预测控制的哲学思想分析及其新架构的提出

发布时间：2019-06-05 17:05

【摘要】：预测控制属于控制科学,从"过去-现在-将来"的哲学视角看其产生与发展存在必然性。预测控制的性能指标是优化求解预测控制的核心,其结构决定预测控制系统的功能。由于系统优化要求,常规预测控制系统的性能指标由被控变量跟踪项与控制能量变化项组成,但仍不全面。本文从马克思主义唯物辩证法出发,由事物普遍联系的规律提出了理应在性能指标结构中引入表征操作变量与约束边界之间距离的函数,来规避由于预测控制系统与外界联系的约束性对控制造成不利的影响;从全局与局部的角度出发,提出了预测控制的多变量控制与区间控制的特性可以为预测控制的结构改进提供前提;从绝对与相对的辨证范畴进一步探讨了性能指标结构的改进结果。基于上述分析,从马克思主义认识论出发,将预测控制性能指标的改进方案与化工过程中裕量的概念相结合,提出了考虑裕量损失函数的预测控制新架构。最后,将理论结合实践,通过仿真实例验证了预测控制新架构的效果。
[Abstract]:Predictive control belongs to control science, and its emergence and development are inevitable from the philosophical perspective of "past-present-future". The performance index of predictive control is the core of optimal solution predictive control, and its structure determines the function of predictive control system. Due to the requirements of system optimization, the performance index of conventional predictive control system is composed of controlled variable tracking term and control energy change term, but it is still not comprehensive. Based on Marxist materialist dialectics and the law of universal connection of things, this paper puts forward a function that should be introduced into the structure of performance index to characterize the distance between operational variables and constraint boundaries. To avoid the adverse effects of the constraints of the predictive control system on the control due to the connection between the predictive control system and the outside world. From the global and local point of view, it is proposed that the multivariable control and interval control characteristics of predictive control can provide the premise for the structural improvement of predictive control. The improved results of performance index structure are further discussed from the absolute and relative dialectical categories. Based on the above analysis and Marxist epistemology, a new predictive control framework considering margin loss function is proposed by combining the improved scheme of predictive control performance index with the concept of margin in chemical process. Finally, the theory is combined with practice, and the effect of the new predictive control architecture is verified by a simulation example.
【作者单位】：中国石油大学(北京)自动化系;
【基金】：国家自然科学基金项目(21676295)
【分类号】：TP273

【相似文献】