Dissertation > Industrial Technology > Automation technology,computer technology > Automation technology and equipment > Automation systems > Automatic control,automatic control system

Research of Decoupling Control System in Engineering

Author WangQiZhi
Tutor WangYongChu
School Huaqiao University
Course Mechanical Manufacturing and Automation
Keywords Degree of coupling Grouping decoupling Class of Feedforward decoupling State feedback Weak coupling Extracellular matrix Wavelet Dynamic Matrix Filter Intelligent decoupling
Type PhD thesis
Year 2002
Downloads 998
Quotes 13
Download Dissertation

Decoupling control is a strong background of the subject of the application, to the researchers a lot more than twenty years, become a very popular research directions in the field of automation. Involved in its range of applications in many fields of the national economy such as iron and steel, petroleum, chemical industry, light industry, machinery, military. As we all know, the production process is an orderly process, a chain, the intricate relationship between the variables of a process variable fluctuations tend to affect more than one variable changes, the coupling is a widespread phenomenon in the industrial production process. Mutual interference between variables called coupling to eliminate this coupling is called decoupling. This paper presents a new concept of a decoupled system of analysis and judgment of universal significance degree of coupling in the absorption of predecessors to improve and perfect the many studies on the basis of the results, such as Bristol relative amplification factor λ ij The method of λ ij deviate from the size of 1 to determine the size of the degree of coupling, but in fact, λ ij is negative, the system is coupled than λ ij considerably more than one situation is much more serious. Compared with this, the degree of coupling this paper, P ij size of truthfully reflect the degree of coupling. The the coupling calculation triangular decoupling network design or grouping decoupling network design combine to weaken the degree of coupling, the effect is quite significant. In packet decoupled process, trying the coupling most severe slip separated. Each separation, the total degree of coupling of the system are reduced, and is often significantly reduced. When the degree of coupling to a certain extent, the system can be stable. Further adjust the controller parameters of the output at their best. This is the weakening system coupled degrees. Production practice, the degree of coupling analysis system based on the use of packet decoupling to weaken the system of coupled model reduction methods to simplify decoupling network model, a better solution in successive weakening process, to decouple equivalent a problem of increasing the order of the control channel. Study a reformer object properties, recovery characteristics equivalent object is distorted greatly simplified packet decoupling embodiment difficulty, the practical value of the works. The advantages of using transfer function describing the system characteristics is straightforward, the drawback is that not all the characteristics of the characterization of the system, the information is incomplete, state space approach is just the opposite. Therefore, these two methods can be said to complement each other, which is especially suitable for the pole configuration of the system. Pole assignment is to choose the state feedback matrix K, the intrinsic polynomial det [sI-(A-BK) = 0, the intrinsic value is the optimum value to obtain the optimal configuration of the root. The article studies Butterworth of Bessel filter decoupling state feedback pole placement. The simulation results show that this root configuration to achieve the good results. This article reveals the state feedback decoupling conditions B * is nonsingular for detB * = 0 detG ≠ Weak coupling system, can not be directly used state feedback decoupling. Some scholars use I = P -1 (S) P (S), so that P -1 (S) and P (S) and the closed-loop transfer function reconstruction makes detB ≠ 0. The method uses the P (S), the main diagonal elements of 1 or s, and the lower triangular element is 0 or αS. In this way, the closed-loop transfer function must factor 1 / s, is not conducive to the stability and robustness of the system. To address this issue, we propose an improved method outside the matrix P (S) can be the main angle element 1 or IIR filter structure (for simplicity, take s beta) lower triangular elements for IIR The structural form of the filter, take α (s β)) or 0 (for simplicity. Thus, the closed loop transfer function will not be provided with a factor 1 / s. Simulation and theoretical analysis prove that this improved method is feasible. In this paper, a class of non-accurate model decoupling control problem. Imprecision or uncertainty of the model a lot of reality control engineering. Meaningful innovation we made three aspects: (1) the introduction of dynamic matrix (DMC) predictive decoupling control system; (2) The application of wavelet transform implementation of the DMC system decoupling; (3) try wavelet transform technology applications decoupling control system fault diagnosis, to get the desired results. Involved in the decoupling field of study, the authors also studied a number of significant decoupling of application characteristics, decoupling methods such as class feedforward. In an industrial process, a 2 x 2 a coupling system, the feedforward decoupling more common, but the disadvantage is that the addition of decoupling network will change the characteristics of the transfer function, and the resulting optimization controller difficulties. This study class feedforward decoupling method to solve this problem. Feedforward decoupling class advantage is the transfer function of the main control channel characteristics remain unchanged after decoupling. In addition, class feed decoupling network matrix transfer function also has the characteristics of the inverse matrix passed completely decoupled. The drawback is that its the decoupling network presents a closed-loop structure, traditionally considered difficult to obtain stable decoupling control system, this approach short shrift, in particular, is higher than the 2X2 coupled system, almost did not see the relevant research. This study shows that, after the introduction of suitable restriction conditions for the controller, the method is widely available value, not only can be used for ZXZffi aggregate systems, the coupling system can also be used for more than 2X2 particularly partial decoupling and system integration method , application easier. In this paper, the thermal power plant coal mill 3X3 decoupled system, a detailed analysis of this design approach. Units decoupling rule is decoupling method, because only theoretical significance and value in use can be calculated decoupling network, but because of its often physically can not be achieved. This paper using a decoupling network factors modify the law to solve this problem. The prevailing view of singular objects is difficult to achieve decoupling control objects matrix G companion) singular, its inverse G '() does not exist, singular object, however, is the most serious coupling objects, decoupling is almost impossible to control. In this paper, a balanced structure method to build a system, not a traditional decoupling decoupling effect of the control structure of the system can be achieved. These the intelligent decoupling, such as neural network decoupling and fuzzy decoupling began to rage issues, methods and applications are not mature, I am also of its research and engineering structures.

Related Dissertations
More Dissertations