeling and optimization practices.

关 键 词：冲天炉 网形图 自适应模糊推理 人工神经网络 粒群优化 随机优化算法 冲天炉熔炼 熔炼过程 模糊推理模型 过程模型 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TG232.1[自动化与计算机技术—控制科学与工程]