本文關(guān)鍵詞：基于多目標(biāo)遺傳算法的上游泵送機械密封優(yōu)化研究　出處：《江蘇大學(xué)》2016年碩士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 上游泵送機械密封 均勻試驗 多元回歸 神經(jīng)網(wǎng)絡(luò) 多目標(biāo)遺傳算法

【摘要】：上游泵送機械密封作為一種典型的非接觸機械密封,因其流體潤滑、泄漏量小甚至無泄漏、使用壽命長等優(yōu)勢而得到推廣應(yīng)用。但是,隨著科學(xué)技術(shù)的發(fā)展和對密封性能要求的不斷提高,如何更加有效地設(shè)計出性能優(yōu)秀的機械密封成為迫切需要解決的課題。機械密封端面造型參數(shù)是涉及密封性能優(yōu)劣的關(guān)鍵,目前大多采用數(shù)值計算方法進行單因素單目標(biāo)優(yōu)化設(shè)計。雖有成效但很難達(dá)到高標(biāo)準(zhǔn)要求。為此本文在國家自然科學(xué)基金項目(項目編號:51279067)和航空科學(xué)基金項目(項目編號:201328R3001)的資助下,以上游泵送機械密封為研究對象,結(jié)合神經(jīng)網(wǎng)絡(luò)和多目標(biāo)遺傳算法進行優(yōu)化設(shè)計研究,主要工作和結(jié)論如下:1.基于空化模型和動網(wǎng)格技術(shù),對上游泵送機械密封微間隙內(nèi)流場進行了數(shù)值計算,分析了密封端面槽型幾何參數(shù)對密封性能的影響。研究表明:槽深h、螺旋角α、槽徑寬徑比β和槽區(qū)寬度比γ是密封性能的主要影響因素。2.采用均勻試驗設(shè)計、多元回歸分析和人工神經(jīng)網(wǎng)絡(luò)相結(jié)合的方法,獲得了上游泵送機械密封端面形貌參數(shù)和密封性能之間的函數(shù)關(guān)系。研究表明:結(jié)合均勻試驗設(shè)計和神經(jīng)網(wǎng)絡(luò)可以獲得槽型幾何參數(shù)與密封性能參數(shù)之間的真實函數(shù)關(guān)系;槽型幾何參數(shù)間存在交互作用且對密封性能有著重要影響;槽型優(yōu)化區(qū)間為6≤h≤10μm,16°≤α≤20°,0.35≤β≤0.55,0.4≤γ≤0.6。3.基于CFD數(shù)值計算和神經(jīng)網(wǎng)絡(luò)預(yù)測模型,采用多目標(biāo)遺傳算法建立多目標(biāo)遺傳優(yōu)化策略,對上游泵送機械密封的性能進行優(yōu)化。研究表明:(1)基于合理的試驗設(shè)計,應(yīng)用神經(jīng)網(wǎng)絡(luò)建立槽型幾何參數(shù)和目標(biāo)函數(shù)的數(shù)學(xué)模型,使得函數(shù)關(guān)系更加真實準(zhǔn)確;(2)優(yōu)化后的密封性能得到了提高,采用多元回歸模型液膜剛度提高了7.5%、泄漏量減小了16.6%,采用神經(jīng)網(wǎng)絡(luò)預(yù)測模型及遺傳算法液膜剛度提高了13.1%、泄漏量減小了18.9%。4.采用MSTS-IV密封試驗臺對機械密封的端面形貌優(yōu)化結(jié)果進行試驗驗證。結(jié)果表明試驗結(jié)果與模擬結(jié)果基本是一致的,通過遺傳算法優(yōu)化的結(jié)果更優(yōu)。
[Abstract]:As a typical non-contact mechanical seal, the upstream pump mechanical seal has been popularized and applied because of its advantages of fluid lubrication, little or no leakage, long service life and so on. With the development of science and technology and the continuous improvement of sealing performance requirements. How to design the mechanical seal with excellent performance more effectively becomes an urgent problem to be solved. The key to the sealing performance is the modeling parameters of the end face of the mechanical seal. At present, most of the numerical calculation methods are used for single factor and single objective optimization design. Although it is effective, it is difficult to meet the high standard. This paper is based on the National Natural Science Foundation of China (Project No.: 51279067). And the Aviation Science Foundation project (project number: 201328R3001). Taking upstream pumping mechanical seal as the research object, combining neural network and multi-objective genetic algorithm to optimize design, the main work and conclusions are as follows: 1. Based on cavitation model and moving grid technology. The flow field in micro-clearance of upstream pumping mechanical seal is numerically calculated and the effect of geometric parameters of seal end groove on seal performance is analyzed. The results show that the groove depth is h and the spiral angle is 偽. The groove width to diameter ratio 尾 and groove width ratio 緯 are the main influencing factors of sealing performance. 2. The uniform test design, multivariate regression analysis and artificial neural network are adopted. The functional relationship between the surface morphology parameters and sealing performance of upstream pumping mechanical seal is obtained. Combined with uniform test design and neural network, the real function relationship between groove geometry parameters and sealing performance parameters can be obtained. There is interaction between geometric parameters of groove type and it has an important effect on sealing performance. The optimum interval of groove type is 6 鈮，

本文編號：1377116