CAJ | 학술논문

乙烯精馏过程具有较强的非线性，常规PID及先进控制技术均难以取得较好的控制效果，而熟练的优秀操作工却可以获得较满意的控制效果。本文基于多Agent(MAS)控制策略，结合PID广义对象自适应预测控制算法(GMPC)，设计了乙烯精馏塔的智能PID广义对象自适应预测控制系统，同时解决了预测控制与规则知识控制的协调问题，既保留了预测控制的优点，又完全可以承担起将熟练操作工、技术人员和专家的知识经验、操作方法等应用于生产过程的操作和控制，实现乙烯精馏过程的高度自动化和智能化。
을희정류과정구유교강적비선성，상규PID급선진공제기술균난이취득교호적공제효과，이숙련적우수조작공각가이획득교만의적공제효과。본문기우다Agent(MAS)공제책략，결합PID엄의대상자괄응예측공제산법(GMPC)，설계료을희정류탑적지능PID엄의대상자괄응예측공제계통，동시해결료예측공제여규칙지식공제적협조문제，기보류료예측공제적우점，우완전가이승담기장숙련조작공、기술인원화전가적지식경험、조작방법등응용우생산과정적조작화공제，실현을희정류과정적고도자동화화지능화。
As a strong nonlinear process, ethylene rectifier isn't easy to be controlled by PID or advanced process control technology. However, excellent operators can do it better with their experiences. In this paper, an adptive predictive control system based on intelligent PID generalized object was designed with GMPC algorithm and MAS control strategy. And the coordinated problem between predictive control and rule knowledge was also resolved. This control system can develope some good qualities of model predictive control(MPC), anymore, it can undertake to put knowledge experiences and operation methods which owned by experienced operators, technicians and experts together to operate and control during operation process. It can realize highly automated and intelligent ethylene distillation process.