周 昊,徐愛俊,周素茵
·農(nóng)業(yè)生物環(huán)境與能源工程·
生豬養(yǎng)殖污水水質(zhì)指標相關(guān)性分析與建模
周 昊,徐愛俊,周素茵※
(1. 浙江農(nóng)林大學(xué)信息工程學(xué)院,杭州 311300; 2. 浙江省林業(yè)智能監(jiān)測與信息技術(shù)研究重點實驗室,杭州 311300)
生豬養(yǎng)殖污水成分復(fù)雜且對環(huán)境存在較大的污染風(fēng)險,常規(guī)實驗室監(jiān)測法準確性高但效率低且時效性差,自動監(jiān)測法速度快但成本高。為尋求一種能兼顧兩種方法優(yōu)點的監(jiān)測方案,該研究以一家規(guī)模生豬養(yǎng)殖場的排放污水為研究對象,對衡量污水水質(zhì)的7個主要指標的變化特征、相關(guān)性和其中2個指標的回歸建模進行了研究。通過對不同季節(jié)及不同氣候條件下30組隨機樣本的檢測與相關(guān)性分析,發(fā)現(xiàn)氨氮、總氮和電導(dǎo)率有相似的變化趨勢且彼此之間均存在強相關(guān)性,相關(guān)系數(shù)分別為0.772、0.775和0.920。基于相關(guān)性分析結(jié)果,對氨氮和總氮分別進行了一元和多元回歸分析建模,并確定了相對最佳的適合于氨氮的“多項式回歸模型”和總氮的“綜合模型”。經(jīng)驗證,兩個模型的決定系數(shù)分別為0.855和0.953,可較好地用于評價生豬養(yǎng)殖污水中氨氮和總氮2個指標的濃度大小?;谶@2個模型,生豬養(yǎng)殖污水需直接檢測的主要指標的數(shù)量可有效減少、檢測難度和成本均明顯降低。因此,模型可為生豬養(yǎng)殖污水高效、低成本的自動監(jiān)測方案的建立提供重要的理論基礎(chǔ)。
生豬養(yǎng)殖污水;水質(zhì);相關(guān)性;回歸分析
生豬養(yǎng)殖污水成分復(fù)雜,其包含的污染物濃度較高,極易造成下游地表和地下水體的富營養(yǎng)化,對生態(tài)環(huán)境構(gòu)成極大的威脅[1-2]。多年來,生豬養(yǎng)殖污水的凈化方法、處理工藝、主要指標的去除方法等一直都是研究的熱點[3-10]。根據(jù)處理工藝的不同,生豬養(yǎng)殖污水的處理方式主要包括生態(tài)處理、工業(yè)處理和集中處理[11-13]。其中工業(yè)處理后的污水因直接排放到自然水體中而存在較大的環(huán)境污染風(fēng)險,而關(guān)于該處理模式下的污水監(jiān)測因受指標數(shù)量多、成本高等諸多因素的限制,目前主要還是采用常規(guī)的取樣監(jiān)測法[14-15]。因此,尋求一種快速、低成本的生豬污水指標監(jiān)測方法是一個十分重要的發(fā)展方向。
衡量畜禽尤其是生豬養(yǎng)殖污水水質(zhì)狀況的主要指標有氨氮(ammonia nitrogen,NH3-N)、總磷(total phosphorus,TP)、總氮(total nitrogen,TN)和化學(xué)需氧量(chemical oxygen demand,COD)等[16]。針對這些指標的檢測,除了傳統(tǒng)的實驗室化學(xué)檢測法,一些新興領(lǐng)域的測量方法和去除方法也相繼問世。陳一輝等[17]基于生物藥物的“酶法”檢測污水中NH3-N含量,該方法操作簡單,精度和靈敏度高,但所用試劑不易保存,且成本相對較高;何金成等[18-19]利用近紅外光譜法先后實現(xiàn)了廢水中COD、五日生化需氧量(five-day biochemical oxygen demand,BOD5)和酸堿度(potential of hydrogen,pH)等指標的快速測量與預(yù)測建模,能較好的反應(yīng)水體中有機物含量;Luo等[20]提出了一種基于人工濕地種植肉芽孢菌高效去除生豬污水中TP的方法;干方群等[21]發(fā)現(xiàn)了高嶺土對畜禽廢水中的TP有較好的凈化效果;基于神經(jīng)網(wǎng)絡(luò)模型的水質(zhì)預(yù)測與評價也是一種較常用的水質(zhì)研究方法[22-24]。
基于相關(guān)性分析方法,尋找水質(zhì)指標與其他因子的關(guān)系的研究也較為普遍。張賢龍等[25]研究發(fā)現(xiàn)艾比湖流域的水質(zhì)指標溶解氧(dissolved oxygen,DO)和COD可通過二維熒光峰值進行快速估算;徐利等[26]得出張家口市清水河中葉綠素a與TP呈極顯著正相關(guān);Copetti等[27]通過實驗室分析法和現(xiàn)場連續(xù)測量相結(jié)合,對意大利米蘭的Seveso河流中微量金屬與濁度、總可溶性固形物等分析,發(fā)現(xiàn)二者相關(guān)程度較高。張苒等[28]通過對廣東省主要流域的2個水質(zhì)自動監(jiān)測站連續(xù)3年的自動監(jiān)測數(shù)據(jù)的分析,發(fā)現(xiàn)pH、濁度、電導(dǎo)率(electrical conductivity,EC)和DO之間均有不同程度的相關(guān),同時EC與TP、COD和NH3-N均具有較強的相關(guān)性,并基于該特點成功實現(xiàn)了污水的預(yù)警監(jiān)測。
綜上所述,國內(nèi)外不乏對水質(zhì)指標的檢測、相關(guān)性分析等研究,但針對生豬養(yǎng)殖污水水質(zhì)指標間關(guān)系的分析等卻鮮見報道。本研究以一家污水處理工藝成熟、豬只生產(chǎn)均衡的規(guī)模生豬養(yǎng)殖場的排放污水為研究對象,對污水中的7個指標(NH3-N、TP、TN、COD、pH、EC和DO)進行了特征分析、相關(guān)性分析和部分指標的回歸建模,分別得到了NH3-N和TN的回歸模型。以這些模型為理論依據(jù),旨在探尋一種高效、低成本的生豬養(yǎng)殖污水水質(zhì)自動監(jiān)測方案。
選擇浙江省北部地區(qū)一家規(guī)模生豬養(yǎng)殖場作為采樣點。該養(yǎng)殖場占地33.35 hm2,各類豬舍建筑面積約30 000 m2,每年生豬出欄量約20 000頭。養(yǎng)殖場的污水處理工程采用較為先進的“養(yǎng)殖廢水低C/N比厭氧沼液高效脫氮除碳處理”工藝。
1.2.1 樣本采集
為提高研究結(jié)論的準確性與可靠性,分別在不同的季節(jié),以及陰雨天、晴天、雪天等不同氣候、不同天氣情況下進行隨機采樣,采集目標為各道污水處理工序中的不同水樣。其中2018年10月和11月,2019年3月和5月各采樣4次;2019年1月、2月和7月各采樣3次;2018年12月、2019年4月和6月各采樣5次,共計樣本數(shù)量40組。將每次采集的樣本中取一組作為檢驗樣本,遴選出建模樣本30組,檢驗樣本10組。
1.2.2 指標檢測及方法
本研究的檢測指標為電導(dǎo)率(electrical conductivity,EC)、酸堿度(potential of hydrogen,pH)和溶解氧(dissolved oxygen,DO),以及《畜禽養(yǎng)殖業(yè)污染物排放標準》中規(guī)定的氨氮(ammonia nitrogen,NH3-N)、總磷(total phosphorus,TP)、總氮(total nitrogen,TN)和化學(xué)需氧量(chemical oxygen demand,COD)。其中,EC、pH和DO等均于現(xiàn)場檢測,之后使用相應(yīng)容器對水樣進行密封,立即運送到實驗室對NH3-N、TP、TN和COD 4個水質(zhì)指標進行實驗室化學(xué)法檢測。
1)指標檢測方法
各指標的檢測方法及儀器如表1所示。
表1 水質(zhì)指標檢測方法及儀器
2)數(shù)據(jù)分析方法
采用SPSS 24.0軟件對污水中各水質(zhì)指標進行Pearson相關(guān)性分析、相關(guān)系數(shù)計算以及數(shù)據(jù)建模;采用Origin Pro 9.0軟件進行圖表繪制。
將檢測后的水質(zhì)指標數(shù)據(jù)輸入SPSS軟件進行分析,得到各自的統(tǒng)計特征如表2所示。由于每組水樣中溶解氧(dissolved oxygen,DO)的大小基本無變化,故將其忽略。
表2 水質(zhì)指標的統(tǒng)計特征
由表2可以看出,氨氮(ammonia nitrogen,NH3-N)、電導(dǎo)率(electrical conductivity,EC)和總氮(total nitrogen,TN)3個指標的變異系數(shù)相對較高,均超過了40%,濃度大小波動較大;而總磷(total phosphorus,TP)、化學(xué)需氧量(chemical oxygen demand,COD)和酸堿度(potential of hydrogen,pH)的變異系數(shù)僅為24%、20%和6%,數(shù)值變化幅度相對較小。
各水質(zhì)指標濃度或數(shù)值變化情況如圖1所示,在不同工序、不同采樣時間以及不同季節(jié)、氣溫等條件下,僅NH3-N、TN和EC的變化趨勢相似,除此之外,其余各指標濃度或大小變化均無明顯規(guī)律可循。
圖1 水質(zhì)指標濃度或數(shù)值變化
綜合表2和圖1可以看出,NH3-N、TN和EC 3個指標的大小總體均呈緩慢下降趨勢,因此推測三者之間可能存在某種聯(lián)系,但直接從圖表中得出具體的數(shù)值關(guān)系,難度系數(shù)較大。
首先對研究的6個水質(zhì)指標進行分組?;诟髦笜说臏y量方法的不同以及所用傳感器檢測成本的高低,將NH3-N、COD、TP和TN 4個在實驗室進行化學(xué)法檢測的指標歸納入組Ⅰ;將pH和EC 2種由儀器直接檢測的指標歸納入組Ⅱ。組Ⅰ內(nèi)各指標檢測成本及復(fù)雜程度總體來說均高于組Ⅱ內(nèi)各指標,對組Ⅰ和組Ⅱ中各指標進行Pearson相關(guān)性分析,分析結(jié)果如表3所示。
表3 各水質(zhì)指標間的相關(guān)系數(shù)
注:*表示<0.05,**表示<0.01。
Note: * Indicates<0.05, ** indicates<0.01.
表3中的相關(guān)性分析結(jié)果表明,在不同條件下采集的水樣樣本中:NH3-N與TN、NH3-N與pH、NH3-N與EC、EC與TN這4組數(shù)據(jù)均存在極顯著相關(guān)關(guān)系(<0.01)。其中TN與EC之間存在極強的顯著正相關(guān)關(guān)系,Pearson相關(guān)系數(shù)高達0.920;NH3-N分別與TN和EC 2個指標存在較強的顯著正相關(guān)關(guān)系(<0.05),Pearson相關(guān)系數(shù)分別為0.772和0.775;此結(jié)果與2.1小節(jié)中NH3-N、EC和TN 3個指標變化規(guī)律大致相似的情況相吻合。
除此之外,NH3-N和pH間雖存在極顯著相關(guān)關(guān)系,但相關(guān)系數(shù)只有0.564;NH3-N和TP、TP和TN、TP和EC之間雖有顯著的相關(guān)關(guān)系,但相關(guān)系數(shù)也都較低,均小于0.50;其余水質(zhì)指標之間則既不存在顯著相關(guān)關(guān)系,Pearson相關(guān)系數(shù)也不高,對本研究的后續(xù)建模沒有參考價值。
2.3.1 模型構(gòu)建
由2.2中的分析可知,NH3-N、TN和EC 3個水質(zhì)指標兩兩之間均存在較高的顯著相關(guān)性?;谠撓嚓P(guān)性,本研究構(gòu)建了三者之間的具體回歸模型。
首先建立3個一元回歸模型:“NH3-N和EC之間的回歸模型”、“TN和EC之間的回歸模型”、“NH3-N和TN之間的回歸模型”。在“NH3-N和EC之間的回歸模型”及“TN和EC之間的回歸模型”中,因EC檢測方便且精度較高,故將其設(shè)為自變量,TN和NH3-N作為組Ⅰ內(nèi)被替代檢測的指標,分別設(shè)作因變量;而在NH3-N和TN之間的回歸模型中,則將NH3-N設(shè)為因變量,TN設(shè)為自變量。表4、表5和表6中分別列出了三者之間不同方法下的擬合效果。
表4 NH3-N和EC擬合效果比較
注:RSS為殘差平方和,MSR為均方回歸,2為決定系數(shù),下同。
Note: RSS is Residual sum of squares, MSR is Mean square regression,2is coefficient of determination, the same below.
由表4可知,NH3-N和EC的3種回歸模型中,線性式的模型擬合效果遠低于另外2種模型,二次式與多項式雖然有相似的RSS和決定系數(shù)2,但多項式模型的MSR僅為90.08,明顯小于二次式模型,因此多項式是NH3-N和EC的最佳回歸模型,其擬合效果如圖2a所示。
表5 TN和EC擬合效果比較
TN和EC之間的回歸模型經(jīng)過篩選分別選用了線性式、冪次式以及多項式。由表5可知,無論是RSS還是MSR,冪次式較其他2種模型都具有極大優(yōu)勢,且冪次式具有較大的決定系數(shù)2,因此冪次式是TN和EC的最佳回歸模型,其擬合效果如圖2b所示。
根據(jù)NH3-N和TN之間的數(shù)據(jù)關(guān)系,也分別選用了線性式、二次式及多項式3種回歸模型。表6的3種模型中,多項式的RSS為56.91,MSR為86.75,相對另外2種模型較小,決定系數(shù)2為0.82,遠高于線性式和二次式。因此NH3-N和TN使用多項式建立模型時,其擬合度最優(yōu),擬合效果如圖2c所示。
表6 NH3-N和TN擬合效果比較
圖2 擬合效果
由圖2b可知,TN和EC的擬合線性度最高,同時圖2c中的NH3-N和TN也有較好的擬合效果,故本研究嘗試將圖2b和圖2c中的兩個模型組合,以EC為自變量、以TN為中間變量、以NH3-N為因變量,建立式(1)中的組合回歸模型,其擬合效果如圖2d所示。
綜上分析,本研究針對NH3-N為因變量而建立的模型為多項式回歸模型以及組合后的回歸模型,分別將其定義為模型Ⅰ和模型Ⅱ;針對TN而建立的模型為冪次式回歸模型,將其定義為模型Ⅲ。
2.3.2 模型驗證
為了檢驗回歸模型的準確性,調(diào)取準備好的10組檢驗樣本數(shù)據(jù),對上文得到的模型Ⅰ、模型Ⅱ、模型Ⅲ進行模型精度檢驗。根據(jù)各模型自變量的要求,將對應(yīng)的檢驗樣本水質(zhì)指標實測值代入到各回歸模型中,得到各回歸模型中因變量的估算值,根據(jù)得到的估算值與相應(yīng)的實測值建立線性擬合圖,通過樣點距離1:1標準參考線的離散程度、擬合線與1:1標準參考線的偏離程度以及2值的大小來衡量回歸模型的準確性[29]。
1)針對NH3-N的模型驗證
模型Ⅰ和模型Ⅱ的NH3-N估算值與實測值線性擬合效果分別如圖3a和圖3b所示。從NH3-N各樣點的分布來看,模型Ⅰ所對應(yīng)的擬合圖中樣點分布的密集程度高于模型Ⅱ;從線性擬合線來看,模型Ⅰ所對應(yīng)的擬合圖中,樣點的線性擬合線與1:1標準參考線的偏離程度更小;從決定系數(shù)2來看,模型Ⅰ對應(yīng)擬合圖中,NH3-N估算值與實測值的決定系數(shù)2為0.855,而模型Ⅱ?qū)?yīng)擬合圖中的決定系數(shù)2僅為0.803。因此從各方面來看,模型Ⅰ的擬合精度高于模型Ⅱ。
2)針對TN的模型驗證
模型Ⅲ中,TN的估算值與實測值線性擬合效果如圖3c所示。TN樣點距1:1標準參考線分布得很近且較為集中,樣點的線性擬合線與1:1標準參考線的偏差較小,TN的估算值與實測值的決定系數(shù)2為0.948,故模型Ⅲ的準確性較高。
圖3 模型驗證
2.3.3 綜合建模
在上述模型建立和驗證過程中,雖已得出針對TN的較高精度回歸模型,但由于缺乏對比參照模型,極易導(dǎo)致結(jié)論的片面性,同時NH3-N的回歸模型精度也有待提高。為此,本研究對NH3-N、TN和EC 3個水質(zhì)指標進行綜合建模:
1)以NH3-N為因變量,EC和TN分別為自變量1、2建立式(2)中的綜合模型
2)以TN為因變量,NH3-N和EC分別為自變量1、2建立式(3)中的綜合建模
同樣,本研究將針對NH3-N建立的綜合模型定義為模型Ⅳ,針對TN建立的綜合模型定義為模型Ⅴ。
圖4 綜合模型驗證
2.3.4 綜合模型驗證
為驗證綜合模型的準確性,本研究再次進行估算值與實測值的相關(guān)性分析,以驗證回歸方程的擬合程度。同樣調(diào)取10組檢驗樣本數(shù)據(jù),將指標真實值與估算值建立線性擬合圖,通過樣點距離1:1標準參考線的離散程度、擬合線與1:1標準參考線的偏離程度以及2值的大小來衡量回歸模型的準確性。
1)針對NH3-N的綜合模型驗證
觀察模型Ⅳ對應(yīng)的NH3-N估算值與實測值線性擬合效果(圖4a)可知,模型Ⅳ樣點分布較散,距離1:1標準參考線較遠,樣點的線性擬合線與1:1標準參考線的偏離程度大,決定系數(shù)2僅為0.607,故模型Ⅳ的擬合精度遠不如模型Ⅰ。
2)針對TN的綜合模型驗證
將模型Ⅴ與模型Ⅲ進行估算值與實測值線性擬合效果的對比(圖3c和圖4b)后發(fā)現(xiàn),模型Ⅴ對應(yīng)的TN樣點分布較模型Ⅲ更為集中,其樣點線性擬合線與1:1標準參考線明顯有著更小的距離偏差,且模型Ⅴ所對應(yīng)的估算值與實測值擬合圖中決定系數(shù)2值更高(0.953)。因此,模型Ⅴ的精確度更高,能更準確地通過NH3-N和EC來描述養(yǎng)殖污水中的TN含量。
評估生豬養(yǎng)殖污水排放是否達標的關(guān)鍵依賴于各水質(zhì)指標的監(jiān)測結(jié)果的大小。但生豬養(yǎng)殖污水成分構(gòu)成極其復(fù)雜,除了本研究中的主要指標外,還含有銅、鐵、鋅、錳等多種微量元素、抗生素、懸浮物等污染物。本研究僅僅是根據(jù)《畜禽養(yǎng)殖業(yè)污染物排放標準》的規(guī)定,對污水中的幾個主要指標氨氮(ammonia nitrogen,NH3-N)、總磷(total phosphorus,TP)、總氮(total nitrogen,TN)、化學(xué)需氧量(chemical oxygen demand,COD)、酸堿度(potential of hydrogen,pH)和電導(dǎo)率(electrical conductivity,EC)等的濃度或數(shù)值進行了檢測、相關(guān)性分析與部分指標的建模。在本研究中的污水指標檢測環(huán)節(jié),為保障數(shù)據(jù)測量的精度和模型的準確性,pH和EC的現(xiàn)場采集分別使用的是具有溫度自動補償功能和較高精度(±0.01 pH和±0.5% F.S)的便攜式儀器,溶解氧(dissolved oxygen,DO)使用的具有鹽度補償功能、精度為±0.01 mg/L的熒光溶氧儀;4個主要指標(NH3-N、TN、TP和COD)的檢測采用的仍是傳統(tǒng)的實驗室化學(xué)檢測法,該方法雖然存在效率低、時效性差、人工成本高等缺點,但數(shù)據(jù)的準確性和可靠性可以得到保障[30]。
基于以上分析,本研究中構(gòu)建的NH3-N和TN的回歸模型在準確性和可靠性上均有較好的基礎(chǔ)數(shù)據(jù)保障。但模型的精度是否可以進一步提高、是否與微量元素等其他污染物之間也存在相關(guān)性、是否存在多因素之間的協(xié)同效應(yīng)與時序影響將是后期值得探討的問題。
高效低成本監(jiān)測意味著基于少于生豬養(yǎng)殖污水常規(guī)檢測指標數(shù)量的傳感設(shè)備及在線檢測設(shè)備便可達到全面監(jiān)測污水指標的目的,同時能兼顧常規(guī)檢測方法的準確性、可靠性和自動檢測方法的便捷性、時效性。關(guān)于2種方法的監(jiān)測效果對比及是否存在較好的一致性,劉京等[30]對15個國家地表水水質(zhì)自動監(jiān)測站中的COD、NH3-N、TP、pH和DO共5個指標進行了連續(xù)3個月的站房外常規(guī)監(jiān)測、站房內(nèi)常規(guī)監(jiān)測和自動監(jiān)測,共獲得6219組有效監(jiān)測數(shù)據(jù),分析發(fā)現(xiàn)常規(guī)監(jiān)測法得到的兩組數(shù)據(jù)基本一致,同時與自動監(jiān)測數(shù)據(jù)有相同的變化趨勢,但一致程度均有所下降。該結(jié)果表明,可以基于自動監(jiān)測數(shù)據(jù)對各水質(zhì)指標建立相應(yīng)的回歸模型,從而實現(xiàn)準確、快速的水質(zhì)指標的監(jiān)測。因此,基于本研究中的模型Ⅰ和模型Ⅴ,構(gòu)建生豬養(yǎng)殖污水水質(zhì)高效、低成本的自動監(jiān)測方案是可行的。方案中涉及的指標包括EC、pH、NH3-N、TP、TN和COD,其中在線檢測難度大、成本高[31]的TN無需直接檢測,其濃度可根據(jù)模型Ⅴ由NH3-N、EC的值計算得出,而檢測難度和成本相對較低的NH3-N濃度可根據(jù)模型I由EC大小得到;EC和pH的檢測采用與表1中相應(yīng)儀器同類型電極,檢測簡單且成本較低;而TP和COD數(shù)據(jù)通過基于自動監(jiān)測數(shù)據(jù)的回歸模型獲得。與傳統(tǒng)監(jiān)測方法相比,該方案中需監(jiān)測的指標數(shù)量明顯減少,使得監(jiān)測總體難度下降、監(jiān)測成本降低、監(jiān)測效率提高。建立生豬污水部分重要監(jiān)測指標的回歸模型,構(gòu)建基于模型并兼顧兩種監(jiān)測方法優(yōu)點的高效、低成本自動監(jiān)測方案是本研究的主要研究目的和創(chuàng)新點。此處只是從監(jiān)測指標的角度進行了探討,關(guān)于監(jiān)測方案的具體實現(xiàn)將是下一步研究的內(nèi)容。
在研究過程中不乏待改善之處。由于實驗條件有限,本研究最終采樣40組,樣本容量偏小,易引發(fā)由于測量值異常而導(dǎo)致模型精度降低的情況,而足夠大的樣本容量有助于研究過程中及時排除異常值,提高模型的精度。另外,本研究在模型建立環(huán)節(jié)選擇了相對簡單的建模方法,在此基礎(chǔ)上進行模型對比分析、模型驗證、綜合建模和綜合模型驗證,滿足了用簡單易測指標推導(dǎo)高難度、高成本測量指標的基本要求。而更精確的回歸模型不僅可以達到替代檢測的基本目的,也可對各替代指標進行精準的估算,對養(yǎng)殖污水,乃至各類工業(yè)污水的監(jiān)測都具有十分重要的意義,也是本研究亟需不斷努力且不斷改進的方向。
通過數(shù)據(jù)分析,確定了生豬養(yǎng)殖污水中NH3-N、TN和EC 3個指標之間均存在極顯著的正相關(guān)關(guān)系。基于這3個指標的強相關(guān)性,首先構(gòu)建出NH3-N、TN和EC兩兩之間的回歸模型,通過比較各模型的決定系數(shù)2、殘差平方RSS和均方回歸MSR篩選出擬合效果最佳的回歸模型(模型Ⅰ和模型Ⅲ);在此基礎(chǔ)上,將模型Ⅰ與模型Ⅲ組合并建立出新的NH3-N回歸模型(模型Ⅱ);隨后研究又對3個指標進行綜合建模,分別得出了NH3-N綜合模型(模型Ⅳ)和TN綜合模型(模型Ⅴ)。至此本研究共建立模型Ⅰ-Ⅴ 5個回歸模型。在模型驗證環(huán)節(jié),利用10組檢驗樣本數(shù)據(jù),對模型的擬合效果進行模型精度驗證,并最終篩選出能較好地反映污水中NH3-N和TN濃度的最佳回歸模型:模型Ⅰ和模型Ⅴ?;谶@2個模型,對生豬養(yǎng)殖污水高效、低成本的自動監(jiān)測方案進行了可行性分析。
[1] 吳建敏,徐俊,翟云忠,等. 畜禽規(guī)模養(yǎng)殖廢水污染因子監(jiān)測評價分析[J]. 家畜生態(tài)學(xué)報,2009,30(4):48-51.
Wu Jianmin, Xu Jun, Zhai Yunzhong, et al. Monitoring and evaluation of wastewater pollution factors in livestock and poultry scale cultivation[J]. Journal of Domestic Animal Ecology, 2009, 30(4): 48-51. (in Chinese with English abstract)
[2] 郭瑞華,靳紅梅,吳華山,等. 規(guī)模豬場污水多級處理系統(tǒng)中重金屬總量及其形態(tài)變化特征[J]. 農(nóng)業(yè)工程學(xué)報,2018,34(6):210-216.
Guo Ruihua, Jin Hongmei, Wu Huashan, et al. Total content of heavy metals and their chemical form changes in multilevel wastewater treatment system in intensive swine farm[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(6): 210-216. (in Chinese with English abstract)
[3] 魯為華,靳省飛,王樹林,等. 伊犁絹蒿種子黏液提取工藝及對畜禽糞便污水的凈化效果[J]. 農(nóng)業(yè)工程學(xué)報,2018,34(21):245-252.
Lu Weihua, Jin Shengfei, Wang Shulin, et al. Extraction process of mucilaginous from seeds of seriphidium transiliense and its purification effect on livestock manure sewage[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(21): 245-252. (in Chinese with English abstract)
[4] Inawati O, Aznah N A, Zaini U, et al. Livestock wastewater treatment using aerobic granular sludge[J]. Bioresource Technology, 2013, 133: 630-634.
[5] Bong-yul T, Bong-sik T, Young-ju K, et al. Optimization of color and COD removal from livestock wastewater by electrocoagulation process: Application of Box–Behnken design (BBD)[J]. Journal of Industrial and Engineering Chemistry, 2015, 28: 307-315.
[6] Waki M, Yasuda T, Fukumoto Y, et al. Treatment of swine wastewater in continuous activated sludge systems under different dissolved oxygen conditions: Reactor operation and evaluation using modelling[J]. Bioresource Technology, 2018, 250: 574-582.
[7] 張彩瑩,王巖,王妍艷. 潛流人工濕地對畜禽養(yǎng)殖廢水的凈化效果[J]. 農(nóng)業(yè)工程學(xué)報,2013,29(17):160-168.
Zhang Caiying, Wang Yan, Wang Yanyan. Purification effect of subsurface flow constructed wetland on livestock wastewater[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2013, 29(17): 160-168. (in Chinese with English abstract)
[8] Robert L K, Victor W E, Robert E B, et al. Constructed wetlands for livestock wastewater management[J]. Ecological Engineering, 2000, 15: 41-55.
[9] Hu Y S, Kumar J L G, Akintunde A O, et al. Effects of livestock wastewater variety and disinfectants on the performance of constructed wetlands in organic matters and nitrogen removal[J]. Environmental Science & Pollution Research, 2011, 18(8): 1414-1421.
[10] 張穎,鄧良偉. 豬場廢水厭氧消化過程中的除磷效果[J]. 生態(tài)與農(nóng)村環(huán)境學(xué)報,2012,28(1):93-97.
Zhang Ying, Deng Liangwei. Phosphorus removal from swine wastewater through anaerobic digestion[J]. Journal of Ecology & Rural Environment, 2012, 28(1): 93-97. (in Chinese with English abstract)
[11] Bilotta Patrícia, Steinmetz R L R, Kunz A, et al. Swine effluent post-treatment by alkaline control and UV radiation combined for water reuse[J]. Journal of Cleaner Production, 2017, 140: 1247-1254.
[12] 于濤,成水平,賀鋒,等. 基于復(fù)合垂直流人工濕地的循環(huán)水養(yǎng)殖系統(tǒng)凈化養(yǎng)殖效能與參數(shù)優(yōu)化[J]. 農(nóng)業(yè)工程學(xué)報,2008,24(2):188-193.
Yu Tao, Cheng Shuiping, He Feng, et al. Performance and optimization of recirculating aquaculture system combined with integrated vertical-flow constructed wetland[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2008, 24(2): 188-193. (in Chinese with English abstract)
[13] 張世羊,常軍軍,高毛林,等. 曝氣對垂直流濕地處理水產(chǎn)養(yǎng)殖廢水脫氮的影響[J]. 農(nóng)業(yè)工程學(xué)報,2015,31(9):235-241.
Zhang Shiyang, Chang Junjun, Gao Maolin, et al. Impact of artificial aeration on nitrogen removal from aquaculture wastewater treated by vertical-flow constructed wetland[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2015, 31(9): 235-241. (in Chinese with English abstract)
[14] 石奧,張建淼,王重慶,等. 北京市規(guī)?;笄蒺B(yǎng)殖場糞便和污水監(jiān)測及承載力估算[J]. 家畜生態(tài)學(xué)報,2018,39(4):63-67,96.
Shi Ao, Zhang Jianmiao, Wang Chongqing, et al. Monitoring and carrying capacity estimation of animal manure and wastewater in Beijing livestock farm[J]. Journal of Domestic Animal Ecology, 2018, 39(4): 63-67, 96. (in Chinese with English abstract)
[15] 杜衛(wèi)莉. 我國南方養(yǎng)豬業(yè)污染現(xiàn)狀監(jiān)測分析[J]. 廣東化工,2016,43(3):86-88,83.
Du Weili. Analysis of China’s southern hoggery pollution monitoring[J]. Guangdong Chemical Industry, 2016, 43(3): 86-88, 83. (in Chinese with English abstract)
[16] 張燕,劉雪蘭,王月明,等. 中國規(guī)?;笄蒺B(yǎng)殖污水處理中人工濕地的研究進展[J]. 環(huán)境科學(xué)與技術(shù),2016,39(1):87-92.
Zhang Yan,Liu Xuelan,Wang Yueming,et al. Research progress of constructed wetland treating intensive livestock and poultry wastewater in China[J]. Environmental Science & Technology, 2016, 39(1): 87-92. (in Chinese with English abstract)
[17] 陳一輝,李偉民,伍培,等. 氨氮測定方法的對比研究[J]. 環(huán)境工程,2011,29(S1):234-236.
Chen Yihui, Li Weimin, Wu Pei, et al. Comparative study on determination methods of ammonia nitrogen[J]. Environmental Engineering, 2011, 29(S1): 234-236. (in Chinese with English abstract)
[18] 何金成,楊祥龍,王立人,等. 近紅外光譜法測定廢水化學(xué)需氧量[J]. 浙江大學(xué)學(xué)報:工學(xué)版,2007,41(5):752-755,789.
He Jincheng, Yang Xianglong, Wang Liren, et al. Determination of chemical oxygen demand in wastewater by near-infrared spectroscopy[J]. Journal of Zhejiang University: Engineering Science, 2007, 41(5): 752-755, 789. (in Chinese with English abstract)
[19] 何金成,楊祥龍,王立人,等. 基于近紅外光譜法的廢水COD、BOD5、pH的快速測量[J]. 環(huán)境科學(xué)學(xué)報,2007,21(12):2105-2108.
He Jincheng, Yang Xianglong, Wang Liren, et al. Rapid determination of chemical oxygen demand (COD), biochemical oxygen demand (BOD5) and pH in wastewater using near-infrared spectroscopy[J]. Acta Scientiae Circumstantiae, 2007, 21(12): 2105-2108. (in Chinese with English abstract)
[20] Luo Pei, Liu Feng, Liu Xinliang, et al. Phosphorus removal from lagoon-pretreated swine wastewater by pilot-scale surface flow constructed wetlands planted with Myriophyllum aquaticum[J]. Science of the Total Environment, 2017, 576: 490-497.
[21] 干方群,徐子昊,楊一帆,等. 高嶺土對畜禽廢水中磷的凈化效果及其費效分析[J]. 生態(tài)與農(nóng)村環(huán)境學(xué)報,2019,35(6):795-800.
Gan Fangqun, Xu Zihao, Yang Yifan, et al. Phosphate removal efficiency of kaolin in livestock and poultry wastewater and its cost-benefit analysis[J]. Journal of Ecology and Rural Environment, 2019, 35(6): 795-800. (in Chinese with English abstract)
[22] 方駿,戴連奎. 基于混合神經(jīng)網(wǎng)絡(luò)模型的污水COD值預(yù)估法[J]. 中國給水排水,2003,19(12):6-10.
Fang Jun, Dai Liankui. Sewage COD value estimation method based on hybrid neural network model[J]. China Water & Wastewater, 2003, 19(12): 6-10. (in Chinese with English abstract)
[23] 王娟,張飛,王小平,等. 平行因子法結(jié)合自組織映射神經(jīng)網(wǎng)絡(luò)的三維熒光特征及其與水質(zhì)的關(guān)系[J]. 光學(xué)學(xué)報,2017,37(7):357-367.
Wang Juan, Zhang Fei, Wang Xiaoping, et al. Three-dimensional fluorescence characteristics of parallel factor method combined with self-organizing map neural network and its relationship with water quality[J]. Acta Optica Sinica, 2017, 37(7): 357-367. (in Chinese with English abstract)
[24] 李宣諭. 基于人工智能對地表水的水質(zhì)預(yù)測與評價研究[D]. 吉林:東北電力大學(xué),2017.
Li Xuanyu. Research on Prediction and Evaluation of Surface Water Quality Based on Artificial Intelligence[D]. Jilin: Northeast Electric Power University, 2017. (in Chinese with English abstract)
[25] 張賢龍,張飛,張海威,等. 艾比湖流域地表水二維熒光峰值與水質(zhì)指標關(guān)系初探[J]. 光譜學(xué)與光譜分析,2018,38(2):481-487.
Zhang Xianlong, Zhang Fei, Zhang Haiwei, et al. The preliminary study on the relationship between two dimensional fluorescence peak value of surface water and water quality indexes in ebinur lake basin[J]. Spectroscopy and Spectral Analysis, 2018, 38(2): 481-487. (in Chinese with English abstract)
[26] 徐利,郝桂珍,甄玉飛,等. 北方人為改造河流葉綠素a和環(huán)境因子時空分布特征及其相關(guān)性分析[J]. 科學(xué)技術(shù)與工程,2018,18(26):147-152.
Xu Li, Hao Guizhen, Zhen Yufei, et al. Spatial and temporal distribution characteristics and correlation of chlorophyll-a and environmental factors in northern China[J]. Science Technology and Engineering, 2018, 18(26): 147-152. (in Chinese with English abstract)
[27] Copetti D, Marziali L, Viviano G, et al. Intensive monitoring of conventional and surrogate quality parameters in a highly urbanized river affected by multiple combined sewer overflows[J]. Water Supply, 2019, 19(3): 953-966.
[28] 張苒,劉京,周偉,等. 水質(zhì)自動監(jiān)測參數(shù)的相關(guān)性分析及在水環(huán)境監(jiān)測中的應(yīng)用[J]. 中國環(huán)境監(jiān)測,2015,31(4):125-129.
Zhang Ran, Liu Jing, Zhou Wei, et al. Study on the correlation of water quality automatic monitoring parameters and its application in aquatic environment monitoring[J]. Environmental Monitoring in China, 2015, 31(4): 125-129. (in Chinese with English abstract)
[29] 馬馳. 基于Sentinel-1雙極化雷達影像的土壤含鹽量反演[J]. 農(nóng)業(yè)工程學(xué)報,2018,34(2):153-158.
Ma Chi. Quantitative retrieval of soil salt content based on Sentinel-1 dual polarization radar image[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2018, 34(2): 153-158. (in Chinese with English abstract)
[30] 劉京,魏文龍,李曉明,等. 水質(zhì)自動監(jiān)測與常規(guī)監(jiān)測結(jié)果對比分析[J]. 中國環(huán)境監(jiān)測,2017,33(5):159-166.
Liu Jing, Wei Wenlong, Li Xiaoming, et al. Comparative analysis of state surface water quality by automatic monitoring and laboratory monitoring[J]. Environmental Monitoring in China, 2017, 33(5): 159-166. (in Chinese with English abstract)
[31] 李文,呂赫,徐明剛,等. 多量程原位水質(zhì)總磷總氮一體式在線監(jiān)測儀[J]. 發(fā)光學(xué)報,2019,40(7):930-940.
Li Wen, Lü He, Xu Minggang, et al. Multi-range in-situ water quality total phosphorus total nitrogen integrated online monitor[J]. Chinese Journal of Luminescence, 2019, 40(7): 930-940. (in Chinese with English abstract)
Correlation analysis and modeling of water quality indexes for swine breeding wastewater
Zhou Hao, Xu Aijun, Zhou Suyin※
(1.,,311300,; 2.,311300,)
According to the difference of treatment process about swine breeding sewage, the treatment methods are divided into ecological treatment, industrial treatment and centralized treatment. The components of sewage treated by industrial treatment are extremely complex, there will be a great risk of environmental pollution if the sewage is directly discharged into the natural water body. It’s very important to monitor sewage quality. The monitoring methods commonly used in swine breeding sewage mainly include laboratory monitoring and automatic monitoring. The laboratory monitoring is traditional, which has the advantage of high data accuracy and the disadvantages of low efficiency and poor timeliness, the sewage indexes can be detected fast but costly using automatic monitoring method. To find a monitoring scheme that combined the advantages of laboratory monitoring method and automatic monitoring method, took the sewage from a large-scale pig farm as the research object, the change characteristics, correlation of seven main indexes of sewage quality and regression modeling of two main indexes were studied. The seven indeices were respectively ammonia nitrogen, total phosphorus, total nitrogen, chemical oxygen demand, the potential of hydrogen, dissolved oxygen and electrical conductivity. Through the detection and correlation analysis of 30 random samples from different seasons and climatic conditions, it was found that ammonia nitrogen, total nitrogen and electrical conductivity had similar variation trends and strong correlation each other, the correlation coefficient of ammonia nitrogen and total nitrogen was 0.772, and that of ammonia nitrogen and electrical conductivity was 0.775, the correlation coefficient of total nitrogen and electrical conductivity was 0.920. Based on the results of correlation analysis, many types of monadic regressive and multivariate regression models for ammonia nitrogen and total nitrogen were established respectively, the relatively optimal “polynomial regression model” (model I) for ammonia nitrogen and the “comprehensive model” (model V) for total nitrogen were determined by comparing the coefficient of determination, residual sum of squares and the mean square regression of each model. The verification results based on 10 sets of data showed that the estimated values of these two models were closest to the measured values, the coefficients of determination of model I and model V were 0.855 and 0.953 respectively. Therefore, these two models could be used to evaluate the concentration of ammonia nitrogen and total nitrogen in swine breeding sewage. The existing studies shown that the data obtained by laboratory monitoring and automatic monitoring had the same change law although the value was different, which meant that there was a good linear relationship between them, hence a linear regression model based on the automatic monitoring data could be established to achieve the monitoring of water quality indexes accurately and rapidly. Based on this conclusion and the above two models, the feasibility of an efficient and low-cost automatic monitoring scheme for swine breeding wastewater quality was analyzed in this study. The indexes involved in the solution included electrical conductivity, the potential of hydrogen, ammonia nitrogen, total phosphorus, total nitrogen, and chemical oxygen demand, the total nitrogen that was difficult and expensive to detect automatically does not require to detect directly, the concentration of which could be calculated by the value of ammonia nitrogen and electrical conductivity according to model V, the concentration of ammonia nitrogen with relatively low difficulty and cost could be obtained by the value of electrical conductivity according to model I, the detection of electrical conductivity and potential of hydrogen was more convenient and the cost was lower, the data of total phosphorus and chemical oxygen demand would be obtained by linear regression model based on automatic monitoring data. Compared with the existing monitoring methods, the number of indexes that needed to be detected directly in this scheme would be significantly reduced, which would make the overall difficulty and the cost of monitoring decreasing, and the monitoring efficiency improved. Consequently, these two models could provide an important theoretical basis for the establishment of an efficient and low-cost automatic monitoring scheme for swine breeding sewage.
swine breeding wastewater; water quality; correlation; regression analysis
周 昊,徐愛俊,周素茵. 生豬養(yǎng)殖污水水質(zhì)指標相關(guān)性分析與建模[J]. 農(nóng)業(yè)工程學(xué)報,2020,36(1):200-207.doi:10.11975/j.issn.1002-6819.2020.01.023 http://www.tcsae.org
Zhou Hao, Xu Aijun, Zhou Suyin. Correlation analysis and modeling of water quality indexes for swine breeding wastewater[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2020, 36(1): 200-207. (in Chinese with English abstract) doi:10.11975/j.issn.1002-6819.2020.01.023 http://www.tcsae.org
2019-10-11
2019-12-27
浙江省公益技術(shù)應(yīng)用研究計劃項目(LGN19F010001)
周 昊,從事污水監(jiān)測等物聯(lián)網(wǎng)方向的研究。Email:1220470928@qq.com
周素茵,講師,從事電子電路的分析與設(shè)計及物聯(lián)網(wǎng)方向的研究。Email:zsy197733@163.com
10.11975/j.issn.1002-6819.2020.01.023
S818.9
A
1002-6819(2020)-01-0200-08