Modular distributed model predictive control of nonlinear neighbor-affine systems

非线性邻域仿射系统的模块化分布式模型预测控制

基本信息

批准号：
310613429
负责人：
Professor Dr. Knut Graichen
金额：
--
依托单位：
Lehrstuhl für Regelungstechnik
依托单位国家：
德国
项目类别：
Research Grants
财政年份：
2016
资助国家：
德国
起止时间：
2015-12-31 至 2018-12-31
项目状态：
已结题

来源：
https://gepris.dfg.de/gepris/projekt/310613429?language=en
关键词：
Modular distributed model predictive control

项目摘要

The steadily growing demands on efficiency and flexibility of modern automation and control systems requires a broader design approach for the overall system that goes beyond the isolated look at and control of single subsystems. Decentral and distributed control schemes follow this holistic design approach by including the interdependencies between the subsystems in the control design.Model predictive control (MPC) appears to be a suitable control approach to tackle these kind of systems. In essence, MPC relies on the numerical solution of a finite-horizon dynamic optimization problem that is repetitively solved according to the sampling rate of the system. An extension of MPC to coupled systems is distributed MPC (DMPC), which assigns a single communicating MPC agent to each subsystem. The goal of the proposed project is to develop a DMPC scheme for nonlinear coupled systems, where each MPC agent contains a neighborhood model that anticipates the dynamical behavior of its neighbors in order to enhance the convergence and robustness of the distributed algorithm. Besides the development and mathematical investigation of the methodology, a further goal of the project is the numerical and experimental realization of the control approach. A particular intention of the project is to develop a modular framework that allows for an easy configuration and adaptation of the couplingstructure for suitable system classes. To achieve these mathematical, numerical and experimental goals, the proposed work programme buildsupon own preliminary work that is detailled in the proposal.

现代自动化和控制系统对效率和灵活性的需求不断增长，需要对整个系统采用更广泛的设计方法，而不仅仅是单个子系统的孤立查看和控制。分散式和分布式控制方案通过在控制设计中包含子系统之间的相互依赖性来遵循这种整体设计方法。模型预测控制（MPC）似乎是处理此类系统的合适控制方法。本质上，MPC 依赖于有限范围动态优化问题的数值解，该问题根据系统的采样率重复求解。 MPC 对耦合系统的扩展是分布式 MPC (DMPC)，它将单个通信 MPC 代理分配给每个子系统。该项目的目标是为非线性耦合系统开发一种 DMPC 方案，其中每个 MPC 代理都包含一个邻域模型，该模型可以预测其邻居的动态行为，以增强分布式算法的收敛性和鲁棒性。除了该方法的开发和数学研究之外，该项目的另一个目标是控制方法的数值和实验实现。该项目的一个特别目的是开发一个模块化框架，允许轻松配置和调整适合合适系统类的耦合结构。为了实现这些数学、数值和实验目标，拟议的工作计划建立在提案中详细介绍的自己的初步工作的基础上。