• 
    

    
    

      99热精品在线国产_美女午夜性视频免费_国产精品国产高清国产av_av欧美777_自拍偷自拍亚洲精品老妇_亚洲熟女精品中文字幕_www日本黄色视频网_国产精品野战在线观看 ?

      基于支持向量機(jī)的滴灌灌水器流量預(yù)測模型建立與驗(yàn)證

      2018-02-28 06:31:52王新端
      關(guān)鍵詞:訓(xùn)練樣本水力灌水

      郭 霖,白 丹,王新端,王 程,周 文,程 鵬

      ?

      基于支持向量機(jī)的滴灌灌水器流量預(yù)測模型建立與驗(yàn)證

      郭 霖1,白 丹1※,王新端1,王 程1,周 文2,程 鵬2

      (1. 西安理工大學(xué)水利水電學(xué)院,西安 710048;2. 華北水利水電大學(xué)水利學(xué)院,鄭州 450011)

      為了直接、準(zhǔn)確預(yù)測灌水器流量,引入支持向量機(jī)預(yù)測方法,取灌水器6個(gè)工作壓力和8個(gè)幾何參數(shù)作為因素,正交設(shè)計(jì)安排300組灌水器訓(xùn)練樣本和30組檢測樣本,并采用精度較高的SST-模型模擬計(jì)算灌水器流量,同時(shí)利用遺傳算法對(duì)支持向量機(jī)參數(shù)進(jìn)行優(yōu)化計(jì)算,得到與模擬流量誤差最小的流量預(yù)測值。結(jié)果表明,懲罰參數(shù)為100、核函數(shù)參數(shù)為20時(shí)檢測樣本的流量預(yù)測值與模擬值的誤差最小,平均相對(duì)誤差為1.91%,決定系數(shù)為0.98,而回歸擬合方法計(jì)算得到的平均相對(duì)誤差為6.45%,決定系數(shù)為0.93,表明支持向量機(jī)預(yù)測流量的優(yōu)越性,且30組試驗(yàn)驗(yàn)證樣本的平均相對(duì)誤差為2.25%,證明支持向量機(jī)預(yù)測的準(zhǔn)確性和可靠性。預(yù)測模型建立可有效地提高灌水器研發(fā)效率,對(duì)水力性能評(píng)估和流道結(jié)構(gòu)設(shè)計(jì)與優(yōu)化提供依據(jù)。

      流量;數(shù)值分析;模型;滴灌灌水器;工作壓力;幾何參數(shù);支持向量機(jī);優(yōu)化

      0 引 言

      灌水器的消能方式主要是在流道內(nèi)部產(chǎn)生局部水頭損失,其原理是借助水力學(xué)中的圓角彎管、折角彎管、以及流道斷面的突擴(kuò)、突縮作用[1-2]通過流道結(jié)構(gòu)[3]與形式[4]的變化,消除進(jìn)口處的多余壓力[5],將有壓水流變成滴水狀[6-7],從而使出流均勻與穩(wěn)定,是保證灌溉質(zhì)量最為重要的部件[8-9]。灌水器流道的不同幾何參數(shù)[10-13]和工作壓力[14-17]對(duì)應(yīng)不同的流量,進(jìn)而影響灌水器的水力性能,設(shè)計(jì)不同額定流量的灌水器產(chǎn)品可滿足不同作物的灌水需求[18-19]。因此,流量測定對(duì)于灌水器性能研究和產(chǎn)品研發(fā)至關(guān)重要。

      流道的幾何參數(shù)和工作壓力影響灌水器的流量[20-21],如何高效和準(zhǔn)確地計(jì)算灌水器的流量是有待解決的問題[22-23]。劉春景等[24-25]采用響應(yīng)曲面法對(duì)流態(tài)指數(shù)、流量系數(shù)與流道關(guān)鍵參數(shù)進(jìn)行了二次多項(xiàng)式擬合,確定關(guān)鍵參數(shù)的最佳組合;Li等[26]采用SPSS軟件對(duì)流量與流道幾何參數(shù)進(jìn)行多元線性回歸,建立二者的經(jīng)驗(yàn)公式;Li等[27]、Celik等[28]通過多元回歸分析了灌水器壓力與流量之間的關(guān)系,并建立流量計(jì)算的經(jīng)驗(yàn)公式;Vekariya等[29]建立了流量與幾何參數(shù)的多元回歸模型,并利用量綱分析法[30]探究了幾何參數(shù)對(duì)流量以及水力性能的影響。

      綜合目前計(jì)算灌水器流量的方法,鮮有對(duì)于灌水器流量直接預(yù)測的研究,一般根據(jù)流量與壓力的經(jīng)驗(yàn)公式先回歸計(jì)算流態(tài)指數(shù)和流量系數(shù),再建立二者與流道幾何參數(shù)的線性或非線性回歸方程,間接計(jì)算流量,統(tǒng)稱為回歸擬合方法,其原理為經(jīng)驗(yàn)風(fēng)險(xiǎn)最小化(empirical risk minimization,ERM)[31]。但回歸擬合方法存在一定的不足,其一,回歸擬合方法的前提是樣本數(shù)趨于無窮大時(shí)的漸進(jìn)理論,即對(duì)于大樣本數(shù)據(jù)可得到較為可信的預(yù)測結(jié)果,但在研究過程中所能提供的樣本數(shù)有限,回歸方程的泛化能力較弱[32];其二,通過流態(tài)指數(shù)和流量系數(shù)間接預(yù)測流量的方法,實(shí)質(zhì)上是雙變量回歸計(jì)算問題,計(jì)算流量多有不便,同時(shí)多次回歸計(jì)算存在較大的擬合誤差。

      因此,尋求更為合理、簡便和直接的灌水器流量預(yù)測方法、彌補(bǔ)計(jì)算流量的不足、提高預(yù)測模型的泛化能力對(duì)于灌水器的深入研究是必要的。近年來,很多學(xué)者借鑒了具有較強(qiáng)泛化能力和廣泛適用性的支持向量機(jī)(support vector machine,SVM)在氣象、水文預(yù)測等領(lǐng)域的突出表現(xiàn)[33-34],對(duì)土壤的水力學(xué)參數(shù)[35]、濕度[36]、水分入滲[37]以及10 m壓力下迷宮式滴頭水力性能[31]等進(jìn)行預(yù)測,并取得了較為理想的結(jié)果,同時(shí)由于SVM是在統(tǒng)計(jì)學(xué)習(xí)和結(jié)構(gòu)風(fēng)險(xiǎn)最小化原理基礎(chǔ)上建立的一種預(yù)測方法,對(duì)于預(yù)測對(duì)象具有一定的普適性,因此,本文采用SVM預(yù)測方法構(gòu)建預(yù)測灌水器流量的SVM響應(yīng)面,直接預(yù)測不同工作壓力、不同幾何參數(shù)的雙向?qū)_流灌水器流量,即給定1組輸入(不同工作壓力、不同幾何參數(shù))如何求得所需的輸出(流量)是本文研究的重點(diǎn),解決這一問題對(duì)增加預(yù)測灌水器流量的可信度、提高灌水器研發(fā)效率、降低研發(fā)成本具有重要意義。

      1 理論基礎(chǔ)

      1.1 Fluent數(shù)值模擬計(jì)算理論

      連續(xù)性方程

      動(dòng)量方程為

      式中為流體速度,m/s;、、分別為流速在、、坐標(biāo)軸上的分量,m/s;為水的密度,kg/m3;為動(dòng)力黏度系數(shù),N·s/m2;為流體的壓力,Pa;div為散度;grad為梯度;F、F、F為微元體在、、坐標(biāo)軸上的體力,N/m3,當(dāng)體力只有重力,且軸豎直向上,則F=0,F=0,F=-。

      由于在前期的研究中對(duì)Fluent模擬精度進(jìn)行了綜合對(duì)比[38],因此不再贅述,本文采用模擬精度較好的SST-模型對(duì)訓(xùn)練樣本和檢測樣本中的灌水器流量進(jìn)行模擬,以獲得與流量真實(shí)值更為吻合的結(jié)果。

      1.2 SVM計(jì)算理論

      假設(shè)存在樣本集為

      其中x∈n是第個(gè)訓(xùn)練樣本的輸入值,且為維列向量,對(duì)應(yīng)的目標(biāo)值為

      ∈且y是第個(gè)訓(xùn)練樣本的輸出值,根據(jù)個(gè)訓(xùn)練樣本值導(dǎo)出對(duì)的依存關(guān)系,同時(shí)引入核函數(shù)[32],核變換后,決策函數(shù)為

      通常情況下最小化置信范圍,決策函數(shù)可轉(zhuǎn)換為最優(yōu)化方程組求解

      引入松弛變量ξξ,可得到

      引入Lagrange函數(shù),最優(yōu)化函數(shù)可變換為

      式中為不敏感損失參數(shù);為懲罰參數(shù),>0;ξξ為松弛變量;α、α為與第個(gè)樣本對(duì)應(yīng)的Lagrange乘子;αα為與第個(gè)樣本對(duì)應(yīng)的Lagrange乘子;(x,x)為核函數(shù)。

      本文核函數(shù)采用徑向基核函數(shù)進(jìn)行計(jì)算,可表示為

      式中為核函數(shù)參數(shù)。

      2 灌水器流量預(yù)測模型

      2.1 灌水器三維模型

      灌水器三維模型如圖1所示。

      1.流道進(jìn)口 2.流道出口 3.外壁面 4.分水件 5.擋水件

      2.2 灌水器流道幾何參數(shù)

      在文獻(xiàn)[15-16]灌水器流道關(guān)鍵幾何參數(shù)的取值基礎(chǔ)上,增加相鄰流道單元距離、流道深度、流道單元數(shù),共設(shè)置8個(gè)灌水器流道幾何參數(shù),流道幾何參數(shù)如圖2所示。

      注:S表示外壁面到分水件之間的流道距離,mm;T表示分水件與擋水件之間的齒尖距離,mm;W表示外壁面到擋水件之間的流道距離,mm;Z表示分水件到擋水件之間的最大流道距離,mm;d表示擋水件的垂直底柱高度,mm;e表示相鄰流道單元距離,mm;D表示流道深度,mm;N表示流道單元數(shù)。下同。

      流道幾何參數(shù)取值范圍包括:、、分別取0.6、0.7、0.8、0.9、1.0 mm,和分別取1.0、1.1、1.2、1.3、1.4 mm,取0、0.3、0.6、0.9、1.2 mm,取0.5、0.6、0.7、0.8、0.9 mm,取18、20、22、24、26。

      2.3 流量預(yù)測模型樣本集建立

      2.3.1 訓(xùn)練樣本集

      流道幾何參數(shù)和工作壓力是影響灌水器流量的因素,且影響因素與流量之間存在復(fù)雜的數(shù)學(xué)關(guān)系,則因素與流量的SVM訓(xùn)練樣本集表示為

      式中和分別為輸入和輸出的訓(xùn)練樣本集;x為第個(gè)訓(xùn)練樣本的輸入值,為多維空間向量,包含9個(gè)變量;y為第個(gè)訓(xùn)練樣本的輸出值;為灌水器的工作壓力,kPa;、、、、、、、均為灌水器的流道幾何參數(shù),mm;q為第個(gè)訓(xùn)練樣本對(duì)應(yīng)的灌水器流量,L/h。

      2.3.2 檢測樣本集

      影響因素與流量的SVM檢測樣本集可表示為

      式中XY分別為輸入和輸出的檢測樣本集;x為第個(gè)檢測樣本的輸入值;y為第個(gè)檢測樣本的預(yù)測值;q為第個(gè)檢測樣本對(duì)應(yīng)灌水器的預(yù)測流量,L/h。

      3 材料與方法

      3.1 SVM訓(xùn)練樣本集設(shè)計(jì)

      為了從全面設(shè)計(jì)方案的樣本點(diǎn)中挑選出部分具有代表性的樣本點(diǎn)作為訓(xùn)練樣本集,依據(jù)正交設(shè)計(jì)的“均勻分散”和“整齊可比”的“正交性”特征,按照正交設(shè)計(jì)表L50(511)安排訓(xùn)練樣本方案,如表1所示,每個(gè)灌水器方案分別在50~250 kPa工作壓力下采用數(shù)值模擬進(jìn)行流量計(jì)算,每隔40 kPa計(jì)算一次,共計(jì)算6次,總共需要計(jì)算50×6=300組流量,并將300組流量模擬值作為SVM預(yù)測流量的訓(xùn)練樣本集,即訓(xùn)練樣本集中9個(gè)因素各有300組輸入,對(duì)應(yīng)訓(xùn)練樣本集有300組輸出。

      3.2 SVM檢測樣本集設(shè)計(jì)

      從全面設(shè)計(jì)方案的樣本點(diǎn)中隨機(jī)挑選5個(gè)灌水器幾何參數(shù)組合作為檢測樣本方案,如表2所示,每個(gè)方案采用數(shù)值模擬計(jì)算6個(gè)工作壓力的流量,總共需要計(jì)算5×6=30組流量,并將30組流量模擬值作為SVM預(yù)測流量的檢測樣本集,即通過SVM的訓(xùn)練學(xué)習(xí)可計(jì)算出檢測樣本集X中的30組輸入所對(duì)應(yīng)的30組灌水器流量的Y預(yù)測值。

      3.3 SVM樣本集數(shù)據(jù)歸一化

      為了消除SVM樣本集中的輸入、輸出項(xiàng)對(duì)預(yù)測結(jié)果的影響,在對(duì)灌水器流量預(yù)測時(shí)應(yīng)對(duì)訓(xùn)練樣本和檢測樣本中各項(xiàng)分別進(jìn)行歸一化處理,其歸一化公式為

      式中AA分別為歸一化處理前和處理后的各方案的變量;min和max分別為每個(gè)樣本對(duì)應(yīng)各項(xiàng)的最小值和最大值。

      表1 訓(xùn)練樣本方案

      表2 檢測樣本方案

      3.4 SVM參數(shù)

      在SVM中對(duì)訓(xùn)練和學(xué)習(xí)效果影響最大的2個(gè)參數(shù)為和,其中參數(shù)直接影響模型的穩(wěn)定性,避免模型在學(xué)習(xí)和推廣過程中產(chǎn)生欠學(xué)習(xí)和過學(xué)習(xí)問題,決定了適應(yīng)誤差的最小化和平滑程度;參數(shù)反映了支持向量之間的相關(guān)程度,直接影響支持向量之間聯(lián)系的松弛度,避免產(chǎn)生欠擬合和過擬合問題,決定模型預(yù)測的推廣能力和泛化性,因此,在對(duì)灌水器流量預(yù)測時(shí)需要對(duì)參數(shù)和進(jìn)行調(diào)節(jié)和優(yōu)化,得到較為理想的預(yù)測結(jié)果。

      3.5 SVM參數(shù)調(diào)節(jié)與優(yōu)化

      采用遺傳算法對(duì)SVM參數(shù)各種群個(gè)體進(jìn)行選擇、交叉、變異,逐代產(chǎn)生新的近似最優(yōu)結(jié)果,最終計(jì)算得到參數(shù)個(gè)體的最優(yōu)解。

      3.5.1 目標(biāo)函數(shù)

      本文對(duì)SVM預(yù)測流量準(zhǔn)確性的評(píng)價(jià)通過流量預(yù)測值與流量模擬值進(jìn)行相對(duì)誤差對(duì)比,以相對(duì)誤差最小為目標(biāo)選擇最優(yōu)參數(shù),目標(biāo)函數(shù)可表示為

      式中為懲罰參數(shù);為核函數(shù)參數(shù);為流量的相對(duì)誤差,%;為流量模擬值,L/h;q為流量預(yù)測值,L/h。

      3.5.2 約束條件

      結(jié)合相關(guān)文獻(xiàn)[39-40]中SVM參數(shù)的取值范圍,將取值范圍作為上下限約束,可表示為

      3.5.3 參數(shù)優(yōu)化模型求解

      采用Matlab遺傳算法工具箱對(duì)目標(biāo)函數(shù)進(jìn)行求解,其中遺傳算法適應(yīng)度反映遺傳算法優(yōu)化計(jì)算中對(duì)應(yīng)解的優(yōu)劣程度,根據(jù)目標(biāo)函數(shù)類型建立適應(yīng)度函數(shù),可表示為

      式中為目標(biāo)函數(shù)界限的保守估計(jì)值。

      遺傳算法程序中求解的主要變量分別設(shè)置為:群體中個(gè)體的數(shù)目NIND為100,每個(gè)變量使用20位表示,即PRECI為20,最大遺傳代數(shù)MAXGEN為200,交叉概率XOVR為1,代溝GGAP為0.9。

      4 結(jié)果與分析

      4.1 SVM訓(xùn)練樣本流量模擬結(jié)果

      對(duì)訓(xùn)練樣本中的300組灌水器進(jìn)行流量模擬計(jì)算(樣本編號(hào)為1~300),由于樣本量較大,且篇幅有限,列出部分模擬結(jié)果僅供參考,訓(xùn)練樣本結(jié)果如表3所示,并分別對(duì)各項(xiàng)進(jìn)行歸一化處理。

      表3 訓(xùn)練樣本結(jié)果

      續(xù)表

      注:表示灌水器的工作壓力,kPa,下同。

      Note:is working pressure of emitter, kPa, same as below.

      4.2 SVM檢測樣本流量預(yù)測結(jié)果與分析

      在約束條件范圍內(nèi)采用遺傳算法工具箱計(jì)算得到當(dāng)為100、為20時(shí),灌水器檢測樣本的SVM流量預(yù)測值與模擬值的相對(duì)誤差最小,為1.91%,檢測樣本集中的30組灌水器樣本(樣本編號(hào)為301~330)的流量預(yù)測結(jié)果如表4所示。

      表4 檢測樣本結(jié)果

      表4分別采用回歸擬合方法和SVM方法預(yù)測灌水器流量,對(duì)2種預(yù)測流量方法進(jìn)行誤差對(duì)比,其結(jié)果如圖3所示?;貧w擬合得到的流量預(yù)測值與模擬值的相對(duì)誤差為0.15%~26.69%,平均相對(duì)誤差為6.45%,決定系數(shù)為0.93,其中檢測樣本307的相對(duì)誤差最大,為26.69%,很大程度偏離了灌水器的真實(shí)流量值;而采用SVM計(jì)算得到的流量預(yù)測值與模擬值的相對(duì)誤差為0.09%~6.43%,平均相對(duì)誤差為1.91%,決定系數(shù)為0.98,預(yù)測值與模擬值的相關(guān)性好,完全滿足對(duì)灌水器流量預(yù)測的需求,從而說明本文提出的采用SVM預(yù)測灌水器流量方法的優(yōu)越性。

      圖3 預(yù)測流量的相對(duì)誤差

      4.3 預(yù)測模型試驗(yàn)驗(yàn)證

      為進(jìn)一步驗(yàn)證SVM預(yù)測流量方法的可靠性,選取5組灌水器流道幾何參數(shù)組合作為試驗(yàn)驗(yàn)證方案,對(duì)比流量試驗(yàn)值與回歸擬合方法和SVM預(yù)測方法計(jì)算的流量預(yù)測值的誤差,驗(yàn)證方案參數(shù)取值及計(jì)算結(jié)果如表5所示,試驗(yàn)采用EM-G32S-X32型高精密雕刻機(jī)同比例加工灌水器樣機(jī),并參照GB/T17187-2009《農(nóng)業(yè)灌溉設(shè)備-滴頭和滴灌管-技術(shù)規(guī)范和試驗(yàn)方法》的技術(shù)要求搭建試驗(yàn)平臺(tái),對(duì)灌水器樣機(jī)進(jìn)行流量測試,每組灌水器樣機(jī)分別安裝5個(gè),并在工作壓力為50~250 kPa范圍內(nèi)同時(shí)進(jìn)行流量測試,每組測試時(shí)間均為5 min,測試3次,取3次流量測試的平均值作為灌水器流量的試驗(yàn)值。

      表5 試驗(yàn)驗(yàn)證方案及結(jié)果

      表5中每組灌水器的驗(yàn)證方案均測試6個(gè)工作壓力的流量,即共有30組試驗(yàn)驗(yàn)證樣本(樣本編號(hào)為331~360),試驗(yàn)驗(yàn)證樣本的相對(duì)誤差如圖4所示。誤差計(jì)算結(jié)果表明回歸擬合得到的流量預(yù)測值與試驗(yàn)值的相對(duì)誤差為0.39%~16.36%,平均相對(duì)誤差為5.52%,與流量試驗(yàn)值的誤差較大;而采用SVM計(jì)算得到的流量預(yù)測值與試驗(yàn)值的相對(duì)誤差為0.14%~5.13%,平均相對(duì)誤差為2.25%,在誤差范圍之內(nèi),決定系數(shù)為0.98,表明建立預(yù)測流量的SVM響應(yīng)面可以準(zhǔn)確地反映灌水器進(jìn)口工作壓力、流道幾何參數(shù)與流量的關(guān)系,并通過SVM模型的建立可直接對(duì)灌水器的流量和水力性能進(jìn)行預(yù)測和評(píng)估,一定程度上縮短灌水器試驗(yàn)周期,提高灌水器研發(fā)效率。

      圖4 試驗(yàn)驗(yàn)證樣本的相對(duì)誤差

      5 結(jié)論與討論

      1)本文基于SVM預(yù)測方法建立灌水器流量預(yù)測響應(yīng)面,并對(duì)SVM中的懲罰參數(shù)和不敏感損失參數(shù)進(jìn)行優(yōu)化,其中最優(yōu)值為100、為20時(shí),SVM預(yù)測流量的平均相對(duì)誤差最小,為1.91%,滿足灌水器流量預(yù)測的精度需求。

      2)對(duì)比回歸擬合方法和SVM方法的預(yù)測流量精度,前者預(yù)測流量的平均相對(duì)誤差為6.45%,后者平均相對(duì)誤差為1.91%,精度更高,說明采用SVM預(yù)測流量的優(yōu)越性。

      3)灌水器流量試驗(yàn)驗(yàn)證得到SVM的流量預(yù)測值與試驗(yàn)值的平均相對(duì)誤差為2.25%,在誤差范圍內(nèi),證明建立預(yù)測灌水器流量的SVM響應(yīng)面的可靠性。

      文中主要從直接預(yù)測灌水器流量的角度提出了SVM響應(yīng)面方法,初步得到SVM預(yù)測流量的平均相對(duì)誤差為1.91%,在預(yù)測精度方面有一定的增強(qiáng),并可降低灌水器的研發(fā)成本,提高研發(fā)效率,有深入研究的必要。但由于灌水器的研究最終是要對(duì)流道結(jié)構(gòu)進(jìn)行優(yōu)化,設(shè)計(jì)性能最優(yōu)的灌水器,因此,后期應(yīng)更多的從灌水器性能角度出發(fā),將SVM流量預(yù)測方法與流道結(jié)構(gòu)優(yōu)化算法相結(jié)合,實(shí)現(xiàn)SVM預(yù)測流量和性能到結(jié)構(gòu)優(yōu)化算法的數(shù)據(jù)傳輸與循環(huán),直接預(yù)測和設(shè)計(jì)灌水器流道最優(yōu)幾何參數(shù),彌補(bǔ)現(xiàn)有灌水器研發(fā)過程中的不足,有助于實(shí)現(xiàn)灌水器研發(fā)的模塊化、數(shù)字化以及精確化,提高灌水器研發(fā)效率和農(nóng)業(yè)工程使用價(jià)值。

      [1] 苑偉靜,魏正英,楚華麗,等. 分流式灌水器結(jié)構(gòu)優(yōu)化設(shè)計(jì)與試驗(yàn)[J]. 農(nóng)業(yè)工程學(xué)報(bào),2014,30(17):117-124.

      Yuan Weijing, Wei Zhengying, Chu Huali, et al. Optimal design and experiment for divided-flow emitter in drip irrigation[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2014, 30(17): 117-124. (in Chinese with English abstract)

      [2] 劉潔,魏青松,史玉升,等. 滴灌灌水器復(fù)雜流道局部阻力特征的試驗(yàn)研究[J]. 中國農(nóng)村水利水電,2011(6):55-60.

      Liu Jie, Wei Qingsong, Shi Yusheng, et al. Experimental research on local resistance characteristics of drip irrigation emitters with complex flow channel[J]. China Rural Water and Hydropower, 2011(6): 55-60. (in Chinese with English abstract)

      [3] Patil S S, Nimbalkar P T, Joshi A. Hydraulic study, design & analysis of different geometries of drip irrigation emitter labyrinth[J]. International Journal of Engineering and Advanced Technology, 2013, 2(5): 455-462.

      [4] De Jesus Souza W, Sinobas L R, Sánchez R, et al. Prototype emitter for use in subsurface drip irrigation: Manufacturing, hydraulic evaluation and experimental analyses[J]. Biosystems Engineering, 2014, 128: 41-51.

      [5] 仵峰,董曉爽,吳玉博,等. 漸縮-突擴(kuò)流道水力特性研究[J]. 華北水利水電大學(xué)學(xué)報(bào):自然科學(xué)版,2017,38(2):61-67.

      Wu Feng, Dong Xiaoshuang, Wu Yubo, et al. Hydraulic properties of the flow in a gradual shrinking and sudden enlarging channel[J]. Journal of North China University of Water Resources and Electric Power: Natural Science Edition, 2017, 38(2): 61-67. (in Chinese with English abstract)

      [6] GB/T 50485-2009,微灌工程技術(shù)規(guī)范[S].

      [7] Tayel M, Lightfoot D, Mansour H. Effects of drip irrigation circuit design and lateral line lengths: I-on pressure and friction loss[J]. Agricultural Sciences, 2012, 3(3): 392-399.

      [8] Zhang Lin, Merkley G P. Relationships between common irrigation application uniformity indicators[J]. Irrigation Science, 2012, 30(2): 83-88.

      [9] 張林,吳普特,朱德蘭,等. 基于制造偏差的滴灌系統(tǒng)綜合流量偏差率[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2013,44(12):135-139.

      Zhang Lin, Wu Pute, Zhu Delan, et al. Integrated flow deviation rate of drip irrigation system based on manufacturing variation[J].Transactions of the Chinese Society for Agricultural Machinery, 2013, 44(12): 135-139. (in Chinese with English abstract)

      [10] 魏正英,苑偉靜,周興,等. 我國壓力補(bǔ)償灌水器的研究進(jìn)展[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2014,45(1):94-101.

      Wei Zhengying, Yuan Weijing, Zhou Xing, et al. Research progress of pressure compensating emitters in micro-irrigation systems in china[J]. Transactions of the Chinese Society for Agricultural Machinery, 2014, 45(1): 94-101. (in Chinese with English abstract)

      [11] Madramootoo C A, Morrison J. Advances and challenges with micro-irrigation[J]. Irrigation and Drainage, 2013, 62(3): 255-261.

      [12] Gilaad Y, Krystal L, Zanker K.Hydraulic and mechanical properties of drippers[C]//Proceedings of the 2nd International Drip Irrigation Congress. Riverside, USA: University of California, 1974.

      [13] 喻黎明. 結(jié)構(gòu)參數(shù)對(duì)梯形流道水力性能及抗堵塞性能的影響[J]. 西北農(nóng)林科技大學(xué)學(xué)報(bào):自然科學(xué)版,2011,39(8):197-202.

      Yu Liming. Influence of the structural parameters of trapezoidal-channel emitters on hydraulic and anti-clogging performance[J]. Journal of Northwest A&F University: Natural Science Edition, 2011, 39(8): 197-202. (in Chinese with English abstract)

      [14] 杜少卿,曾文杰,施澤,等. 工作壓力對(duì)滴灌管迷宮流道灌水器水力性能的影響[J]. 農(nóng)業(yè)工程學(xué)報(bào),2011,27(增刊2):55-60.

      Du Shaoqing, Zeng Wenjie, Shi Ze, et al. Effects of working pressure on hydraulic performances of labyrinth path emitters[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2011, 27(Supp.2): 55-60. (in Chinese with English abstract)

      [15] 郭霖,白丹,王新端,等.雙向?qū)_流滴灌灌水器水力性能與消能效果[J]. 農(nóng)業(yè)工程學(xué)報(bào),2016,32(17):77-82.

      Guo Lin, Bai Dan, Wang Xinduan, et al. Hydraulic performance and energy dissipation effect of two-ways mixed flow emitter in drip irrigation[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2016, 32(17): 77-82. (in Chinese with English abstract)

      [16] 郭霖,白丹,王新端,等. 雙向?qū)_流灌水器水力性能和消能機(jī)理模擬與驗(yàn)證[J]. 農(nóng)業(yè)工程學(xué)報(bào),2017,33(14):100-107.

      Guo Lin, Bai Dan, Wang Xinduan, et al. Numerical simulation and verification of hydraulic performance and energy dissipation mechanism of two-ways mixed flow emitter[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(14): 100-107. (in Chinese with English abstract)

      [17] 范永申,仵峰,宰松梅,等. 新型微壓滴灌灌水器水力性能試驗(yàn)研究[J]. 灌溉排水學(xué)報(bào),2006,25(5):39-41.

      Fan Yongshen, Wu Feng, Zai Songmei, et al. Experimental study on hydraulic property of emitters with tiny water pressure[J]. Journal of Irrigation and Drainage, 2006, 25(5): 39-41. (in Chinese with English abstract)

      [18] 蔡耀輝,吳普特,張林,等. 設(shè)計(jì)流量和土壤質(zhì)地對(duì)微孔陶瓷灌水器入滲特性的影響[J]. 農(nóng)業(yè)工程學(xué)報(bào),2017,33(7):100-106.

      Cai Yaohui, Wu Pute, Zhang Lin, et al. Effects of designed flow rate and soil texture on infiltration characteristics of porous ceramic irrigation emitters[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2017, 33(7): 100-106. (in Chinese with English abstract)

      [19] 黎安,劉潔,魏青松,等. 滴灌灌水器多類型流道的微壓水力性能研究[J]. 中國農(nóng)村水利水電,2012(2):4-7.

      Li An, Liu Jie, Wei Qingsong, et al. The hydraulic properties of multiple flow channels of drip emitters under micro-pressure[J]. China Rural Water and Hydropower, 2012(2): 4-7. (in Chinese with English Abstract)

      [20] Hezarjaribi A, Dehghani A A, Helghi M M, et al. Hydraulic performances of various trickle irrigation emitters[J]. Journal of Agronomy, 2008, 7(3): 265-271.

      [21] 李治勤,馬靜. 迷宮灌水器水流流態(tài)試驗(yàn)[J]. 農(nóng)業(yè)工程學(xué)報(bào),2012,28(1):82-86.

      Li Zhiqin, Ma Jing. Experiment on flow pattern in labyrinth emitter[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2012, 28(1): 82-86. (in Chinese with English abstract)

      [22] 田濟(jì)揚(yáng),白丹,任長江,等. 滴灌雙向流流道灌水器水力特性分析[J]. 農(nóng)業(yè)工程學(xué)報(bào),2013,29(20):89-94.

      Tian Jiyang, Bai Dan, Ren Changjiang, et al. Analysis on hydraulic performance of bidirectional flow channel of drip irrigation emitter[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2013, 29(20): 89-94. (in Chinese with English abstract)

      [23] 李向明,楊建國. 微孔陶瓷灌水器流量影響因素研究[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2016,47(4):73-78.

      Li Xiangming, Yang Jianguo. Factors influencing flow rate of microporous ceramic irrigation emitters[J]. Transactions of the Chinese Society for Agricultural Machinery, 2016, 47(4): 73-78. (in Chinese with English abstract)

      [24] 劉春景,唐敦兵,鄭加強(qiáng),等. 滴灌梯形迷宮滴頭流道水力性能的響應(yīng)曲面法優(yōu)化[J]. 農(nóng)業(yè)工程學(xué)報(bào),2011,27(2):46-51.

      Liu Chunjing, Tang Dunbing, Zheng Jiaqiang, et al. Optimization of hydraulic performance for drip irrigation trapezoidal labyrinth channel of emitter using response surface methodology[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2011, 27(2): 46-51. (in Chinese with English abstract)

      [25] 劉春景,唐敦兵,何華,等. 滴灌三角形迷宮滴頭水力性能穩(wěn)健性分析[J]. 農(nóng)業(yè)機(jī)械學(xué)報(bào),2013,44(1):67-72.

      Liu Chunjing, Tang Dunbing, He Hua, et al. Hydraulic performance robustness analysis for drip irrigation triangle labyrinth channel of emitter[J]. Transactions of the Chinese Society for Agricultural Machinery, 2013, 44(1): 67-72. (in Chinese with English abstract)

      [26] Li Guangyong, Wang Jiandong, Alam M, et al. Influence of geometrical parameters of labyrinth flow path of drip emitters on hydraulic and anti-clogging performance[J]. Transactions of the ASABE, 2006, 49(3): 637-643.

      [27] Li Yunkai, Yang Peiling, Xu Tingwu, et al. Hydraulic property and flow characteristics of three labyrinth flow paths of drip irrigation emitters under micro-pressure[J]. Transactions of the ASABE, 2009, 52(4): 1129-1138.

      [28] Celik H K, Karayel D, Caglayan N, et al. Rapid prototyping and ?ow simulation applications in design of agricultural irrigation equipment: Case study for a sample in-line drip emitter[J]. Virtual and Physical Prototyping, 2011, 6(1): 47-56.

      [29] Vekariya P B, Subbaiah R, Mashru H H. Hydraulics of microtube emitters: A dimensional analysis approach[J]. Irrigation Science, 2011, 29(4): 341-350.

      [30] Demir V, Yurdem H, Degirmencioglu A. Development of prediction models for friction losses in drip irrigation laterals equipped with integrated in-line and on-line emitters using dimensional analysis[J]. Biosystems Engineering, 2007, 96(4): 617-631.

      [31] 劉春景,鄭加強(qiáng),周宏平,等. 滴灌滴頭水力性能預(yù)測模型[J]. 水利學(xué)報(bào),2010,41(11):1353-1359.

      Liu Chunjing, Zheng Jiaqiang, Zhou Hongping, et al. Prediction model for hydraulic performance of trickle irrigation emitters[J]. Journal of Hydraulic Engineering, 2010, 41(11): 1353-1359. (in Chinese with English abstract)

      [32] 王建國,張文興. 支持向量機(jī)建模及其智能優(yōu)化[M]. 北京:清華大學(xué)出版社,2015.

      [33] Vapnik V N. The Nature of Statistical Learning Theory[M]. New York: Springer-Verlag, 1999.

      [34] Deo R C, Kisi O, Singh V P. Drought forecasting in eastern Australia using multivariate adaptive regression spline, least square support vector machine and M5Tree model[J]. Atmospheric Research, 2017, 184: 149-175.

      [35] 楊紹鍔,黃元仿. 基于支持向量機(jī)的土壤水力學(xué)參數(shù)預(yù)測[J]. 農(nóng)業(yè)工程學(xué)報(bào),2007,23(7):42-47.

      Yang Shao’e, Huang Yuanfang. Prediction of soil hydraulic characteristic parameters based on support vector machine[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2007, 23(7): 42-47. (in Chinese with English abstract)

      [36] 張強(qiáng),黃生志,陳曉宏. 基于支持向量機(jī)的土壤濕度模擬及預(yù)測研究[J]. 土壤學(xué)報(bào),2013,50(1):59-67.

      Zhang Qiang, Huang Shengzhi, Chen Xiaohong. Simulation and prediction of soil moisture based on support vector machine technique[J]. Acta Pedologica Sinica, 2013, 50(1): 59-67. (in Chinese with English abstract)

      [37] 雷國慶,樊貴盛. 基于支持向量機(jī)的土壤水分入滲參數(shù)預(yù)測研究[J]. 節(jié)水灌溉,2015(12):28-30.

      Lei Guoqing, Fan Guisheng. Research on soil infiltration parameters prediction based on support vector machine[J]. Water Saving Irrigation, 2015(12): 28-30. (in Chinese with English abstract)

      [38] Guo Lin, Bai Dan, Zhou Wen, et al. Evaluation of numerical simulation accuracy for two-ways mixed flow drip irrigation emitter based on CFD[J]. International Journal of Heat and Technology, 2017, 35(2): 384-392.

      [39] Anguita D, Ridella S, Rivieccio F, et al. Hyperparameter design criteria for support vector classifiers[J]. Neurocomputing, 2003, 55(1): 109-134.

      [40] Tharwat A, Hassanien A E, Elnaghi B E. A ba-based algorithm for parameter optimization of support vector machine[J]. Pattern Recognition Letters, 2017, 93: 13-22.

      Establishment and validation of flow rate prediction model for drip irrigation emitter based on support vector machine

      Guo Lin1, Bai Dan1※, Wang Xinduan1, Wang Cheng1, Zhou Wen2, Cheng Peng2

      (1.,,71004,; 2.,,450011,)

      To carry out the prediction and calculation of the flow rate for further study the hydraulic performance and the structure optimization of the flow channel in drip irrigation emitter is of great significance. In order to predict and calculate the flow rate of the emitter accurately, in this study, the prediction and calculation method of Support Vector Machine (SVM) with strong generalization ability was introduced, and the flow rate prediction model of the SVM was built. We chose six working pressures and eight geometric parameters of the flow channel as factors, and arranged 300 sets of emitter schemes as training sample of SVM according to the orthogonal experimental design method, and 30 sets of schemes as test sample. Based on these, the prediction model sample set of flow rate of SVM was established. The flow rate of the emitter was simulated by the SST k-ω model with high precision in the sample set, and compared with the predicted value of flow rate of the SVM. The pressure and geometric parameter of the emitter was taken as the input item, and the flow rate was taken as the output item of SVM. The prediction and simulation of the flow rate were carried out in State Key Laboratory Base of Eco-hydraulic Engineering in Arid Area, Xi’an University of Technology. In order to eliminate the impact of each factor on the predicted results, the input and output item in the emitter sample were normalized before predicting flow rate. At the same time, the Genetic Algorithm was used to optimize the C and δ parameter in the Radial Basis Function (RBF) kernel of the SVM, and then the minimum error between the predicted value and simulated value of flow rate was obtained. The results showed that the relative error between the predicted value of flow rate using SVM and the simulated value was from 0.09% to 6.43%, the average relative error was 1.91%, and the determination coefficient was 0.98 when the optimal values of SVM parameterandwere 100 and 20, respectively. The predicted value of flow rate of SVM had a good correlation with the simulated value, which satisfied the predicted demand for the flow rate of the emitter. However, when the regression fitting method was adopted and calculated, the relative error between the predicted value and the simulated value was from 0.15% to 26.69%, the average relative error was 6.45%, and the determination coefficient was 0.93, which indicated excellent superiority based on SVM. To further verify the reliability of SVM, the five experimental verification schemes were chosen, and manufactured by using high-precision engraving technology. The flow rate value of experimental verification sample was tested under different pressure range, and was compared with the predicted value of flow rate. The relative error between the predicted value of flow rate using SVM and the experimental value was from 0.14% to 5.13%, and the average relative error was 2.25%, which were within the error range, verifying the accuracy and reliability of predicting flow rate using SVM. The establishment of the flow rate prediction response surface based on SVM can effectively improve the development efficiency of the emitter, and provide the evidence and guidance for the hydraulic performance evaluation, the flow channel structure design and optimization.

      flow rate; numerical analysis; models; drip irrigation emitter; working pressure; geometric parameter; support vector machine; optimization

      10.11975/j.issn.1002-6819.2018.02.010

      S275.6

      A

      1002-6819(2018)-02-0074-09

      2017-10-17

      2017-12-20

      國家自然科學(xué)基金資助項(xiàng)目(51279156、41571222);高等學(xué)校博士學(xué)科點(diǎn)專項(xiàng)科研基金聯(lián)合資助課題(20116118110010)

      郭 霖,博士生,主要從事節(jié)水灌溉技術(shù)研究。Email:guolinedu@126.com

      白 丹,教授,博士生導(dǎo)師,主要從事節(jié)水灌溉理論與技術(shù)研究。Email:baidan@xaut.edu.cn

      郭 霖,白 丹,王新端,王 程,周 文,程 鵬. 基于支持向量機(jī)的滴灌灌水器流量預(yù)測模型建立與驗(yàn)證[J]. 農(nóng)業(yè)工程學(xué)報(bào),2018,34(2):74-82. doi:10.11975/j.issn.1002-6819.2018.02.010 http://www.tcsae.org

      Guo Lin, Bai Dan, Wang Xinduan, Wang Cheng, Zhou Wen, Cheng Peng. Establishment and validation of flow rate prediction model for drip irrigation emitter based on support vector machine[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(2): 74-82. (in Chinese with English abstract) doi:10.11975/j.issn.1002-6819.2018.02.010 http://www.tcsae.org

      猜你喜歡
      訓(xùn)練樣本水力灌水
      水力全開
      灌水取球
      番茄灌水掌握技巧
      冬季棚菜灌水四關(guān)鍵
      人工智能
      球墨鑄鐵管的水力計(jì)算
      寬帶光譜成像系統(tǒng)最優(yōu)訓(xùn)練樣本選擇方法研究
      融合原始樣本和虛擬樣本的人臉識(shí)別算法
      基于稀疏重構(gòu)的機(jī)載雷達(dá)訓(xùn)練樣本挑選方法
      水力噴射壓裂中環(huán)空水力封隔全尺寸實(shí)驗(yàn)
      芒康县| 道孚县| 青冈县| 德州市| 盐亭县| 西盟| 碌曲县| 杭锦旗| 垣曲县| 凌源市| 海门市| 铅山县| 孝感市| 东安县| 龙江县| 靖宇县| 连云港市| 潍坊市| 尚志市| 房产| 濉溪县| 安达市| 阳泉市| 阿瓦提县| 蓝田县| 电白县| 九江县| 寿阳县| 尚志市| 柘城县| 庆元县| 富蕴县| 望城县| 唐山市| 连州市| 高台县| 松阳县| 资源县| 大同市| 黄梅县| 延边|