Predictive Control In Networked Control Systems To Achieve

Networked control systems (NCSs) are new distributed and real time feedback control systems. The primary advantages of NCSs are the sharing of the information, framework simplification, ease of diagnosing and maintaining. Ethernet is selected as the research background based on the analysis of three control buses, which are Ethernet, ControlNet and DeviceNet. Predictive control algorithm is introduced as the research methodology based on the survey of recent NCS control methodologies.Three predictive control methodologies are designed in this thesis for different plants in NCSs. Generalized predictive control (GPC) + queuing strategy (QS) method is designed for linear plants. While nonlinear predictive control (NPC) + QS method and approximate predictive control (APC) + QS method are designed for nonlinear plants. In local area network of Zhejiang University, two actual NCS experiments are implemented. One is a linear system made of electrocircuit, and the other is a nonlinear system of tank level.The proposed methods have several advantages inherited from GPC. Good performance of the methods is showed in simulations and experiments.