沸騰換熱論文基于神經(jīng)網(wǎng)絡(luò)的豎直矩形細(xì)通道內(nèi)沸騰換熱汽液兩相流

1、 沸騰換熱論文：基于神經(jīng)網(wǎng)絡(luò)的豎直矩形細(xì)通道內(nèi)沸騰換熱汽液兩相流型識別研究【中文摘要】在沸騰換熱的研究中發(fā)現(xiàn),汽液兩相流動介質(zhì)的相界面分布狀況,即流型,極大地影響著汽液兩相流的流動特性和傳熱性能,同時也對流動參數(shù)的準(zhǔn)確測量以及兩相流系統(tǒng)運行特性的確定具有很大的影響作用。因此,沸騰換熱汽液兩相流流型識別的研究一直是兩相流參數(shù)分析一個重要組成部分。本課題針對流型識別存在的不足,提出了利用神經(jīng)網(wǎng)絡(luò)進(jìn)行流型識別的方法。首先,本課題以去離子水為被加熱工質(zhì),對槽道寬度分別為2mm、1.5mm、1mm和0.5mm的實驗件進(jìn)行沸騰換熱實驗。實驗過程實現(xiàn)了被測實驗件內(nèi)部沸騰狀態(tài)的可視化,并且測量了工質(zhì)的體積流量

2、、實驗段入口處溫度和壓力和實驗件的壓差波動信號,同時在實驗件上沿工質(zhì)流動方向設(shè)置多組熱電偶,測量不同位置的溫度值。然后,運用matlab小波降噪的方法對獲取的信號進(jìn)行處理。選擇小波模塊為wavelet 2-D,選擇母小波為haar,選擇閾值模式為unscaled white noise和horizontal details coefs,選擇閾值函數(shù)為軟閾值。經(jīng)過處理后的信號可以更好的輔助實驗者確定流型的類別,減少主觀判斷錯誤的發(fā)生。最后,通過實驗獲得的各點溫度、液體體積流量和壓差等數(shù)據(jù),計算得到一個無量綱數(shù)。并以此無量綱數(shù)和熱電偶處測得的溫度值為輸入向量,以各向量所對應(yīng)的三種流型單相流(001

3、),彈狀流(011),受限彈狀流(111)為輸出向量。本文建立了四種神經(jīng)網(wǎng)絡(luò)：BP,RBF, SOM和Elman神經(jīng)網(wǎng)絡(luò)。實驗過程中測得的數(shù)據(jù)其中一部分用于神經(jīng)網(wǎng)絡(luò)的建立和訓(xùn)練,另外一部分?jǐn)?shù)據(jù)用于對所建立神經(jīng)網(wǎng)絡(luò)進(jìn)行驗證。神經(jīng)網(wǎng)絡(luò)的識別結(jié)果表明：利用BP和Elman神經(jīng)網(wǎng)絡(luò)進(jìn)行流型的識別效果較好,識別率在90%以上,而利用SOM進(jìn)行識別正確率低于50%,RBF神經(jīng)網(wǎng)絡(luò)的識別能力介于兩者之間。從而,BP和Elman神經(jīng)網(wǎng)絡(luò)可以作為流型的分類器。針對這兩種神經(jīng)網(wǎng)絡(luò)的良好的分類效果,本文建立了流型識別的用戶界面,用戶可以在輸入欄中輸入數(shù)據(jù),直接得到輸出結(jié)果。本文提供了一種識別流型的新方法,與其他識別

4、方法相比,減少了因主觀而造成的識別誤差,提高了流型的識別率?！居⑽恼縄t is found in the studies of boiling heat transfer that the vapor-liquid two-phase flow situation on the interface-flow pattern, can greatly affect not only the two-phase flow and heat transfer characteristics of the vapor-liquid, but also the accurate measuremen

5、t of flow parameters and the determination of the operating characteristics of two-phase system. So the analysis of the pattern recognition of boiling heat transfer is an important part for the vapor-liquid two-phase flow.Firstly, water is taken as the working fluid. The width of the channel is set

6、to be 2mm,1.5mm,1mm and 0.5mm for the boiling heat transfer experiments. The channel is covered with plexi-glass, which is nature to be transparent to achieve the visualization of the experiment. In the study, mass flow rate, temperature, pressure at the entrance of the test section, and fluctuation

7、 signal of the differential pressure are measured, besides, thermo-couples are set along the direction of flow to measure the temperature value of different positions.Then the signal obtained use wavelet denoising methods of matlab for processing. The wavelet mode selects wavelet 2-D, the mother wav

8、elet selects haar, the threshold mode selects the unscaled white noise and horizontal details coefs, threshold function selects the soft threshold. After treatment, the signal can greatly assist the experimenter to determine the flow pattern, and reduce subjective errors. After the signal analysis a

9、nd processing, it can be learned in this experiment that there are three types of flow pattern:single-phase flow, slug flow and limited slug flow.Finally, a dimensionless number is gained by computing the experiments temperature, fluid velocity and differential pressure and other datum. The dimensio

10、nless number is taken as the input vector with the temperature measured by thermo-couples; three corresponding flow:single-phase flow (0 01), slug flow (011) and limited slug flow (111) are taken as the output vectors. In this paper, four neural networks:BP, RBF, SOM, and Elman neural networks are e

11、stablished. Part of the datum is used for the establishment and tra i n i ng of neural networks, other part of the datum i s used for neural network validation. The results show that the recognition of flow pattern of BP and Elman neural network are better than another two networks, with the rate of

12、 90% or more, while the correct rate of SOM is less than 50%. And the recognition rate of RBF neural network is just between them. According to the classification of the good results of the two neural networks, the user interface is established, so the users can output the data directly.This paper p

13、rovides a new method of identifying flow patterns. Comparing with other identification methods, the subjective recognition errors are decreased and the recognition rate of flow can be improved.【關(guān)鍵詞】沸騰換熱 流型 神經(jīng)網(wǎng)絡(luò) 識別【英文關(guān)鍵詞】boiling heat transfer flow pattern neural networks recognition【目錄】基于神經(jīng)網(wǎng)絡(luò)的豎直矩形細(xì)通道

14、內(nèi)沸騰換熱汽液兩相流型識別研究摘要13-14ABSTRACT14-15第1章 緒論16-221.1 課題研究的背景及意義16-171.2 汽液兩相流流型識別的主要方法17-201.2.1 流型的直接測量法17-181.2.2 流形的間接測量法18-191.2.3 基于神經(jīng)網(wǎng)絡(luò)的流形識別方法19-201.3 本文的主要研究內(nèi)容20-22第二章 細(xì)通道內(nèi)沸騰汽液兩相流動的實驗研究22-342.1 實驗系統(tǒng)簡介22-252.2 實驗件252.3 所需實驗儀器及精度25-262.4 實驗步驟26-292.4.1 需要測量的參數(shù)272.4.2 熱電偶布置272.4.3 溫度的測量27-282.4.4 壓

15、差的測量282.4.5 實驗數(shù)據(jù)28-292.5 實驗所觀察到的流型29-322.5.1 流型的定義29-302.5.2 實驗中觀察到的流型30-322.6 本章小結(jié)32-34第三章 基于連續(xù)小波變換的信號處理34-423.1 連續(xù)小波變換34-393.1.1 連續(xù)小波的二維特征34-353.1.2 連續(xù)小波變換的性質(zhì)35-393.1.3 小波變換的依據(jù)393.2 連續(xù)小波變換的圖片處理39-413.2.1 連續(xù)小波的GUI39-403.2.2 信號的降噪處理40-413.3 本章小結(jié)41-42第四章 基于神經(jīng)網(wǎng)絡(luò)的沸騰汽液兩相流流型識別42-724.1 神經(jīng)網(wǎng)絡(luò)基本理論42-494.1.1

16、神經(jīng)網(wǎng)絡(luò)發(fā)展歷程42-444.1.2 神經(jīng)網(wǎng)絡(luò)研究內(nèi)容444.1.3 神經(jīng)元的模型44-464.1.4 神經(jīng)元的連接方式46-494.2 BP神經(jīng)網(wǎng)絡(luò)模型49-554.2.1 BP神經(jīng)網(wǎng)絡(luò)基礎(chǔ)49-504.2.2 BP神經(jīng)網(wǎng)絡(luò)的學(xué)習(xí)算法504.2.3 BP神經(jīng)網(wǎng)絡(luò)的流型識別50-554.3 徑向基函數(shù)網(wǎng)絡(luò)模型55-594.3.1 徑向基函數(shù)神經(jīng)網(wǎng)絡(luò)基礎(chǔ)55-564.3.2 基于徑向基神經(jīng)網(wǎng)絡(luò)的學(xué)習(xí)算法56-584.3.3 基于徑向基神經(jīng)網(wǎng)絡(luò)的流型識別58-594.4 自組織神經(jīng)網(wǎng)絡(luò)模型59-624.4.1 自組織神經(jīng)網(wǎng)絡(luò)基礎(chǔ)59-604.4.2 自組織特征映射神經(jīng)網(wǎng)絡(luò)(SOM)結(jié)構(gòu)60-614.4.3 自組織特征映射神經(jīng)網(wǎng)絡(luò)學(xué)習(xí)算法614.4.4 基于自組織神經(jīng)網(wǎng)絡(luò)的流型識別61-624.5 反饋型神經(jīng)網(wǎng)絡(luò)模型62-684.5.1 反饋型神經(jīng)網(wǎng)絡(luò)基礎(chǔ)62