原矿浆制备过程的智能优化设定方法及应用

来源期刊：冶金自动化2009年第4期

论文作者：曾宇 李强 陈建

文章页码：5 - 8

关键词：优化设定； 案例推理； 专家规则； 原矿浆； 生产指标

Key words：optimizing setting； case-based reasoning； expert rule； ore pulp； production index

摘    要：针对中国铝业山西分公司氧化铝分厂原矿浆制备过程机理复杂、关键被控变量的优化设定值不易获得从而使生产指标难以控制在期望目标值范围内的问题,将案例推理和专家系统技术相结合,提出一种多变量智能优化设定方法。预设定模型根据生产指标的目标值和工况的变化给出控制回路的预设定值,基于专家规则的修正模型根据实际工况的变化对预设定值进行校正,从而实现了控制回路设定值的在线自动调整。研发了基于分布式控制系统的原矿浆制备过程设定软件,此系统应用于原矿浆制备过程的生产实际,能够适应工况的频繁变化,自动更新回路设定值,从而实现了生产指标的优化控制,取得了明显成效。

Abstract: Because optimizing setting values of key controlled variables were difficult to find in preparation process of ore pulp,and production indices were difficult to be controlled within target ranges,a multivariable intelligent optimizing setting method is presented combining case-based reasoning(CBR) with expert system.Presetting values of control loops are given by presetting model according to production target and change of working conditions,and corrected by revising model based on expert rules according to change of working conditions,so online automatic adjustment of control loop setting values was realized.Setting software of pulp preparation process based on distributed control system was developed.Industrial application has proved that it can adapt frequently changed working conditions and update automatically loop setting values.Optimizing control of production index was achieved and obvious effect was obtained.