Font Size: a A A

Study On Several Control Problems In Distributed Networked Control Systems

Posted on:2009-01-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:X ZhangFull Text:PDF
GTID:1118360272978587Subject:Power electronics and electric drive
Abstract/Summary:PDF Full Text Request
Along with modern industry developing rapidly, with the production of society and the level of science and technology improving continually, the scale of process system of industry get more huge, the system function get more complex and the geographical distribution of system get more dispersive. The rapid development of process system of industry induces that the scale of control problem get more huge, the structure of control system get more complex and the control requirement put forward get more high, traditional control structure,theory and method haven't adapted to the requirement of the large scale system in modeling,analysis,synthesis and design. The distributed networked control system integrates computer, control, communication and network technology etc . , adapts to the developing trend of industrial control system in decentralization, network and intelligence etc., becomes a hotspot of industrial automatization and achieves all kinds of functions in synthetical automatization. The distributed networked control system is studied from the direction of control, in the round and in depth.Firstly, analyzing the structure of typical distributed networked control system, for example, the distributed control system, the fieldbus control system and the industrial Ethernet control system etc. , the uniform structural model of the distributed networked control system is built, and the model is the base of the system study. The basic structural model of the distributed networked control system consists of five elements of task of process, a whole control chain is formed by sensor node, link of sensor to controller, controller node, link of controller to actuator, actuator node, in turn. The transferred behavior of state of basic structural mode is analyzed and studied from both a control and a computer engineering perspective. The characteristic of transmit delay of forward channel and feedback channel, control performance how to be influenced and bad result how to be eliminated are studied from control engineering perspective. The problem of delay of deadline and jitter in control scheduling are studied from computer engineering perspective. The problem of delay that is focus in the distributed networked control system is analyzed and studied from control delay, control period, delay jitter and transmit error of data, thus, the delay how to be produced and influenced is discussed in detail. Because the delay of the distributed networked control system is key to system, accurate model of delay becomes precondition to control. For obtaining the model of delay, the experiments that measure the characteristic of transmit delay of CAN bus and Ethernet are introduced in detail, the metrical data of delay under various load are analyzed and processed and three model of delay applied in various condition are obtained. These delays are constant delay, independent random delay between two transfers and correlative random delay based on a Markov chain between two transfers, respectively. Under discrete time domain, the mathematic model of control object of the distributed networked control system driven by time and event are presented, respectively, and the criteria of control performance needed by designing the parameters of control system are presented, too. Based on the above-mentioned, The domestic and overseas status of study and work on the distributed networked control system are introduced, in brief.Secondly, Based on the problem of delay of the distributed networked control system and basic structural model of the system, the mathematic model of networked generalized control object made up by the part of communication and control object in the distributed networked control system is built. The distributed networked control system is divided into two kinds: one is the double channel delay system that the forward channel and the feedback channel of control system have transferred delay, the other is the feedback channel system that the feedback channel of control system has transferred delay. Two mathematic models of the distributed networked control system are built from two directions, respectively: one is time delay system, the other is large delay system . Under continuous time domain and discrete time domain, two mathematic models are built for two systems, respectively. For simplifying system model and easy to analyze system, two systems work on time driven mode in discrete time domain, namely sensor, controller and actuator are driven by time. Based on the first mathematic model of the distributed networked control system, for output feedback control, existent theorems of stability and designs of parameter of stable controller are presented. Thirdly, the system is designed based on the second mathematic model of the distributed networked control system. Smith predicator control and optimizing control are put forward. For Smith predicator control, designing controller for output feedback control and state feedback control, output feedback controller based on Smith predicator and state feedback controller based on Smith predicator with state estimation are obtained, respectively. For optimizing control, equivalent transform of LQ criterion between delay system and non-delay system is presented. Based on above matter, optimizing control of infinite time is designed for the distributed networked control system. Optimizing control of communication and control co-design for networked control system is put forward. Delayed input is transformed to extended state, analysis of stability and design of state feedback control are achieved. Jump system model is built for networked control system in the presence of packet losses, analysis of stability and design of state feedback control are achieved based on the model. Control of input multirate for networked control system is studied.Lastly, the T-S fuzzy model with uncertain state and uncertain, delayed input is built for the nonlinear distributed networked control system. The delayed input of the model is compensated by Smith predicator, then, the design of state feedback control is achieved based on the model without delay. The design of T-S fuzzy time-delayed control is put forward based on the model.
Keywords/Search Tags:the distributed networked control system, transmitted delay, time delay system, Smith predicator, optimizing control, T-S fuzzy control
PDF Full Text Request
Related items