農(nóng)業(yè)大棚溫室智能化自動(dòng)控制中英文對(duì)照外文翻譯文獻(xiàn)_第1頁(yè)
農(nóng)業(yè)大棚溫室智能化自動(dòng)控制中英文對(duì)照外文翻譯文獻(xiàn)_第2頁(yè)
農(nóng)業(yè)大棚溫室智能化自動(dòng)控制中英文對(duì)照外文翻譯文獻(xiàn)_第3頁(yè)
農(nóng)業(yè)大棚溫室智能化自動(dòng)控制中英文對(duì)照外文翻譯文獻(xiàn)_第4頁(yè)
農(nóng)業(yè)大棚溫室智能化自動(dòng)控制中英文對(duì)照外文翻譯文獻(xiàn)_第5頁(yè)
已閱讀5頁(yè),還剩20頁(yè)未讀, 繼續(xù)免費(fèi)閱讀

下載本文檔

版權(quán)說(shuō)明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請(qǐng)進(jìn)行舉報(bào)或認(rèn)領(lǐng)

文檔簡(jiǎn)介

農(nóng)業(yè)大棚溫室智能化自動(dòng)控制中英文對(duì)照外文翻譯文獻(xiàn)農(nóng)業(yè)大棚溫室智能化自動(dòng)控制中英文對(duì)照外文翻譯文獻(xiàn)(文檔含英文原文和中文翻譯)翻譯:農(nóng)業(yè)大棚溫室智能化自動(dòng)控制摘要:確定控制溫室作物生長(zhǎng)歷來(lái)使用約束優(yōu)化或應(yīng)用人工智能技術(shù),解決了軌跡的問(wèn)題。已被用作經(jīng)濟(jì)利潤(rùn)的最優(yōu)化研究的主要標(biāo)準(zhǔn),以獲得足夠的作物生長(zhǎng)的氣候控制設(shè)定值。本文針對(duì)溫室作物生長(zhǎng)的問(wèn)題,通過(guò)分層控制體系結(jié)構(gòu)由一個(gè)高層次的多目標(biāo)優(yōu)化方法,在解決這個(gè)問(wèn)題的辦法是找到白天和夜間溫度參考軌跡(氣候相關(guān)的設(shè)定值)和電導(dǎo)率(fertirrigation相關(guān)設(shè)定值)。的目標(biāo)是利潤(rùn)最大化,果實(shí)品質(zhì),水分利用效率,這些目前正在培育的國(guó)際規(guī)則。結(jié)果說(shuō)明,選擇從那些獲得工業(yè)的溫室,在過(guò)去的八年中示出和描述。關(guān)鍵詞:農(nóng)業(yè);分層系統(tǒng);過(guò)程控制;優(yōu)化方法;產(chǎn)量?jī)?yōu)化1介紹現(xiàn)代農(nóng)業(yè)是時(shí)下在質(zhì)量和環(huán)境影響方面的規(guī)定,因此它是一個(gè)自動(dòng)控制技術(shù)的應(yīng)用領(lǐng)域已經(jīng)增加了很多,在過(guò)去的幾年里,溫室產(chǎn)生的空氣系統(tǒng)的是一個(gè)復(fù)雜的物理,化學(xué)和生物學(xué)過(guò)程,同時(shí),使具有不同的響應(yīng)時(shí)間和模式的環(huán)境因素,其特征在于由許多相互作用,它必須加以控制,以以獲得最佳效果的種植者。作物生長(zhǎng)過(guò)程是最重要的,主要受周圍環(huán)境的氣候變量(光合有效輻射-PAR,溫度,濕度,二氧化碳濃度,里面的空氣),水和化肥,灌溉,病蟲害提供量,和文化的勞動(dòng)力,如修剪和農(nóng)藥的治療等等。溫室是適合作物生長(zhǎng),因?yàn)樗鼧?gòu)成了一個(gè)封閉的環(huán)境中,可以控制氣候和肥料灌溉變量。氣候和肥料灌溉是兩個(gè)獨(dú)立的系統(tǒng),不同的控制問(wèn)題和目標(biāo)。根據(jù)經(jīng)驗(yàn),不同作物品種的水和養(yǎng)分的要求是已知的,在實(shí)際上,第一個(gè)自動(dòng)化系統(tǒng)控制這些變量。另一方面,市場(chǎng)價(jià)格的波動(dòng)和環(huán)境的規(guī)則,以提高水的利用效率或其他方面加以考慮,減少肥料殘留在土壤中的(如硝酸鹽含量)。因此,優(yōu)化生產(chǎn)過(guò)程,可概括為一個(gè)溫室大氣系統(tǒng)的問(wèn)題,達(dá)到以下目標(biāo):的最佳作物生長(zhǎng)(一個(gè)更大的生產(chǎn)與質(zhì)量更好)聯(lián)營(yíng)公司的成本(主要是燃料,電力和化肥,減少),減少殘留物(主要是殺蟲劑和離子在土壤中),和水的利用效率的提高。許多方法已被應(yīng)用到這個(gè)問(wèn)題,例如,處理的溫室氣候管理中的最優(yōu)控制字段。2M0優(yōu)化作物生產(chǎn)一個(gè)MO優(yōu)化問(wèn)題可以定義為尋找決策變量的向量,它滿足約束條件和優(yōu)化的目標(biāo)函數(shù)一個(gè)向量,其元素。特點(diǎn)是競(jìng)爭(zhēng)的措施,表現(xiàn)或目標(biāo)的問(wèn)題被視為MO優(yōu)化問(wèn)題,其中n目標(biāo)姬(p)在變量的向量P∈P的同時(shí)最小化或最大化。問(wèn)題往往沒(méi)有最佳的解決方案,同時(shí)優(yōu)化所有目標(biāo),但它有一組作為一個(gè)Pareto最優(yōu)集。其中一個(gè)折衷的解決方案可以選自已知的不理想的或不占主導(dǎo)地位的替代解決方案設(shè)置一個(gè)決策過(guò)程。不同的標(biāo)準(zhǔn),如物理產(chǎn)量,作物品質(zhì),產(chǎn)品質(zhì)量,生產(chǎn)過(guò)程中的時(shí)間不同,或生產(chǎn)成本和風(fēng)險(xiǎn),可配制于溫室作物管理。這些標(biāo)準(zhǔn)往往會(huì)產(chǎn)生有爭(zhēng)議的的氣候和肥料灌溉要求,必須要解決的或明或暗地在所謂的戰(zhàn)術(shù)層面上,種植者有幾個(gè)相互沖突的目標(biāo)做出決定。該解決方案的這個(gè)MO優(yōu)化過(guò)程,的p∈P,是最佳的日間和夜間的當(dāng)前和未來(lái)的參考軌跡的溫度,XTA,導(dǎo)電性,XEC,作物周期的其余部分。即,沿著優(yōu)化的時(shí)間間隔內(nèi)的空氣溫度是一個(gè)向量,并沿著優(yōu)化的時(shí)間間隔的電導(dǎo)率(EC)是一個(gè)矢量。請(qǐng)注意,在植物生長(zhǎng)的PAR輻射(晝夜的條件)的影響下,進(jìn)行光合作用過(guò)程。此外,溫度成為影響糖的生產(chǎn)速度通過(guò)光合作用,從而輻射和溫度具有較高的輻射水平的方式,對(duì)應(yīng)于較高的溫度達(dá)到平衡。所以,在晝夜條件下的溫度維持在較高的水平是必要的。在夜間條件下的植物都沒(méi)有激活(作物不生長(zhǎng)),所以它不是必要的,以維持這樣高的溫度。出于這個(gè)原因,通常被認(rèn)為是兩個(gè)溫度設(shè)定點(diǎn):日間和夜間。這是必要的,以反白顯示,雖然在連續(xù)時(shí)間的過(guò)程優(yōu)化,解決了在離散的時(shí)間間隔為一個(gè)優(yōu)化地平線化,且(k)項(xiàng)(該層是可變的,代表剩余的時(shí)間段,直到結(jié)束的農(nóng)業(yè)季節(jié))。因此,解向量,其中k是當(dāng)前離散時(shí)間瞬間獲得。需要注意的是,對(duì)于提出的優(yōu)化問(wèn)題,溫室作物生產(chǎn)的模型是必需的,以估計(jì)內(nèi)的的氣候行為和作物的生長(zhǎng),該算法通過(guò)不同的步驟,并涉及不同的功能目標(biāo)決策變量。溫室內(nèi)的微氣候的動(dòng)態(tài)行為是涉及能量轉(zhuǎn)移(輻射和熱)和質(zhì)量平衡(水蒸汽通量和二氧化碳濃度)的物理過(guò)程的組合。另一方面,主要取決于作物的生長(zhǎng)和產(chǎn)量,在其他情況中,如灌溉和化肥,在溫室內(nèi)的溫度,PAR輻射,CO2濃度。因此,無(wú)論是氣候條件和作物生長(zhǎng)的相互影響,其動(dòng)態(tài)行為特征,可以通過(guò)不同的時(shí)間尺度。其中XCL=XCL(t)是一個(gè)n1的維向量的溫室氣候狀態(tài)變量的(主要的內(nèi)部空氣的溫度和濕度,二氧化碳濃度,PAR輻射,土壤表面溫度,蓋溫度,和植物溫度),XGR=XGR(叔)是作物生長(zhǎng)狀態(tài)變量(主要是數(shù)量的主莖上,葉面積指數(shù)(LAI)或表面土壤面積的葉片,總干物質(zhì)代表所有植物成分的根,莖節(jié)點(diǎn)N2-維向量,葉,花和果實(shí),不包括水,水果干物質(zhì)生物量的水果,不包括水,和成熟的果實(shí)干物質(zhì)或成熟果實(shí)生物量的積累),U=U(t)是m維向量輸入變量(天然通風(fēng)孔和加熱系統(tǒng),在這項(xiàng)工作中),D=D(t)是干擾(外界溫度,濕度,風(fēng)速和風(fēng)向,室外輻射,雨)鄰維向量,V=V(t)的一類q維向量,系統(tǒng)變量的(蒸騰,縮合,和其他進(jìn)程有關(guān)),系統(tǒng)常數(shù),C是r維向量,t是時(shí)間,XCL,i和XGR,在初始時(shí)刻ti,i是已知的狀態(tài)整箱整箱(t)是一個(gè)非線性函數(shù)的基礎(chǔ)上的傳質(zhì)和傳熱的結(jié)余的fgr=的fgr(t)是一個(gè)非線性函數(shù)的基礎(chǔ)上的植物的基本的生理過(guò)程。地中海地區(qū),已開發(fā)了線性和非線性模型的物理定律。這些模型可以發(fā)現(xiàn)深解釋拉米雷斯·阿里亞斯,羅德里格斯,Berenguel和費(fèi)爾南德斯(水模),拉米雷斯-阿里亞斯等。(增長(zhǎng)模型),羅德里格斯等人。(氣候模式),羅德里格斯和Berenguel。這些模型過(guò)于復(fù)雜,這里詳述,但主要的增長(zhǎng)模型方程問(wèn)題的目標(biāo)和最終的MO優(yōu)化問(wèn)題的解釋在下面的章節(jié)將描述。這些方程將用來(lái)展示如何在不同的目標(biāo)(成本函數(shù))表示為決策變量的函數(shù)的優(yōu)化問(wèn)題(目前和未來(lái)的溫度和EC的設(shè)定值)。2.1利潤(rùn)最大化利潤(rùn)的計(jì)算作為新鮮水果的銷售收入,并關(guān)聯(lián)到他們的生產(chǎn)成本之間的差異VPR(t)是產(chǎn)量估計(jì)從市場(chǎng)的銷售價(jià)格,XFFP(t)是獲得作物生長(zhǎng)模型的VCO(T)的新鮮水果生產(chǎn),所產(chǎn)生的費(fèi)用由供熱,電力,化肥,水,t是時(shí)間,ti是作物周期的初始時(shí)間,th是最新的收獲時(shí)間,同時(shí)選擇由種植者。請(qǐng)注意,在實(shí)踐中,有多個(gè)番茄作物收獲在生長(zhǎng)季節(jié)。出于這個(gè)原因,日式代表了最新的收獲時(shí)間。

另一種方法是考慮在未來(lái)的收獲時(shí)間(TN),成本函數(shù),并再次重新啟動(dòng)優(yōu)化過(guò)程,一旦前收割工作已經(jīng)產(chǎn)生。這兩種替代品的有效期為多收獲。收入取決于番茄果實(shí)的價(jià)格(千克-1,€公斤-1),收獲日期,并在每表面單位鮮重的產(chǎn)量(公斤米2)。價(jià)格政策需要市場(chǎng)模型或歷史數(shù)據(jù),這是一個(gè)非常困難的預(yù)測(cè)問(wèn)題。下面的小節(jié)描述如何新鮮水果生產(chǎn),XFFP(T),以及工藝成本,壓控振蕩器(T),可以預(yù)計(jì)相關(guān)的決策變量。2.2質(zhì)量最大化利潤(rùn)最大化,雖然可以被理解為主要目標(biāo)從種植者的角度來(lái)看,這不能總是被用來(lái)作為唯一的一個(gè)。種植者通常屬于合作社或農(nóng)業(yè)社會(huì),有利于引入園藝產(chǎn)品進(jìn)入市場(chǎng)。這些協(xié)會(huì)修復(fù)的政策,優(yōu)質(zhì)的產(chǎn)品,根據(jù)不同的市場(chǎng)需求,因此,種植者必須適應(yīng)其生產(chǎn)這些政策的過(guò)程中,為了達(dá)到一些最低限度的質(zhì)量水平。食品質(zhì)量擁抱感覺(jué)屬性很容易察覺(jué)到人的感官和隱藏屬性,如健康和營(yíng)養(yǎng)。在水果和蔬菜的感官性能由糖類,有機(jī)酸,揮發(fā)性化合物的量,以及顏色,形狀和紋理。然而,糖和酸那些反映整體一個(gè)水果口味喜好。對(duì)于番茄作物,可溶性固形物已涉及到糖和可滴定酸度主要有機(jī)酸,因此它們可以作為果實(shí)品質(zhì)的指標(biāo)。堅(jiān)定的水果是另一種重要的質(zhì)量參數(shù)鏈中的種植者經(jīng)銷商消費(fèi)者。然而,一些作品已經(jīng)表明,園藝蔬菜,如西紅柿或鮮花,感官質(zhì)量的一些重要參數(shù)是在沖突與產(chǎn)量。番茄果實(shí)可溶性固形物,滴定酸度,果實(shí)硬度和大小可以使用下面的線性方法([Dorais等人,2001年XTA(t)和XEC(T)(決策變量)]和[Sonneveld和面包車博格,1991])(15)Y(T)=A+B(X(T)-G(X(T)))其中Y(t)為變量的計(jì)算(可溶性固形物,滴定酸度,果實(shí)硬度,或大?。?,X(t)是相關(guān)的決策變量(XEC為VSSol(T)(T),腹側(cè)被蓋區(qū)(T),vfs的(t)的和XTA(??t)的VFF(t))的,在Y(t)的系數(shù),是一個(gè)常數(shù)增量,b為增量在Y(t)的系數(shù),在X(t)的單位的增量,并G(X(t))的代表在Y(T),其中有一個(gè)增量的X(t)的閾值是一個(gè)分段函數(shù)。2.3水利用效率的最大化這個(gè)目標(biāo)優(yōu)化問(wèn)題明確納入環(huán)境有關(guān)的目的。在半干旱的氣候,如地中海的,水是非常稀缺和昂貴的資源,主要是在一些一年四季。有些作者認(rèn)為,在這樣的地區(qū),是由生產(chǎn)力可用的水和用水效率使用。這樣,適當(dāng)管理的水是必需的。與顯式包含這一目標(biāo),種植者可以選擇提供的期望的耗水量,在生長(zhǎng)周期從帕累托前沿的解決方案。這一目標(biāo)的嘗試使用的水量足以作物生長(zhǎng)發(fā)育的密切關(guān)系,所提供的營(yíng)養(yǎng)液的濃度。在本文中,水分利用效率被認(rèn)為是類似的生物量的效率之間的關(guān)系定義為新鮮水果的物質(zhì)生產(chǎn)與供給的水。2.4多目標(biāo)優(yōu)化問(wèn)題所有這些目標(biāo)中的變量是空氣溫度,XTA和/或歐盟,XEC,(XFFP(T)的FSF(T),西南(T),腹側(cè)被蓋區(qū)(T),VSSol(T)的功能,VFS(T),VFF(T)),以及衡量的干擾,如PAR輻射或二氧化碳濃度。也就是說(shuō),目標(biāo)函數(shù)可以表示為對(duì)于i=1,2,3,是沿著優(yōu)化的時(shí)間間隔內(nèi)的空氣溫度的向量是一個(gè)向量沿著優(yōu)化的時(shí)間間隔的EC,Θ是一個(gè)向量測(cè)的擾動(dòng)具有沿水平優(yōu)化預(yù)測(cè)。MO優(yōu)化問(wèn)題的解決提供了歐共體內(nèi)的空氣溫度控制地平線其余的日間和夜間的設(shè)定軌跡。恒定的日間和夜間的設(shè)定點(diǎn)定義,穩(wěn)定狀態(tài)模型的溫室氣候和番茄作物,總結(jié)在Eqs.Although幾種技術(shù)已被評(píng)估為解決MO優(yōu)化問(wèn)題,在這種情況下,一個(gè)目標(biāo)實(shí)現(xiàn)算法已被用于(序貫二次規(guī)劃SQP)。確定每個(gè)目標(biāo)的重點(diǎn),通過(guò)使用權(quán)重,按順序在每個(gè)迭代修改。的約束被定義為從專家的知識(shí)獲得的最大和最小的溫度和EC值表明“最佳”番茄的生長(zhǎng)溫度和通過(guò)分析局部數(shù)據(jù)從歷史系列。由此產(chǎn)生的約束條件改變整個(gè)每年的時(shí)間與過(guò)去的二十年收集的數(shù)據(jù)的基礎(chǔ)上設(shè)計(jì)的圖案。3多級(jí)遞階控制結(jié)構(gòu)動(dòng)態(tài)參與溫室生產(chǎn)過(guò)程中呈現(xiàn)出不同的時(shí)間尺度上,如上所述,即內(nèi)部溫室氣候,作物快速動(dòng)力學(xué)(即蒸騰作用,光合作用和呼吸作用),和緩慢的的作物發(fā)育(即作物生長(zhǎng)和果實(shí)的變化)。因此,多層分級(jí)控制架構(gòu)已經(jīng)提出并使用(Rodriguez等人,2003年和羅德里格斯等人,2008])3.1作物生長(zhǎng)控制層考慮到長(zhǎng)期目標(biāo)(市場(chǎng)價(jià)格,收獲日期和所需的質(zhì)量)和長(zhǎng)期預(yù)測(cè)的增長(zhǎng)狀態(tài),使用修改后的模型(拉米雷斯-阿里亞斯等人,2004)進(jìn)行優(yōu)化計(jì)算的溫室內(nèi)溫度的設(shè)定值軌跡和歐盟一起考慮控制范圍內(nèi)(通常是65天為一個(gè)淡旺季-260決策變量-或120天為一個(gè)漫長(zhǎng)的賽季-480決策變量)。灌溉模型也已開發(fā),控制和優(yōu)化的目的。長(zhǎng)期天氣預(yù)測(cè),這是邏輯上具有較高程度的不確定性的要素之一,是使用一個(gè)軟件工具,訪問(wèn)由西班牙國(guó)家氣象局的天氣預(yù)測(cè),未來(lái)八天向前,產(chǎn)生模式在幾個(gè)指標(biāo)(清晰度,最大,平均和最低氣溫,太陽(yáng)輻射),在本地搜索歷史氣候序列數(shù)據(jù)庫(kù)生成模式,更好地適合。以這種方式,以所選擇的序列作為短期天氣預(yù)報(bào),估計(jì)作物周期的其余部分被從該短序列和使用從歷史數(shù)據(jù)庫(kù)中的一個(gè)數(shù)據(jù)窗口生成。通過(guò)滾動(dòng)的方法,在第二層進(jìn)行修改,降低不確定性的相關(guān)程度高。3.2設(shè)定適應(yīng)層在這一層中,被發(fā)送到下層為第二天的設(shè)定值被修改和更新,以避免不可行性問(wèn)題,并允許達(dá)到參考值??紤]在上層,短期內(nèi)的天氣預(yù)報(bào)(具有較低程度的不確定性),當(dāng)前狀態(tài)的作物產(chǎn)生的軌跡,這些修改和短期種植者目標(biāo)(考慮到他/她的技能和作物狀態(tài),這是必要的自由度,讓種植者的分層控制系統(tǒng)進(jìn)行交互)。然后,該信息是用在上面描述的模型,以模擬的溫室的行為,并評(píng)價(jià),如果所提供的設(shè)定點(diǎn)可以達(dá)到。在優(yōu)化過(guò)程被重復(fù)修改(減少或增加設(shè)定值),根據(jù)仿真結(jié)果的約束。當(dāng)設(shè)定點(diǎn)是可到達(dá)的,它們被發(fā)送到下層。3.3氣候控制和營(yíng)養(yǎng)層從上層使用的溫度和EC設(shè)定點(diǎn),控制器計(jì)算的適當(dāng)?shù)目刂菩盘?hào),致動(dòng)器。所開發(fā)的控制算法包括范圍寬,從饋控制,自適應(yīng)控制,預(yù)測(cè)控制,混合控制。這顯然是有限的引用列表和溫度控制上的許多重要文件都沒(méi)有提到,由于空間的限制。4。結(jié)論在這項(xiàng)工作中,一個(gè)MO優(yōu)化問(wèn)題已經(jīng)提出,溫室作物生長(zhǎng)管理測(cè)試,獲得三個(gè)目標(biāo):經(jīng)濟(jì)利益的最大化,果實(shí)品質(zhì),水分利用效率的折中解決方案。這個(gè)優(yōu)化方案已經(jīng)集成到一個(gè)層次的控制架構(gòu),使日間和夜間的溫度和EC通過(guò)整個(gè)作物周期(使用滾動(dòng)戰(zhàn)略)的設(shè)定值自動(dòng)生成。結(jié)果表明短期和長(zhǎng)期兩個(gè)作物周期的邏輯軌跡。在未來(lái)8年,提供實(shí)時(shí)的結(jié)果在工業(yè)溫室進(jìn)行建模,仿真,控制和優(yōu)化的溫室作物生產(chǎn)工作總結(jié)研究。原文:AgriculturalgreenhousesgreenhouseintelligentautomaticcontrolAbstract:Theproblemofdeterminingthetrajectoriestocontrolgreenhousecropgrowthhastraditionallybeensolvedbyusingconstrainedoptimizationorapplyingartificialintelligencetechniques.Theeconomicprofithasbeenusedasthemaincriterioninmostresearchonoptimizationtoobtainadequateclimaticcontrolsetpointsforthecropgrowth.Thispaperaddressestheproblemofgreenhousecropgrowththroughahierarchicalcontrolarchitecturegovernedbyahigh-levelmultiobjectiveoptimizationapproach,wherethesolutiontothisproblemistofindreferencetrajectoriesfordiurnalandnocturnaltemperatures(climate-relatedsetpoints)andelectricalconductivity(fertirrigation-relatedsetpoints).Theobjectivesaretomaximizeprofit,fruitquality,andwater-useefficiency,thesebeingcurrentlyfosteredbyinternationalrules.Illustrativeresultsselectedfromthoseobtainedinanindustrialgreenhouseduringthelasteightyearsareshownanddescribed.Keywords:Agriculture;Hierarchicalsystems;Processcontrol;Optimizationmethods;Yieldoptimization1.IntroductionModernagricultureisnowadayssubjecttoregulationsintermsofqualityandenvironmentalimpactandthusitisafieldwheretheapplicationofautomaticcontroltechniqueshasincreasedalotduringthelastfewyearsThegreenhouseproductionagrosystemisacomplexofphysical,chemicalandbiologicalprocesses,takingplacesimultaneously,reactingwithdifferentresponsetimesandpatternstoenvironmentalfactors,andcharacterizedbymanyinteractions(Challa&vanStraten,1993),whichmustbecontrolledinordertoobtainthebestresultsforthegrower.Cropgrowthisthemostimportantprocessandismainlyinfluencedbysurroundingenvironmentalclimaticvariables(PhotosyntheticallyActiveRadiation—PAR,temperature,humidity,andCO2concentrationoftheinsideair),theamountofwaterandfertilizerssuppliedbyirrigation,pestsanddiseases,andculturelaborssuchaspruningandpesticidetreatmentsamongothers.AgreenhouseisidealforcropgrowingsinceitconstitutesaclosedenvironmentinwhichclimaticandFertilizerirrigationvariablescanbecontrolled.ClimateandFertilizerirrigationaretwoindependentsystemswithdifferentcontrolproblemsandobjectives.Empirically,thewaterandnutrientrequirementsofthedifferentcropspeciesareknownand,infact,thefirstautomatedsystemswerethosethatcontrolthesevariables.Ontheotherhand,themarketpricefluctuationsandtheenvironmentrulestoimprovethewater-useefficiencyorreducethefertilizerresiduesinthesoil(suchasthenitratecontents)areotheraspectstobetakenintoaccount.Therefore,theoptimalproductionprocessinagreenhouseagrosystemmaybesummarizedastheproblemtoreachingthefollowingobjectives:anoptimalcropgrowth(abiggerproductionwithabetterquality),reductionoftheassociatecosts(mainlyfuel,electricity,andfertilizers),reductionofresidues(mainlypesticidesandionsinsoil),andtheimprovementofthewateruseefficiency.Manyapproacheshavealreadybeenappliedtothisproblem,forinstance,dealingwiththemanagementofgreenhouseclimateintheoptimalcontrolfield,e.g.Challaandvan2.MOoptimizationincropproductionAnMOoptimizationproblemcanbedefinedasfindingavectorofdecisionvariableswhichsatisfiesconstraintsandoptimizesavectorwhoseelementsrepresentobjectivefunctionsTheproblemscharacterizedbycompetingmeasuresofperformanceorobjectivesareconsideredasMOoptimizationproblems,wherenobjectivesJi(p)inthevectorofvariablesp∈Paresimultaneouslyminimized(ormaximized)。Theproblemoftenhasnooptimalsolutionthatsimultaneouslyoptimizeallobjectives,butithasasetofsuboptimalornon-dominatedalternativesolutionsknownasaParetooptimalset,whereacompromisesolutionmaybeselectedfromthatsetbyadecisionprocess.Differentcriteria,suchasphysicalyield,cropquality,productquality,timingoftheproductionprocess,orproductioncostsandrisks,canbeformulatedwithingreenhousecropmanagement.Thesecriteriawilloftengiverisetocontroversialclimateand肥料灌溉requirements,whichhavetobesolvedexplicitlyorimplicitlyattheso-calledtacticallevel,wherethegrowerhastomakedecisionsaboutseveralconflictingobjectives.ThesolutionofthisMOoptimizationprocess,p∈P,istheoptimaldiurnalandnocturnalpresentandfuturereferencetrajectoriesoftemperature,Xta,andelectricalconductivity,XEC,fortherestofthecropcycle.Thatis,whereisavectoroftheinsideairtemperaturealongtheoptimizationintervals,andisavectoroftheelectricalconductivity(EC)alongtheoptimizationintervals.NoticethattheplantsgrowundertheinfluenceofthePARradiation(diurnalconditions),performingthephotosynthesisprocess.Furthermore,thetemperatureinfluencesthespeedofsugarproductionbyphotosynthesis,andthusradiationandtemperaturehavetobeinbalanceinthewaythatahigherradiationlevelcorrespondstoahighertemperature.So,underdiurnalconditionsitisnecessarytomaintainthetemperatureatahighlevel.Innocturnalconditions,theplantsarenotactive(thecropdoesnotgrow),soitisnotnecessarytomaintainsuchahightemperature.Forthisreason,twotemperaturesetpointsareusuallyconsidered:diurnalandnocturnal.Itisnecessarytohighlightthatalthoughtheprocessoptimizationispresentedincontinuoustime,itissolvedindiscretetimeintervalsforanoptimizationhorizon,Nf(k)(thishorizonisvariableandrepresentstheremainingintervalsuntiltheendoftheagriculturalseason).Thus,thesolutionvectorsandareobtainedaswherekisthecurrentdiscretetimeinstant.Noticethat,fortheproposedoptimizationproblem,agreenhousecropproductionmodelisrequiredinordertoestimatetheinnerclimatebehaviorandthecropgrowththroughthedifferentstepsofthealgorithmandrelatethedifferentfunctionobjectivestothedecisionvariables.Thedynamicbehaviorofthemicroclimateinsidethegreenhouseisacombinationofphysicalprocessesinvolvingenergytransfer(radiationandheat)andmassbalance(watervaporfluxesandCO2concentration).Ontheotherhand,thecropgrowthandyieldmainlydepend,amongotherconditionssuchasirrigationandfertilizers,ontheinsidetemperatureofthegreenhouse,thePARradiation,andtheCO2concentration.Thus,bothclimateconditionsandcropgrowthinfluenceeachotherandtheirdynamicbehaviorcanbecharacterizedbydifferenttimescales.Hence,thecropgrowthinresponsetotheenvironmentcanbedescribedbytwodynamicmodels,representedbytwosystemsofdifferentialequationswithatimescaleassociatedtotheirdynamics,whichcanberepresentedbywhereXcl=Xcl(t)isann1-dimensionalvectorofgreenhouseclimatestatevariables(mainlytheinsideairtemperatureandhumidity,CO2concentration,PARradiation,soilsurfacetemperature,covertemperature,andplanttemperature),Xgr=Xgr(t)isann2-dimensionalvectorofcropgrowthstatevariables(mainlynumberofnodesonthemainstem,leafareaindex(LAI)orsurfaceofleavesbysoilarea,totaldrymatterwhichrepresentsalltheplantconstituents–root,stem,leaves,flowerandfruit–excludingwater,fruitdrymatterbeingthebiomassofthefruitsexcludingwater,andmaturefruitdrymatterormaturefruitbiomassaccumulation),U=U(t)isanm-dimensionalvectorofinputvariables(naturalventsandheatingsysteminthiswork),D=D(t)isano-dimensionalvectorofdisturbances(outsidetemperatureandhumidity,windspeedanddirection,outsideradiation,andrain),V=V(t)isaq-dimensionalvectorofsystemvariables(relatedtotranspiration,condensation,andotherprocesses),Cisanr-dimensionalvectorofsystemconstants,tisthetime,Xcl,iandXgr,iaretheknownstatesattheinitialtimeti,fcl=fcl(t)isanonlinearfunctionbasedonmassandheattransferbalances,andfgr=fgr(t)isanon-linearfunctionbasedonthebasicphysiologicalprocessesoftheplants.FortheMediterraneanarea,theauthorshavedevelopedlinearandnonlinearmodelsusingphysicallaws.Thesemodelsaretoocomplextobedetailedhere,butthemaingrowthmodelequationswillbedescribedinthefollowingsectionswheretheproblemobjectivesandthefinalMOoptimizationproblemareexplained.Theseequationswillbeusedtoshowhowthedifferentobjectives(costfunctions)areexpressedasfunctionsofthedecisionvariablesoftheoptimizationproblem(presentandfuturetemperatureandECsetpoints).2.1.MaximizationofprofitsProfitsarecalculatedasthedifferencebetweentheincomefromthesellingofthefreshfruitsandthecostsassociatedtotheirproductionwhereVpr(t)isthesellingpriceoftheproduction(estimatedfromthemarket),XFFP(t)isthefreshfruitproductionobtainedfromthecropgrowthmodelVcos(t)arethecostsincurredbyheating,electricity,fertilizers,andwater,tisthetime,tiistheinitialtimeofcropcycle,andthisthelatestharvestingtime,bothselectedbythegrower.Noticethatinpractice,thetomatocrophasmultipleharvestduringthegrowingseason.Forthatreason,threpresentsthelatestharvestingtimeinEq.Analternativeistoconsiderthenextharvestingtime(tn)inthecostfunctionandrestartingtheoptimizationprocessagainoncethepreviousharvesthasbeenproduced.Bothalternativesarevalidformultipleharvest.Theincomedependsonthepriceoftomatofruits($kg?1,€kg?1),theharvestingdates,andontheyieldinfreshweightpersurfaceunit(kgm?2).Thepricepolicyrequiresmarketmodelsorhistoricaldata,thisbeingaverydifficultpredictionproblem.Thefollowingsubsectionsdescribehowthefreshfruitproduction,XFFP(t),andtheprocesscosts,Vcos(t),canbeestimatedandrelatedwiththedecisionvariables,.2.2.MaximizationofqualityAlthoughmaximizingtheprofitscanbeunderstoodasthemainobjectivefromthegrowers’pointofview,thiscannotalwaysbeusedastheonlyone.Thegrowersusuallybelongtocooperativesoragrariansocietiesthatfacilitatetheintroductionofthehorticulturalproductsintothemarket.Theseassociationsfixthepoliciesonqualityproductsbasedonthedifferentmarketrequirements,andthusthegrowersmustadapttheirproductionprocesstothosepoliciesinordertoreachsomeminimumqualitylevels.Foodqualityembracesbothsensoryattributesthatarereadilyperceivedbythehumansensesandhiddenattributessuchashealthinessandnutrition(Shewfelt,1999).Infruitsandvegetables,thesensorypropertiesaredeterminedbytheamountofsugars,organicacids,andvolatilecompounds,aswellascolor,shape,andtexture.However,sugarsandacidsarethosereflectingoveralltastepreferencesforafruit.Foratomatocrop,solublesolidshavebeenrelatedtosugars([Lietal.,2001]and[SonneveldandvanderBurg,1991])andtitratableaciditytomainorganicacids([Auerswaldetal.,1999]and[SonneveldandvanderBurg,1991]);thustheycanbeusedasindicatorsoffruitquality.Firmnessofthefruitisanotherimportantqualityparameterinthechaingrower–dealer–consumer.Nevertheless,someworkshaveshownthatinhorticulturalvegetables,suchastomatoorflowers,someimportantparametersofsensoryqualityareinconflictwithyield([Doraisetal.,2001],[Lietal.,2001]and[SonneveldandvanderBurg,1991]).Hence,thefruitqualitycanbeexpressedas(14)whereVSSol(t)isthesolublesolidsconcentrationinthefruit,Vta(t)isthetitratableacidityinfruits,Vff(t)isthefruitfirmness,Vfs(t)isfruitsize,andwssol,wta,wff,andwfsareweightingparameters.Intomatofruits,solublesolids,titratableacidity,fruitfirmnessandsizemayberelatedtoXta(t)andXEC(t)(decisionvariables)usingthefollowinglinearapproach([Doraisetal.,2001]and[SonneveldandvanderBurg,1991])(15)Y(t)=a+b(X(t)?g(X(t)))whereY(t)isthevariabletobecalculated(solublesolids,titratableacidity,fruitfirmness,orsize),X(t)istherelateddecisionvariable(XEC(t)forVSSol(t),Vta(t),Vfs(t);andXta(t)forVff(t)),aisaconstantincrementcoefficientinY(t),bistheincrementcoefficientinY(t)perunitofincrementinX(t),andg(X(t))isapiecewisefunctionrepresentingathresholdofX(t)overwhichthereisanincrementinY(t).2.3.Maximizationofwater-useefficiencyTheexplicitinclusionofthisobjectiveintheoptimizationproblemhasanenvironment-relatedpurpose.Insemi-aridclimates,suchasMediterraneanones,waterisaveryscarceandexpensiveresource,mainlyduringsomeseasonsoftheyear.Someauthorsmaintainthattheproductivityinsuchregionsisdeterminedbytheavailablewaterandthewaterefficiencyuse(Hsiao&Xu,2000).Thus,anadequatemanagementofwaterisrequired.Withtheexplicitinclusionofthisobjective,thegrowercanselectasolutionfromtheParetofrontprovidingthedesiredwaterconsumptionduringthegrowingcycle.Thisobjectivetriestousethewaterquantitiesadequatetothecropgrowthincloserelationshiptothesuppliedconcentrationofnutrientsolution.Inthispaper,water-useefficiencyisconsideredlikethebiomassefficiencydefinedastherelationshipbetweenthefreshfruitmatterproductionandthewatersupplied.2.4.MultiobjectiveoptimizationproblemAllthevariablespresentedintheseobjectivesarefunctionsoftheairtemperature,Xta,and/ortheEC,XEC,(XFFP(t),Fsf(t),Wsw(t),Vta(t),VSSol(t),Vfs(t),Vff(t)),aswellasofmeasurabledisturbancessuchasPARradiationorCO2concentration.Thatis,theobjectivefunctionscanbeexpressedasfori=1,2,3,whereisavectoroftheinsideairtemperaturealongtheoptimizationinterval,isavectoroftheECalongtheoptimizationinterval,andΘisavectorofthemeasurabledisturbancesthathavetobepredictedalongtheoptimizationhorizon.ThesolutiontotheMOoptimizationproblemprovidesbothdiurnalandnocturnalsetpointtrajectoriesofECandinsideairtemperaturefortherestofthecontrolhorizon.Constantdiurnalandnocturnalsetpointsaredefined,andsteadystatemodelsofgreenhouseclimateandtomatocrop,summarizedinEqs.AlthoughseveraltechniqueshavebeenevaluatedtosolvetheMOoptimizationproblem(Liuetal.,2003),inthiscase,agoalattainmentalgorithmhasbeenused(sequentialquadraticprogramingSQP-based).Prioritiesforeachobjectivearedeterminedbyusingweightsthataresequentiallymodifiedineachiteration.TheconstraintsaredefinedbymaximumandminimumvaluesoftemperatureandECobtainedfromexperts’knowledgethatindicate“optimal”growingtemperaturesfortomatoandbyanalyzinglocaldatafromhistoricalseries.Theresultingconstraintsarechangingthroughouttimewithayearlypatterndesignedonthebasisofthelasttwentyyearscollecteddata.3.Multilevelhierarchicalcontrolarchitecture.Thedynamicsinvolvedinthegreenhouseproductionprocesspresentdifferenttimescalesasdescribedabove,namely,internalgreenhouseclimate,fastcropdynamics(i.e.transpiration,photosynthesis,andrespiration),andslowcropdevelopment(i.e.cropgrowthandfruitchanges).Hence,amultilayerhierarchicalcontrolarchitecturehasbeenproposedandused。3.1.CropgrowthcontrollayerTakingintoaccountthelong-termobjectives(marketprices,harvestingdates,andrequiredquality)andthelong-termpredictionsofthegrowthstateusingthemodifiedTomgromodel(Ramírez-Ariasetal.,2004)(fortheestimationofyieldandprofits),theoptimizationisperformedtocalculatethesetpointtrajectoriesoftheinsidegreenhousetemperatureandtheECalongtheconsideredcontrolhorizon(typically65daysforashortseason?260decisionvariables—or120daysforalongseason?480decisionvariables).Modelsforirrigationhavealsobeendevelopedforcontrolandoptimizationpurposes。Thelong-termweatherprediction,whichislogicallyoneoftheelementswithahigherdegreeofuncertaintyandisperformedusingasoftwaretoolthataccessestheweatherpredictionsgivenbytheSpanishNationalInstituteofMeteorologyforthenexteightdaysforward,generatespatternsbasedonseveralindexes(clarity,maximum,meanandminimumtemperatures,andsolarradiation),andsearcheswithinalocalhistoricaldatabaseforaclimaticsequencethatbetterfitsthegeneratedpatterns.Inthisway,takingtheselectedsequenceasashorttermweatherpredict

溫馨提示

  • 1. 本站所有資源如無(wú)特殊說(shuō)明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請(qǐng)下載最新的WinRAR軟件解壓。
  • 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請(qǐng)聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
  • 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁(yè)內(nèi)容里面會(huì)有圖紙預(yù)覽,若沒(méi)有圖紙預(yù)覽就沒(méi)有圖紙。
  • 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
  • 5. 人人文庫(kù)網(wǎng)僅提供信息存儲(chǔ)空間,僅對(duì)用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對(duì)用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對(duì)任何下載內(nèi)容負(fù)責(zé)。
  • 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請(qǐng)與我們聯(lián)系,我們立即糾正。
  • 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時(shí)也不承擔(dān)用戶因使用這些下載資源對(duì)自己和他人造成任何形式的傷害或損失。

最新文檔

評(píng)論

0/150

提交評(píng)論