【摘要】：多目標(biāo)優(yōu)化算法廣泛應(yīng)用于化工領(lǐng)域,諸如過(guò)程控制與操作優(yōu)化、化工設(shè)備設(shè)計(jì)、環(huán)境工程等。近年來(lái),越來(lái)越多的學(xué)者將多目標(biāo)優(yōu)化算法與流程模擬器相結(jié)合解化工優(yōu)化問(wèn)題。由于流程模擬需要花費(fèi)大量時(shí)間進(jìn)行收斂計(jì)算,因此,優(yōu)化算法必須能夠在較少的目標(biāo)函數(shù)評(píng)價(jià)次數(shù)的情況下,也能快速達(dá)到收斂。非支配遺傳算法是多目標(biāo)優(yōu)化領(lǐng)域應(yīng)用最廣的算法,但其必須經(jīng)過(guò)上萬(wàn)次的目標(biāo)函數(shù)評(píng)價(jià)才能得到較好的結(jié)果,而且其本身也存在容易過(guò)早收斂、局部搜索能力不強(qiáng)等缺點(diǎn)。因此,本文旨在提出一種高效的多目標(biāo)優(yōu)化算法,將其應(yīng)用于化工過(guò)程的優(yōu)化。本文的工作主要有以下幾個(gè)方面:(1)闡述了多目標(biāo)優(yōu)化算法的研究背景和意義,從科學(xué)研究和工程應(yīng)用兩個(gè)方面,介紹了多目標(biāo)進(jìn)化算法的發(fā)展,并對(duì)列隊(duì)競(jìng)爭(zhēng)算法的研究和應(yīng)用進(jìn)展進(jìn)行了簡(jiǎn)單介紹;(2)介紹了多目標(biāo)優(yōu)化問(wèn)題的相關(guān)概念與定義,詳細(xì)描述了列隊(duì)競(jìng)爭(zhēng)算法LCA和非支配遺傳算法NSGA-II的計(jì)算流程與關(guān)鍵算子,引入了解集收斂性和均勻性評(píng)價(jià)指標(biāo);(3)提出了一種多目標(biāo)列隊(duì)競(jìng)爭(zhēng)算法MOLCA,采用多種策略,在降低目標(biāo)函數(shù)評(píng)價(jià)次數(shù)的同時(shí)達(dá)到快速收斂。對(duì)MOLCA主要參數(shù)的設(shè)定進(jìn)行了討論,然后利用經(jīng)典測(cè)試函數(shù)對(duì)MOLCA進(jìn)行測(cè)試分析,與NSGA-II相比,該方法表現(xiàn)更優(yōu)秀。將MOLCA應(yīng)用于催化裂化主分餾塔的操作參數(shù)優(yōu)化,以總經(jīng)濟(jì)效益和系統(tǒng)能耗為兩目標(biāo),給出了優(yōu)化的操作方案;(4)針對(duì)NSGA-II存在容易收斂于局部最優(yōu)解和計(jì)算時(shí)間久的問(wèn)題,提出了一種基于多目標(biāo)列隊(duì)競(jìng)爭(zhēng)算法和非支配遺傳算法的混合算法MOLCA-NSGA-II。經(jīng)過(guò)經(jīng)典測(cè)試函數(shù)的測(cè)試分析表明,該算法在運(yùn)算時(shí)間、收斂性和分布性方面都要明顯優(yōu)于NSGA-II。將MOLCA-NSGA-II應(yīng)用于甲醇制烯烴分離過(guò)程優(yōu)化,結(jié)果給出了一系列Pareto最優(yōu)解,可以根據(jù)不同的生產(chǎn)要求,綜合考慮能耗和收率,選擇適宜的操作條件。
[Abstract]:Multi-objective optimization algorithm is widely used in chemical engineering fields, such as process control and operation optimization, chemical equipment design, environmental engineering and so on. In recent years, more and more scholars combine multi-objective optimization algorithm with process simulator to solve chemical optimization problem. Because the process simulation takes a lot of time to calculate convergence, the optimization algorithm must be able to achieve convergence quickly with less evaluation times of objective function. The non-dominated genetic algorithm is the most widely used algorithm in the field of multi-objective optimization, but it must pass through tens of thousands of objective function evaluation to get a better result, and its own shortcomings such as easy premature convergence, weak local search ability and so on. Therefore, this paper proposes an efficient multi-objective optimization algorithm and applies it to the optimization of chemical processes. The main work of this paper is as follows: (1) the research background and significance of multi-objective optimization algorithm are expounded, and the development of multi-objective evolutionary algorithm is introduced from two aspects: scientific research and engineering application. The research and application of queue competition algorithm are briefly introduced. (2) the concept and definition of multi-objective optimization problem are introduced, the calculation flow and key operators of LCA and NSGA-II are described in detail, and the evaluation index of convergence and uniformity of solution set is introduced. (3) A multi-objective queue competition algorithm (MOLCA,) is proposed, which adopts many strategies to reduce the number of evaluation of the objective function and achieve rapid convergence. The setting of the main parameters of MOLCA is discussed, and then the classical test function is used to test and analyze the MOLCA. Compared with NSGA-II, this method performs better than NSGA-II. MOLCA was applied to the optimization of the operation parameters of the main fractionator of FCC. With the total economic benefit and the energy consumption of the system as the two objectives, the optimized operation scheme was given. (4) in view of the problem that NSGA-II is easy to converge to the local optimal solution and the computation time is long, a hybrid algorithm MOLCA-NSGA-II. based on multi-objective queue competition algorithm and non-dominated genetic algorithm is proposed. The test results of classical test function show that the algorithm is superior to NSGA-II. in computing time, convergence and distribution. MOLCA-NSGA-II was applied to the optimization of the separation process of methanol to olefin. A series of optimal solutions of Pareto were given. According to different production requirements, energy consumption and yield could be considered synthetically, and suitable operating conditions could be selected.
【學(xué)位授予單位】：武漢理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2015
【分類(lèi)號(hào)】：TQ015.9

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 吳勇;程明;項(xiàng)敏建;;基于NSGA-Ⅱ的催化裂化分餾塔的多目標(biāo)優(yōu)化[J];計(jì)算機(jī)測(cè)量與控制;2015年01期

2 龍愛(ài)偉;;多目標(biāo)優(yōu)化在化工領(lǐng)域中的應(yīng)用進(jìn)展[J];廣東化工;2014年21期

3 周草臣;陳自郁;何中市;;高維多目標(biāo)優(yōu)化算法分析研究[J];計(jì)算機(jī)科學(xué);2014年S1期

4 鄭金華;賴(lài)念;郭觀七;;多目標(biāo)進(jìn)化算法中基于角度偏好的ε-Pareto支配策略[J];模式識(shí)別與人工智能;2014年06期

5 施辰斐;趙霞;邵惠鶴;;甲醇四塔精餾系統(tǒng)的多目標(biāo)優(yōu)化[J];石油化工自動(dòng)化;2013年06期

6 付博;裴軍;陳玉石;錢(qián)學(xué)勤;陳韶輝;;芳烴液液抽提過(guò)程模擬和優(yōu)化研究[J];計(jì)算機(jī)與應(yīng)用化學(xué);2013年11期

7 敦劍;馮霄;何暢;王東亮;;煤制天然氣酚氨廢水汽提過(guò)程經(jīng)濟(jì)和環(huán)境多目標(biāo)優(yōu)化[J];化工學(xué)報(bào);2013年12期

8 張宇;鄢烈祥;李國(guó)建;史彬;;非支配排序進(jìn)化策略求解煤氣化多目標(biāo)優(yōu)化問(wèn)題[J];化工學(xué)報(bào);2013年12期

9 李建隆;婁曉燕;劉穎;王宇光;;MTO產(chǎn)品分離工藝的模擬與優(yōu)化[J];計(jì)算機(jī)與應(yīng)用化學(xué);2013年05期

10 周建淞;陳益;張曉麗;韓榮榮;仇麗霞;武俊青;;基于向量評(píng)估遺傳算法的多目標(biāo)優(yōu)化效果評(píng)價(jià)及程序測(cè)試[J];中國(guó)衛(wèi)生統(tǒng)計(jì);2012年02期

相關(guān)博士學(xué)位論文 前1條

1 楊娟;煤炭礦區(qū)節(jié)能減排多目標(biāo)優(yōu)化決策研究[D];中國(guó)地質(zhì)大學(xué);2014年

相關(guān)碩士學(xué)位論文 前6條

1 詹雪蘭;催化裂化主分餾塔和吸收穩(wěn)定系統(tǒng)的全流程模擬與優(yōu)化[D];華東理工大學(xué);2014年

2 王康;多目標(biāo)進(jìn)化算法解集分布性評(píng)價(jià)指標(biāo)及其應(yīng)用[D];湘潭大學(xué);2013年

3 伍思敏;多目標(biāo)粒子群優(yōu)化算法的改進(jìn)及應(yīng)用研究[D];江南大學(xué);2013年

4 劉楠楠;基于進(jìn)化算法的多目標(biāo)優(yōu)化算法及應(yīng)用研究[D];南京航空航天大學(xué);2010年

5 汪采萍;蟻群算法的應(yīng)用研究[D];合肥工業(yè)大學(xué);2007年

6 曾威;多目標(biāo)進(jìn)化算法及其在約束優(yōu)化中的應(yīng)用研究[D];中南大學(xué);2007年

，

本文編號(hào)：2395174