面向窑压控制的基于改进STS模型的非线性广义预测控制研究

发布时间：2018-05-01 04:12

本文选题：窑压控制 + 简化T-S模型　；参考：《电子学报》2017年07期

【摘要】：窑压是玻璃窑炉运行过程中重要的被控指标之一,由于受到大量不确定因素影响,现有控制算法的跟踪超调大且波动频繁.鉴于简化T-S模型(STS)的强在线自适应学习能力和减法聚类可以解决最优模型结构的不稳定问题,本文研究一种基于减法聚类与简化T-S模型相融合的非线性模糊广义预测方法实现窑压的控制.利用该方法可以快速的完成窑压跳变后对象的逼近,克服由于跳变所引起的系统不稳定现象;在改进的STS模型的基础上,针对窑压阶跃跟踪超调过大的问题,结合整体最优滚动优化函数及最优控制增量快速求解方法设计控制律,从本源出发抑制超调,计算量小,并易于实施.
[Abstract]:Kiln pressure is one of the important control indexes in the operation process of glass kiln. Due to a large number of uncertain factors, the tracking of existing control algorithms is overshoot and fluctuates frequently. In view of the strong online adaptive learning ability of the simplified T-S model (STS) and the subtraction clustering can solve the problem of the optimal model structure, this paper studies one kind of problem. The nonlinear fuzzy generalized prediction method based on the fusion of subtractive clustering and simplified T-S model is used to control the kiln pressure. The method can quickly complete the approximation of the object after the kiln pressure jump and overcome the instability of the system caused by the jump. On the basis of the improved STS model, the overshoot overshoot is too large for the kiln pressure step tracking. The control law is designed by combining the overall optimal rolling optimization function and the optimal control increment fast solution method. The overshoot is suppressed from the source, and the calculation is small, and it is easy to implement.

【作者单位】：北京化工大学信息科学与技术学院;清华大学自动化系;
【基金】：国家自然基金(No.51375038) 高等学校博士学科点专项科研基金博导类(No.20130010110009) 吉林大学符号计算与知识工程教育部重点实验室开放课题(No.93K172014K05)
【分类号】：TQ171.623

【相似文献】