Font Size: a A A

Research On Advanced Control And Optimization Of Complex Thermal Processes

Posted on:2009-05-30Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y ZhangFull Text:PDF
GTID:1102360245494935Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
Focusing on the complex thermal processes, aiming at the hotspot problems of the controlling and optimization investigation, The dissertation includes two parts of research contents. The first one is robust stability analysis of predictive PID controller, based on first-order plus dead-time model, second-order plus dead-time model and non-minimum phase model, are proposed respectively. The theoretic analysis and simulation results demonstrate that this control algorithm has better robustness than PID control algorithm when the process parameters are far away from the normal value.The ranges of largest process parameters are given when the system's parameters are uncertainty. The second one is the research of optimizing the set-points of supervisory level, which deal with the designing of the two-level controller and the optimization analysis of the set-point.Industry automatization has come into computer ages. Just as the Computer Integrated Processes System (CIPS). The boundary of the process system automatization was enlarged greatly. That is change from the basic controlling and advanced controlling to the supervisory controlling, manufacture project and attemper, information management and decision making. The assignment of the automation has been transformed from the stand-alone mode to the integration automatic mode, which is multilayer, multi-pattern, multi-view, for the purpose of the development of the information and the economy. This is a bran-new philosophic theory and idea.Referring to literatures, two aspects of work above mentioned are put forward in this paper, which accords to current research direction of control theory and it is the hotspot and the difficult spot in the field. The main contents and innovation points of this include:1,The actual thermal industry processes are always in the uncertain condition and they are usually affected by some beforehand undiscovered incertitude factors, the matching is not always exist between the model and the controlled object. The best control value based on the model perhaps will lead the performance of the system to worse when it was used in the actual process. Therefore it is necessary to study the robustness of the Predictive PID Control. The input and output robust stability analysis of predictive PID control is proposed, aimed to the first-order plus dead-time model, by using Kharitonov theorem and Edge theorem. The ranges of largest process parameters are given when the system's parameters are uncertainty. The theoretic analysis and simulation results demonstrate that this control algorithm has better robustness than PID control algorithm when the process parameters are far away from the normal value.2,Because some complex thermal processes can not be replaced by the simple first-order plus dead-time system, in this paper, aimed to the second-order plus dead-time model, by using Kharitonov theorem and Edge theorem. The ranges of largest process parameters are given when the system's parameters are uncertainty. The theoretic analysis and simulation results demonstrate that this control algorithm has better robustness than PID control algorithm when the process parameters are far away from the normal value.3,The non-minimum phase model is representative in the industry field, such as the feedwater system of the boiler. Therefore, it is very important to study the robust stability of the non-minimum phase system. In this paper, the robust stability of the predictive PID system is researched, which provides a certain directive effect for the design of DCS system. Under the circumstance of the uncertainty of the parameters, there is not a better method to study the Schur stability of the system. In this paper, the discrete characteristic polynomial can be transform into the continuous characteritic polynomial by using the double linear transform. And the Hurwitz stability can be validate by using Kharitonov theroy and the edge theory to this continuous characteritic polynomial.4,Under the different production condition point of view from the control, it is the optimization contents, such as adequately exerting the equipment's most potential, minimally consuming the energy sources and raw materials, getting the tiptop high yield guaranteeing the quality. Whereas, processes optimization relates to a set or a plant, so a great deal of information is necessary, the complex model should be set up and a great lot of operation should be carry through. All of these can be realized by the supervisory level. In the control level, the general PID control is replaced by predictive PID strategy and the general predictive control is used in the supervisory level. The general predictive control has the function of display the intending dynamic action of the system, the information is supplied by using the predictive model and then the input is determined, the object is to make the ouput of the process close to the anticipative result. The good control performance and the economy objective have verified by applying this method to a boiler system.5,There being severe load variation, the boiler variables(such as the steam pressure and the water level) under go some major fluctuations. Such fluctuations are difficult to minimize using classical control techniques. A two-level structure control method has been put forward in this study to improve the performance of the boiler response during such disturbance, in which the direct level consists of a predictive PID controller; and the set-points of the drum steam pressure and the water level deviation are modified whenever necessary by a supervisor partially based on fuzzy theory according to the current operation condition. The good control performance and the economy objective have verified by applying this method to a boiler system.6,Generally, in the industry manufacture processes, the set-points of the loop are decided by the operator according to the experience. As a result, the set-point is very random. But the quality, benefit of the processes is connected closely with the proper set-point. The optimization of the set-point is very important to the economic benefit. In the complex industry, there are so many control loops, the manufacture processes can be guarantee if only the controlled parameters are kept in a certain range. But the controlled parameters are related to the consuming of raw material and energy. So another set of parameters perhaps will lead to the different produce cost and different yield efficiency. Most of the processes have the self balance capacity. Therefore, the optimization of the set-point has an important meaning. In this paper, the supervisory level is setting on the control level. The predictive control is used in the control level. The DCS is the artery in the industry field, the combination of several PID modules has the control function of the predictive PID based model, which can improve the perforcance of the control level. The Genetic Algorithm theory is used in the supervisory level. The Genetic Algorithm theory has the characteristic of better robust, simply compute and strong function. It can deal with the problem with constrain. The good control performance and the economy objective have been verified by applying this method to a boiler system.
Keywords/Search Tags:DCS (Distributed Control Systems), Predictive Control, Robust Stability, Non-Minimum Phase System, Supervisory Control, Predictive PID Control, Fuzzy Control, Genetic Algorithm
PDF Full Text Request
Related items