基于在线智能辨识的模糊专家控制方法及其应用

来源期刊：中南大学学报(自然科学版)2004年第6期

论文作者：李劼 张文根 丁凤其 刘业翔

文章页码：911 - 914

关键词：铝;电解;过程控制;智能控制;模糊控制;模糊专家控制

Key words：aluminum; electrolysis; process control; intelligent control; fuzzy control; fuzzy expert control

摘    要：提出一种综合应用在线智能辨识、多控制模式与自决策机制以及变论域技术的新型模糊专家控制方法,并应用于铝电解槽氧化铝浓度等重要工艺参数的实时控制。基于该控制方法的智能控制器在实施控制的同时,对被控对象施加“激励”并通过模糊推理对可测变量与不可测的被控变量间的特征关系(特征曲线)及工作点在特征曲线上的位置进行在线智能辨识,同时依据辨识结果自动调整模糊变量的论域和进行控制模式选择,实现自适应、自学习功能并增强鲁棒性。在铝电解槽的应用结果表明,采用该控制方法能较好地控制氧化铝浓度。

Abstract: A new fuzzy expert control method was brought forward by synthetically using the techniques of online intelligent identification, multi-control-mode, self-decision mechanism and self-adjustment of universe of discourse. The method was successfully used in the real-time control of important technical parameters such as alumina concentration in aluminum reduction cells. While carrying on its control functions, the intelligent controller based on the method can put“exciting signal” on its control object, carrying out the on-line intelligent identification of the characteristic relationships (the characteristic curves) between the measurable variables and the immeasurable ones as well as the locations of the operating points on the characteristic curves, automatically adjusting the universes of discourse of fuzzy variables and deciding the choice of control modes according to the results of the on-line intelligent identification, thus, realizing the functions of self-adaptation and self-study and increasing the robustness. The application in aluminum reduction cells shows that the control method acquires very satisfactory effects on the control of alumina concentration.