【摘要】：眾所周知,原油是一種富含高價(jià)值但又極其復(fù)雜的混合物,主要是復(fù)雜的烴類和非烴類的混合物,因此必須通過各種加工手段才能使其轉(zhuǎn)化為符合質(zhì)量要求的產(chǎn)品。一般情況下,通過一次加工,采用常減壓蒸餾后,能夠得到10~40%的輕質(zhì)油品,如汽油和柴油等等,而大部分余下的是利用價(jià)值比較低的重質(zhì)油和殘?jiān)�。由于國民�?jīng)濟(jì)和國防需要大量的輕質(zhì)油品,若不將重油進(jìn)行二次加工,而只通過原油常減壓蒸餾得到的輕質(zhì)油是不能滿足需要的。為了滿足國民經(jīng)濟(jì)和國防的需要,從而使催化裂化技術(shù)得到了發(fā)展。反應(yīng)-再生部分是催化裂化裝置(Fluid Catalytic Cracking Unit,FCCU)的核心部分,關(guān)于催化裂化的優(yōu)化與控制一直是研究者關(guān)注的焦點(diǎn)。針對(duì)反應(yīng)-再生部分的控制研究,本文主要進(jìn)行了如下幾方面的研究工作:(1)FCCU反應(yīng)-再生部分機(jī)理模型研究。本章對(duì)反應(yīng)-再生部分的機(jī)理模型和求解方法進(jìn)行了研究分析,建立了FCCU優(yōu)化和控制所需的數(shù)學(xué)模型,然后對(duì)該模型模擬計(jì)算,實(shí)現(xiàn)了FCCU反應(yīng)-再生部分的集總濃度與反應(yīng)溫度仿真分析。(2)FCCU反應(yīng)-再生部分BP(Back Propagation,又稱誤差反向傳播)神經(jīng)網(wǎng)絡(luò)建模。根據(jù)神經(jīng)網(wǎng)絡(luò)能夠無限接近于實(shí)際對(duì)象的特點(diǎn),通過MATLAB軟件編程,建立FCCU反應(yīng)-再生部分BP神經(jīng)網(wǎng)絡(luò)模型。所建模型的誤差較小,且訓(xùn)練網(wǎng)絡(luò)時(shí)收斂非�？�。(3)FCCU反應(yīng)-再生部分PID(Proportion Integration Differentiation,比例-積分-微分)控制方案設(shè)計(jì)及仿真。明確了FCCU反應(yīng)-再生部分的控制目標(biāo)、操縱變量和被控變量。利用MATLAB軟件對(duì)反應(yīng)-再生部分進(jìn)行PID控制方案設(shè)計(jì)并仿真。結(jié)果表明,PID控制的控制效果良好,可以滿足催化裂化的工業(yè)生產(chǎn)目標(biāo)。(4)FCCU反應(yīng)-再生部分預(yù)測(cè)控制設(shè)計(jì)及性能分析。通過探索模型預(yù)測(cè)控制(Model Predictive Control,簡(jiǎn)稱MPC)的基本原理,將神經(jīng)網(wǎng)絡(luò)與預(yù)測(cè)控制結(jié)合,借助Simulink神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)控制工具箱,進(jìn)行反應(yīng)-再生部分神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)控制研究方案設(shè)計(jì)。并和PID控制相比較,結(jié)果表明,FCCU反應(yīng)-再生部分的神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)控制的控制質(zhì)量明顯優(yōu)于PID控制,且振蕩過程較短,超調(diào)量也較小,可實(shí)現(xiàn)FCCU反應(yīng)-再生的優(yōu)化控制。
[Abstract]:As we all know, crude oil is a kind of high value but extremely complex mixture, mainly complex hydrocarbon and non-hydrocarbon mixture, so it must be converted into quality products by various processing methods. In general, 10% of the light oil, such as gasoline and diesel oil, can be obtained by one processing and atmospheric and vacuum distillation, while most of the rest are heavy oil and residual oil with low value. Because the national economy and national defense need a large number of light oil products, if the heavy oil is not reprocessed, the light oil obtained only by atmospheric and vacuum distillation of crude oil can not meet the needs. In order to meet the needs of national economy and national defense, catalytic cracking technology has been developed. The reaction-regeneration part is the core part of catalytic cracking unit (Fluid Catalytic Cracking Unit,FCCU, and the optimization and control of catalytic cracking has always been the focus of researchers' attention. Aiming at the control of reaction-regeneration part, the main research work in this paper is as follows: (1) the mechanism model of FCCU reaction-regeneration part is studied. In this chapter, the mechanism model and solution method of reaction-regeneration part are studied and analyzed, and the mathematical model needed for FCCU optimization and control is established, and then the model is simulated and calculated to realize the simulation analysis of lumped concentration and reaction temperature of FCCU reaction-regeneration part. (2) FCCU reaction-regeneration part BP (Back Propagation, also known as error back propagation) neural network modeling. According to the fact that the neural network can be infinitely close to the actual object, the BP neural network model of FCCU reaction-regeneration part is established by MATLAB software programming. The error of the model is small, and the convergence of the training network is very fast. (3) Design and simulation of FCCU reaction-regenerated part PID (Proportion Integration Differentiation, proportional integral-differential) control scheme. The control objectives, manipulation variables and controlled variables of FCCU reaction-regeneration part are defined. The PID control scheme of the reaction-regeneration part is designed and simulated by MATLAB software. The results show that the control effect of PID control is good and can meet the industrial production goal of catalytic cracking. (4) Predictive control design and performance analysis of FCCU reaction-regeneration part. By exploring the basic principle of model predictive control (Model Predictive Control,), combining neural network with predictive control, and with the help of Simulink neural network predictive control toolbox, the research scheme of reaction-regeneration neural network predictive control is designed. Compared with PID control, the control quality of neural network predictive control in FCCU reaction-regeneration part is obviously better than that in PID control, and the oscillation process is shorter and the overshoot is small. The optimal control of FCCU reaction-regeneration can be realized.
【學(xué)位授予單位】：新疆大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2015
【分類號(hào)】：TE96
，

本文編號(hào)：2500714