国产日韩欧美一区二区三区三州_亚洲少妇熟女av_久久久久亚洲av国产精品_波多野结衣网站一区二区_亚洲欧美色片在线91_国产亚洲精品精品国产优播av_日本一区二区三区波多野结衣 _久久国产av不卡

?

基于CGA-BP神經(jīng)網(wǎng)絡(luò)的好氧堆肥曝氣供氧量預(yù)測(cè)模型

2023-06-12 05:22:04丁國(guó)超施雪玲
關(guān)鍵詞:氧量遺傳算法神經(jīng)網(wǎng)絡(luò)

丁國(guó)超,施雪玲,胡 軍

基于CGA-BP神經(jīng)網(wǎng)絡(luò)的好氧堆肥曝氣供氧量預(yù)測(cè)模型

丁國(guó)超1,施雪玲1,胡 軍2※

(1. 黑龍江八一農(nóng)墾大學(xué)信息與電氣工程學(xué)院,大慶 163319; 2. 黑龍江八一農(nóng)墾大學(xué)工程學(xué)院,大慶 163319)

為提高好氧堆肥曝氣供氧量的曝氣效率以及預(yù)測(cè)精度,該研究利用遺傳算法(genetic algorithm, GA)對(duì)標(biāo)準(zhǔn)反向傳播(back propagation, BP)神經(jīng)網(wǎng)絡(luò)的初始權(quán)值和閾值進(jìn)行優(yōu)化,再利用克隆選擇算法(clonal genetic algorithm, CGA)優(yōu)化遺傳算法中的變異算子并復(fù)制算子,加快獲取最優(yōu)參數(shù)的速度,構(gòu)建基于CGA-BP神經(jīng)網(wǎng)絡(luò)的曝氣供氧量預(yù)測(cè)模型。為驗(yàn)證CGA-BP模型的有效性,與BP模型、GA-BP模型預(yù)測(cè)結(jié)果進(jìn)行對(duì)比。試驗(yàn)結(jié)果表明:克隆遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)能加快獲得最優(yōu)解,效率相比BP模型和GA-BP模型分別提高了75.36%、51.30%;在曝氣供氧量預(yù)測(cè)模型中,CGA-BP模型具有更準(zhǔn)確的預(yù)測(cè)效果,預(yù)測(cè)精度為99.65%,而BP模型與GA-BP模型預(yù)測(cè)精度分別為96.99%、99.26%;CGA-BP模型評(píng)價(jià)指標(biāo)的均方誤差、平均絕對(duì)誤差、平均絕對(duì)百分誤差分別為0.003 4、0.038 9和0.350 6,均小于BP神經(jīng)網(wǎng)絡(luò)和GA-BP神經(jīng)網(wǎng)絡(luò)模型評(píng)價(jià)指標(biāo)的誤差;利用CGA-BP好氧堆肥曝氣供氧量預(yù)測(cè)模型對(duì)好氧堆肥發(fā)酵過程進(jìn)行精準(zhǔn)曝氣,提高了3.22%的曝氣控制效率。由此可知CGA-BP神經(jīng)網(wǎng)絡(luò)模型有更好的預(yù)測(cè)效果,可滿足好氧堆肥在發(fā)酵過程中曝氣供氧量的需求,提高曝氣效率,為精準(zhǔn)控制曝氣提供更直接有效的方法。

模型;試驗(yàn);遺傳算法;好氧堆肥;曝氣供氧;BP神經(jīng)網(wǎng)絡(luò);CGA-BP神經(jīng)網(wǎng)絡(luò)

0 引 言

中國(guó)年產(chǎn)各類有機(jī)廢棄物大約有45~50億t,其中農(nóng)業(yè)廢棄物9.8億t、林業(yè)廢棄物1.6億t、有機(jī)生活垃圾1.5億t、畜禽糞污19億t左右[1-2]。隨意棄置未被處理的有機(jī)廢棄物不僅浪費(fèi)大量資源,而且造成環(huán)境污染。好氧堆肥是對(duì)農(nóng)業(yè)有機(jī)廢棄物無害化處理和資源化利用的有效方式之一[3-5]。好氧堆肥發(fā)酵過程中,堆肥中存在的微生物起著重要的作用,能促進(jìn)堆肥反應(yīng)的正常進(jìn)行并充分反應(yīng)并腐熟的堆肥才可作為有機(jī)肥產(chǎn)品用于農(nóng)業(yè)施肥,提高作物產(chǎn)量而曝氣影響好氧堆肥中微生物的活性,能促進(jìn)對(duì)有機(jī)物的氧化分解,曝氣裝置以及曝氣量在一定程度上決定著好氧堆肥發(fā)酵中有機(jī)物分解的效率和腐熟程度。因此,曝氣量在好氧堆肥發(fā)酵過程中起著至關(guān)重要的作用[6]。

國(guó)內(nèi)外學(xué)者在精準(zhǔn)控制曝氣方面已進(jìn)行了一些研究,既往大多利用曝氣控制系統(tǒng)對(duì)曝氣量進(jìn)行控制。李升等[7]分析自動(dòng)曝氣系統(tǒng)運(yùn)行情況,表明供氧量的控制精度穩(wěn)定在1%左右,水中溶解氧的穩(wěn)定程度超過0.9。沈軍等[8]利用精確曝氣系統(tǒng)研究污水池中的含氧量,通過參數(shù)控制,曝氣池內(nèi)的含氧量變化不超過±0.5 mg/L,水質(zhì)達(dá)標(biāo)率提升5%左右。唐維等[9]利用污水進(jìn)出水的指標(biāo)建立遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)模型,預(yù)測(cè)生化池的曝氣量,測(cè)試樣本數(shù)據(jù)預(yù)測(cè)誤差在5%之內(nèi)的比例達(dá)到98.67%。CHEN等[10]提出多變量最優(yōu)控制模型,以曝氣過程中的最小耗能作為目標(biāo)函數(shù),保證系統(tǒng)正常供氧量時(shí)實(shí)現(xiàn)最優(yōu)控制,能耗降低20%左右。上述研究均是對(duì)污水進(jìn)行曝氣,與對(duì)好氧堆肥進(jìn)行曝氣的原料、方式有差異,好氧堆肥曝氣量的研究大多是基于經(jīng)驗(yàn)對(duì)堆肥定量定時(shí)進(jìn)行曝氣,按照一定通風(fēng)速率間歇通氣[11-12],曝氣效率低。綜上,曝氣量作為影響堆肥發(fā)酵的重要因素,利用堆肥發(fā)酵過程中的多項(xiàng)指標(biāo)與曝氣量之間的關(guān)系建立曝氣量預(yù)測(cè)模型,為能夠有效指導(dǎo)堆肥順利進(jìn)行、提高曝氣控制效率奠定基礎(chǔ)。

本文從提高曝氣效率出發(fā),利用克隆選擇算法(clonal genetic algorithm, CGA)的并行性、自適應(yīng)性等優(yōu)點(diǎn),保持群體的多樣性,搜索最優(yōu)解的速度較快,并結(jié)合遺傳算法優(yōu)化BP(back propagation)神經(jīng)網(wǎng)絡(luò)初始權(quán)值和閾值[13-16],建立CGA-BP神經(jīng)網(wǎng)絡(luò)的曝氣供氧量預(yù)測(cè)模型,以預(yù)測(cè)好氧堆肥發(fā)酵所需的曝氣供氧量,縮短好氧堆肥發(fā)酵時(shí)間,為精準(zhǔn)控制曝氣提供有力的技術(shù)支撐。

1 材料與方法

1.1 試驗(yàn)裝置

好氧堆肥試驗(yàn)以牛糞、牛糞沼渣、雞糞和玉米秸稈為好氧堆肥原料,由北京市密云區(qū)海華沼氣廠提供。其試驗(yàn)設(shè)備放置于北京市農(nóng)林科學(xué)院科技成果展示示范溫室,各原料初始理化參數(shù)如表1所示。

表1 堆肥原料初始理化參數(shù)

影響好氧堆肥發(fā)酵的因素主要包括堆體的溫度、有機(jī)質(zhì)含量、含水率、pH值、發(fā)氧濃度以及碳氮比等[17-19]。本文試驗(yàn)利用影響堆肥發(fā)酵的6個(gè)指標(biāo)(溫度、濕度、氧氣濃度、室溫、pH值和EC(electrical conductivity)值與曝氣量的關(guān)系建立曝氣量預(yù)測(cè)模型。各指標(biāo)參數(shù)數(shù)據(jù)信息通過發(fā)酵罐內(nèi)的不同類型傳感器采集,好氧堆肥試驗(yàn)裝置如圖1所示,利用自主研發(fā)的控制器獲取、傳輸并存儲(chǔ)傳感器數(shù)據(jù)。主要的傳感器參數(shù)如表2所示,其中EP-200傳感器是北京農(nóng)林科學(xué)院根據(jù)DS18B20溫度傳感器與ECH2O土壤水分傳感器自主研發(fā)的無線溫濕度傳感器。

圖1 好氧堆肥試驗(yàn)裝置結(jié)構(gòu)簡(jiǎn)圖

表2 溫濕度、氧氣濃度性能參數(shù)

1.2 數(shù)據(jù)來源及處理

1.2.1 數(shù)據(jù)來源及測(cè)定

試驗(yàn)選取的數(shù)據(jù)來源于2019年1月4日-22日的1號(hào)反應(yīng)器數(shù)據(jù)。從堆肥發(fā)酵開始,每間隔2 h利用EP-200溫濕度傳感器以及O2S-FR-T2-18X氧氣傳感器分別采集一次數(shù)據(jù),并在3個(gè)不同高度的取樣口進(jìn)行取樣,均勻混合后取出等量樣品180 g進(jìn)行實(shí)驗(yàn)室化驗(yàn)分析,pH值與EC值分別采用按照國(guó)標(biāo)土壤pH的測(cè)定(NY/T 1377-2007)和電導(dǎo)法測(cè)定。

1.2.2 數(shù)據(jù)預(yù)處理

好氧堆肥測(cè)定的指標(biāo)數(shù)據(jù)單位均不同,為加快程序運(yùn)行的收斂速度,對(duì)樣本數(shù)據(jù)做歸一化處理,本文采用mapminmax函數(shù)使數(shù)據(jù)均勻分布在[0,1]之間,其計(jì)算式為

式中表示輸入數(shù)據(jù),表示輸出數(shù)據(jù);min和max分別表示的最小值和最大值;min和max分別為的最小值和最大值;new和new分別為歸一化后的輸入和輸出數(shù)據(jù)。

試驗(yàn)修復(fù)樣本數(shù)據(jù)、剔除異常值后,選取數(shù)據(jù)樣本共268組,隨機(jī)選擇其中的218組數(shù)據(jù)作為測(cè)試集,余下的50組數(shù)據(jù)作為檢驗(yàn)集。通過MATLAB分別建立BP、GA-BP、CGA-BP神經(jīng)網(wǎng)絡(luò)曝氣供氧量預(yù)測(cè)模型,并計(jì)算模型誤差,分析模型預(yù)測(cè)精度、平均絕對(duì)誤差、平均絕對(duì)百分誤差、均方根誤差。

1.3 曝氣供氧量預(yù)測(cè)模型的建立

1.3.1 BP神經(jīng)網(wǎng)絡(luò)

本次試驗(yàn)的BP神經(jīng)網(wǎng)絡(luò)的輸入層、輸出層的神經(jīng)元個(gè)數(shù)分別設(shè)置為6和1。隱含層神經(jīng)元個(gè)數(shù)會(huì)影響神經(jīng)網(wǎng)絡(luò)的性能,本文采用經(jīng)驗(yàn)公式法來確定BP神經(jīng)網(wǎng)絡(luò)隱含層的節(jié)點(diǎn)數(shù),選取網(wǎng)絡(luò)誤差最小時(shí)對(duì)應(yīng)的隱含層節(jié)點(diǎn)數(shù)。常用的經(jīng)驗(yàn)公式為

式中表示隱含層節(jié)點(diǎn)數(shù)目,為輸入層節(jié)點(diǎn)數(shù)目,為輸出層節(jié)點(diǎn)數(shù)目,為1~13之間的整數(shù),根據(jù)式(3)確定隱含層神經(jīng)元個(gè)數(shù)的范圍為3~15。

運(yùn)行程序,在樣本集和訓(xùn)練次數(shù)相同的情況下,計(jì)算得出不同隱含層節(jié)點(diǎn)數(shù)的均方誤差,如表3所示。從表3對(duì)比均方誤差可知,均方誤差最小,為0.000 447,對(duì)應(yīng)的隱含層節(jié)點(diǎn)數(shù)為14。因此文中神經(jīng)網(wǎng)絡(luò)模型隱含層節(jié)點(diǎn)數(shù)設(shè)為14。

通過上述分析,本次試驗(yàn)建立的BP神經(jīng)網(wǎng)絡(luò)好氧堆肥曝氣量的預(yù)測(cè)模型的輸入層、隱含層和輸出層的神經(jīng)元個(gè)數(shù)分別為6、14、1,網(wǎng)絡(luò)結(jié)構(gòu)如圖2所示。

1.3.2 GA-BP神經(jīng)網(wǎng)絡(luò)

本文將遺傳算法與BP神經(jīng)網(wǎng)絡(luò)相結(jié)合,建立基于遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的曝氣供氧量預(yù)測(cè)模型,記作GA-BP神經(jīng)網(wǎng)絡(luò)模型[20]。遺傳算法從任一初始種群出發(fā),通過隨機(jī)選擇、交叉和變異操作獲取的最優(yōu)參數(shù)對(duì)BP神經(jīng)網(wǎng)絡(luò)的權(quán)值和閾值進(jìn)行賦值和訓(xùn)練[21],計(jì)算種群適應(yīng)度值,最終找出基于GA-BP神經(jīng)網(wǎng)絡(luò)的曝氣供氧量預(yù)測(cè)模型的最優(yōu)個(gè)體。

表3 不同隱含層節(jié)點(diǎn)數(shù)對(duì)應(yīng)的均方誤差

圖2 BP神經(jīng)網(wǎng)絡(luò)模型結(jié)構(gòu)圖

GA優(yōu)化BP神經(jīng)網(wǎng)絡(luò)算法參數(shù)主要有適應(yīng)度函數(shù)、種群規(guī)模、迭代次數(shù)、交叉概率和變異概率,下面對(duì)主要參數(shù)設(shè)置進(jìn)行確定:

1)適應(yīng)度函數(shù)

用來度量種群中個(gè)體適應(yīng)性的函數(shù)為適應(yīng)度函數(shù)[22],試驗(yàn)中的適應(yīng)度函數(shù)為:預(yù)測(cè)輸出和期望之間的誤差平方和的倒數(shù)。

2)種群規(guī)模

初始解的個(gè)數(shù)即為種群規(guī)模,與個(gè)體基因的復(fù)雜程度有關(guān)。取值越大越利于找尋全局最優(yōu)解,但運(yùn)行時(shí)間過長(zhǎng)。為了讓初始解在解空間分布均勻,考慮時(shí)間及效率成本取種群規(guī)模為10~80[23],本文的種群規(guī)模由優(yōu)化精度比較確定,結(jié)果如表4所示。

表4 不同種群規(guī)模、迭代次數(shù)算法誤差比較

根據(jù)表4中的種群規(guī)模的優(yōu)化精度算法對(duì)比,種群規(guī)模為40時(shí),試驗(yàn)優(yōu)化精度值最高。

3)迭代次數(shù)

遺傳算法收斂時(shí)達(dá)到的精度和系統(tǒng)的性能會(huì)受到迭代次數(shù)的影響,需要平衡算法的精度和執(zhí)行效率。利用適應(yīng)度函數(shù)控制迭代次數(shù),當(dāng)?shù)螖?shù)達(dá)到30時(shí),算法已收斂(表4)。由表4中不同迭代次數(shù)的算法誤差比較可知,設(shè)置迭代次數(shù)為30,此時(shí)算法誤差最小為0.003 694。

4)交叉概率

通過交叉操作產(chǎn)生新的個(gè)體,可提高遺傳算法的全局搜索能力[24]。交叉概率的大小要求盡可能不破壞個(gè)體結(jié)構(gòu)和群體中優(yōu)良的模式,但要有效產(chǎn)生較好的新個(gè)體模式。因此,交叉概率一般設(shè)置在0.40~0.99之間[23]。在此范圍內(nèi),不同的交叉概率所對(duì)應(yīng)算法誤差比較如表5所示。

表5 不同交叉概率的算法誤差比較

根據(jù)表5對(duì)比不同交叉概率所得算法誤差可得,當(dāng)交叉概率為0.8時(shí),算法誤差較小,為0.013 289。

5)變異概率

遺傳算法的局部搜索能力會(huì)因設(shè)置變異概率的大小而變化。變異概率較大時(shí),能夠產(chǎn)生較多的新個(gè)體,但也可能改變較好的模式結(jié)構(gòu)導(dǎo)致近似于隨即搜索算法的性能[15];若變異概率取值太小,會(huì)抑制產(chǎn)生新的個(gè)體和抑制早熟現(xiàn)象的能力。一般限定在0.000 1~0.1之間以有效維持群體的多樣性[23],試驗(yàn)中需要產(chǎn)生較多新個(gè)體搜索最優(yōu)解,變異概率設(shè)置值最大時(shí)的算法模式結(jié)構(gòu)并未被破壞,因此,試驗(yàn)的變異概率設(shè)置為0.1。

1.3.3 CGA-BP神經(jīng)網(wǎng)絡(luò)好氧堆肥的曝氣供氧量預(yù)測(cè)模型的建立

為提高GA-BP算法運(yùn)行效率,縮短GA-BP模型訓(xùn)練時(shí)間,本文利用克隆免疫算法優(yōu)化GA-BP模型,該算法是2002年DE等通過仿生免疫反應(yīng)中的親和度的成熟過程提出的??寺∵x擇算法作為克隆免疫算法中的一種,具有自學(xué)習(xí)、記憶機(jī)制和并行性等優(yōu)點(diǎn),并成功應(yīng)用于多模態(tài)函數(shù)優(yōu)化、組合優(yōu)化等方面[25]。BP神經(jīng)網(wǎng)絡(luò)經(jīng)過克隆遺傳算法優(yōu)化后,記作CGA-BP神經(jīng)網(wǎng)絡(luò)模型。相比GA-BP神經(jīng)網(wǎng)絡(luò)模型、BP神經(jīng)網(wǎng)絡(luò)模型,CGA-BP神經(jīng)網(wǎng)絡(luò)模型加快了搜索最優(yōu)解的速度,提升算法的運(yùn)算效率。

克隆遺傳算法以遺傳算法為主,當(dāng)進(jìn)行遺傳交叉和變異之后,利用克隆、高頻變異保持群體的多樣性,不斷迭代得到最優(yōu)個(gè)體[26]。本文建立的CGA-BP神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)曝氣供氧量模型的流程圖,如圖3所示。通過克隆選擇算法對(duì)BP神經(jīng)網(wǎng)絡(luò)和GA的參數(shù)進(jìn)行優(yōu)化,采用比例復(fù)制算子和比例變異算子保持群體的多樣性,同時(shí)利用記憶單元平衡全局與局部搜索的能力[27]。通過上述克隆遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò),CGA-BP神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)曝氣供氧量模型的參數(shù)設(shè)置如表6。

圖3 基于CGA-BP神經(jīng)網(wǎng)絡(luò)的流程圖

表6 CGA-BP神經(jīng)網(wǎng)絡(luò)模型參數(shù)設(shè)置

1.4 模型評(píng)價(jià)指標(biāo)

2 結(jié)果與分析

2.1 模型預(yù)測(cè)結(jié)果

為檢驗(yàn)克隆遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的效果,本文利用測(cè)試集對(duì)訓(xùn)練好的BP模型、GA-BP模型和CGA-BP模型的好氧堆肥曝氣供氧量預(yù)測(cè)效果進(jìn)行對(duì)比分析,并分別將3種模型的預(yù)測(cè)值與真實(shí)值進(jìn)行對(duì)比,結(jié)果如圖4所示。

CGA-BP模型與真實(shí)值的擬合程度最好,預(yù)測(cè)精度達(dá)到99.65%,比BP模型、GA-BP模型的預(yù)測(cè)精度96.99%、99.26%分別提高2.66個(gè)百分點(diǎn)、0.39個(gè)百分點(diǎn)。CGA-BP神經(jīng)網(wǎng)絡(luò)算法、GA-BP神經(jīng)網(wǎng)絡(luò)算法和BP算法的曝氣供氧量變化趨勢(shì)與真實(shí)值變化趨勢(shì)整體是一致的。但是本文提出的CGA-BP模型的預(yù)測(cè)精度相比其他2種模型都有一定提升。

本文通過影響堆肥發(fā)酵的多指標(biāo)與曝氣量的關(guān)系直接建立BP曝氣供氧量模型、GA-BP曝氣供氧量模型和CGA-BP曝氣供氧量預(yù)測(cè)模型,3種模型中,CGA-BP模型預(yù)測(cè)精度最高。試驗(yàn)結(jié)果表明,CGA-BP模型預(yù)測(cè)曝氣量能夠更準(zhǔn)確預(yù)測(cè)好氧堆肥發(fā)酵過程中堆肥實(shí)際需要的曝氣量,提高堆肥曝氣控制效率。

2.2 模型誤差對(duì)比

BP神經(jīng)網(wǎng)絡(luò)算法、GA-BP神經(jīng)網(wǎng)絡(luò)和CGA-BP神經(jīng)網(wǎng)絡(luò)3種算法的曝氣供氧量的絕對(duì)誤差值比較,如圖5所示。從圖5中3個(gè)模型的絕對(duì)誤差值可知,BP神經(jīng)網(wǎng)絡(luò)與GA-BP神經(jīng)網(wǎng)絡(luò)的誤差波動(dòng)范圍較大,BP神經(jīng)網(wǎng)絡(luò)模型誤差波動(dòng)在±0.5,GA-BP神經(jīng)網(wǎng)絡(luò)模型誤差波動(dòng)在±0.4,而CGA-BP神經(jīng)網(wǎng)絡(luò)模型的誤差穩(wěn)定在±0.1之間,相比其他2個(gè)模型,絕對(duì)誤差波動(dòng)范圍更小,預(yù)測(cè)結(jié)果更準(zhǔn)確。

此外,將3種好氧堆肥的曝氣供氧量預(yù)測(cè)模型的評(píng)價(jià)指標(biāo)進(jìn)行對(duì)比,結(jié)果見表7。CGA-BP模型的MAE值、MAPE值和MSE值分別為0.038 9、0.350 6、0.003 4,相比其他2個(gè)模型的性能指標(biāo)也都有很大提升。GA-BP神經(jīng)網(wǎng)絡(luò)模型的MAE值、MAPE值和MSE值與BP模型相比,分別提高了29.49%、30.16%和53.25%,CGA-BP模型的MAE值、MAPE值和MSE值比GA-BP模型分別提高了53.16%、52.56%和74.43%。在搜索最優(yōu)解的速度上,CGA-BP模型的效率也更高,比BP模型和GA-BP模型分別提高了75.36%、51.30%。

圖4 BP、GA-BP和CGA-BP模型預(yù)測(cè)值和真實(shí)值比較

圖5 BP、GA-BP和CGA-BP神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)模型誤差

表7 BP、GA-BP和CGA-BP模型評(píng)價(jià)指標(biāo)對(duì)比

基于上述分析得出,克隆遺傳算法起到了優(yōu)化作用,CGA-BP好氧堆肥曝氣供氧量預(yù)測(cè)模型優(yōu)于GA-BP神經(jīng)網(wǎng)絡(luò)模型和BP神經(jīng)網(wǎng)絡(luò)模型,能夠更好地預(yù)測(cè)曝氣供氧量。

另一方面,CGA-BP模型是利用好氧堆肥實(shí)驗(yàn)裝置控制傳感器采集到的數(shù)據(jù)進(jìn)行預(yù)測(cè)堆肥所需的曝氣量,數(shù)據(jù)異常會(huì)影響曝氣量預(yù)測(cè)結(jié)果的準(zhǔn)確性,需要及時(shí)對(duì)各傳感器進(jìn)行定期維護(hù)和堆肥發(fā)酵過程中異常值的監(jiān)測(cè),保證預(yù)測(cè)結(jié)果的準(zhǔn)確性,減少誤差。

2.3 試驗(yàn)驗(yàn)證

選取北京市密云區(qū)海華沼氣廠2019年2月2組堆肥數(shù)據(jù)進(jìn)行分析。在整個(gè)好氧堆肥發(fā)酵過程中,一組數(shù)據(jù)根據(jù)CGA-BP模型預(yù)測(cè)的曝氣量進(jìn)行控制曝氣,另一組數(shù)據(jù)直接利用實(shí)際值曝氣作為對(duì)照組,并將CGA-BP模型預(yù)測(cè)曝氣量和實(shí)際曝氣量進(jìn)行對(duì)比分析。兩組好氧堆肥反應(yīng)均完成后,對(duì)照組曝氣量總共為3 366.67 m3/h,預(yù)測(cè)模型曝氣量為3 257.82 m3/h,利用CGA-BP模型預(yù)測(cè)曝氣量提高了約3.22%的曝氣控制效率,并且根據(jù)CGA-BP模型預(yù)測(cè)曝氣供氧量對(duì)堆肥進(jìn)行精準(zhǔn)曝氣試驗(yàn),比對(duì)照組提前6 h結(jié)束發(fā)酵過程,縮短了好氧堆肥發(fā)酵的時(shí)間,達(dá)到了節(jié)能減排的效果。

基于上述分析,通過建模方法的CGA-BP神經(jīng)網(wǎng)絡(luò)的好氧堆肥曝氣供氧量預(yù)測(cè)模型能夠?qū)崿F(xiàn)精準(zhǔn)曝氣的效果并提高堆肥發(fā)酵效率。

3 結(jié) 論

本研究針對(duì)好氧堆肥曝氣量的預(yù)測(cè)精度,利用樣本數(shù)據(jù)訓(xùn)練神經(jīng)網(wǎng)絡(luò),根據(jù)程序運(yùn)行結(jié)果調(diào)整模型的參數(shù),最終建立3種好氧堆肥曝氣供氧量預(yù)測(cè)模型,對(duì)3種模型進(jìn)行對(duì)比分析,結(jié)論如下:

1)克隆遺傳算法對(duì)BP神經(jīng)網(wǎng)絡(luò)起到了優(yōu)化作用,加快獲得最優(yōu)解,效率相比BP模型和GA-BP模型分別提高了75.36%、51.30%。

2)在好氧堆肥曝氣供氧量預(yù)測(cè)模型中,CGA-BP模型的預(yù)測(cè)結(jié)果更佳,預(yù)測(cè)精度達(dá)到99.65%,比BP模型、GA-BP模型的預(yù)測(cè)精度分別提高2.66、0.39個(gè)百分點(diǎn)。CGA-BP模型的均方誤差、平均絕對(duì)誤差、平均絕對(duì)百分誤差分別為0.003 4、0.038 9和0.350 6,均小于BP神經(jīng)網(wǎng)絡(luò)和GA-BP神經(jīng)網(wǎng)絡(luò)模型評(píng)價(jià)指標(biāo)的誤差。

3)CGA-BP好氧堆肥曝氣供氧量預(yù)測(cè)模型的預(yù)測(cè)精度較高,能對(duì)好氧堆肥發(fā)酵過程實(shí)現(xiàn)精準(zhǔn)曝氣,提高了3.22%的曝氣控制效率,并縮短了堆肥發(fā)酵時(shí)間。

[1] TOKARSKI D, ?IME?KOVá J, KU?ERíK J, et al. Detectability of degradable organic matter in agricultural soils by thermogravimetry[J]. Journal of Plant Nutrition and Soil Science, 2019, 182(5): 729-740.

[2] GUO H, GU J, WANG X, et al. Negative effects of oxytetracycline and copper on nitrogen metabolism in an aerobic fermentation system: Characteristics and mechanisms[J]. Journal of Hazardous Materials, 2021, 403: 123890.

[3] CUI X, GUO L, LI C, et al. The total biomass nitrogen reservoir and its potential of replacing chemical fertilizers in China[J]. Renewable and Sustainable Energy Reviews, 2021, 135: 110215.

[4] 沈玉君,張朋月,孟海波,等. 通風(fēng)方式對(duì)豬糞堆肥主要臭氣物質(zhì)控制的影響研究[J]. 農(nóng)業(yè)工程學(xué)報(bào),2019,35(7):203-209.

SHEN Yujun, ZHANG Pengyue, MENG Haibo, et al. Study on the influence of ventilation mode on the control of main odorous substances in pig manure compost[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2019, 35(7): 203-209. (in Chinese with English abstract)

[5] 張秧, 艾為黨, 馮海艷,等. 小麥秸桿好氧堆肥過程中微生物多樣性與優(yōu)勢(shì)菌群分析[J]. 農(nóng)業(yè)工程學(xué)報(bào), 2021, 37(11):206-212.

ZHANG Yang, AI Weidang, FENG Haiyan, et al. Analysis of microbial diversity and dominant microbiota during aerobic composting of wheat straw[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2021, 37(11): 206-212. (in Chinese with English abstract)

[6] 王雅雅,陳高攀,孫浩,等. 基于EDEM的分倉(cāng)立式好氧堆肥反應(yīng)器設(shè)計(jì)與試驗(yàn)[J]. 農(nóng)業(yè)工程學(xué)報(bào),2023,39(3):171-179.

WANG Yaya, CHEN Gaopan, SUN Hao, et al. Design and test of separate silo vertical aerobic composting reactor based on EDEM[J]. Transactions of the Chinese Society of Agricultural Engineering, 2023, 39(3): 171-179. (in Chinese with English abstract)

[7] 李升,胡曉東,許飛飛,等. 馬頭崗污水廠精確曝氣系統(tǒng)的實(shí)施及應(yīng)用[J]. 中國(guó)給水排水,2016,32(15):26-31.

LI Sheng, HU Xiaodong, XU Feifei, et al. Implementation and application of precision aeration system of Matougang sewage plant[J]. Water Supply and Drainage in China, 2016, 32(15): 26-31. (in Chinese with English abstract)

[8] 沈軍. 精確曝氣系統(tǒng)(AVS)在 AAO 工藝中的運(yùn)行分析[J]. 凈水技術(shù),2016,35(5):73-78.

SHEN Jun. Operational analysis of precision aeration system (AVS) in AAO processes[J]. Water Purification Technology, 2016, 35(5): 73-78.(in Chinese with English abstract)

[9] 唐維,陶鈺欣,郝啟文,等.基于遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的曝氣量預(yù)測(cè)[J].控制工程,2022,29(7):1600-1605.

TANG Wei, TAO Yuxin, HAO Qiwen, et al. Optimization of aeration prediction of BP neural network based on genetic algorithm[J]. Control Engineering, 2022, 29(7):1600-1605.(in Chinese with English abstract)

[10] CHEN Y S, ZHANG H J, YIN Y F, et al. Smart energy savings for aeration control in wastewater treatment[J]. Energy Reports, 2022, 8: 1711-1721.

[11] 王麗麗,許雷,姚紀(jì)宇,等.堆肥廢氣余熱回用對(duì)寒區(qū)好氧堆肥的影響[J].農(nóng)業(yè)工程學(xué)報(bào),2022,38(7):237-244.

WANG Lili, XU Lei, YAO Jiyu, et al. Effect of waste heat reuse of compost waste gas on aerobic composting in cold areas[J]. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 2022, 38(7): 237-244. (in Chinese with English abstract)

[12] 王友玲,邱慧珍,PHILIP Ghanney,等.通風(fēng)方式對(duì)牛糞堆肥氨氣排放與氮素轉(zhuǎn)化的影響[J].農(nóng)業(yè)機(jī)械學(xué)報(bào),2020,51(11):313-320.

WANG Youling, QIU Huizhen, PHILIP Ghanney, et al. Effects of ventilation on ammonia emission and nitrogen conversion of cow manure compost[J]. Transactions of the Chinese Society for Agricultural Machinery, 2020, 51(11): 313-320. (in Chinese with English abstract)

[13] 李成杰,王濤,李海漪,等. 基于BP神經(jīng)網(wǎng)絡(luò)的堆肥物料抗剪強(qiáng)度預(yù)測(cè)模型[J]. 中國(guó)給水排水,2020,36(7):108-113.

LI Chengjie, WANG Tao, LI Haiyi, et al. Shear strength prediction model of compost materials based on BP neural network[J]. Water Supply and Drainage in China, 2020, 36(7): 108-113. (in Chinese with English abstract)

[14] 聞新. 智能故障診斷技術(shù):MATLAB應(yīng)用[M]. 北京:北京航空航天大學(xué)出版社,2015.

[15] ZHANG X, LIAO Z, MA L, et al. Hierarchical multistrategy genetic algorithm for integrated process planning and scheduling[J]. Journal of Intelligent Manufacturing, 2022, 33: 223-246.

[16] ZHENG Y C, XUE Y Y, YA L L, et al. Throughput optimization in cognitive wireless network based on clone selection algorithm[J]. Computers and Electrical Engineering, 2016, 52(5): 328-336.

[17] TRAN H T, LIN C, BUI X T, et al. Aerobic composting remediation of petroleum hydrocarbon-contaminated soil. Current and future perspectives[J]. Science of the Total Environment, 2021, 753: 142250.

[18] 劉娟,沈玉君,羅文海,等. 鹽含量對(duì)餐廚垃圾好氧堆肥腐殖化過程及微生物演變的影響[J].農(nóng)業(yè)工程學(xué)報(bào),2022,38(19):190-201.

LIU Juan, SHEN Yujun, LUO Wenhai, et al. Effects of salt content on aerobic compost humification process and microbial evolution of food waste[J]. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 2022, 38(19): 190-201. (in Chinese with English abstract)

[19] 楊佳,王國(guó)英,唐若蘭,等. 生物炭和菌劑對(duì)羊糞微好氧堆肥腐熟度和溫室氣體排放的影響[J].農(nóng)業(yè)工程學(xué)報(bào),2022,38(10):224-231.

YANG Jia, WANG Guoying, TANG Ruolan, et al. Effects of biochar and fungicides on decay and greenhouse gas emissions of microaerobic compost of sheep manure[J]. Transactions of the Chinese Society of Agricultural Engineering(Transactions of the CSAE), 2022, 38(10): 224-231. (in Chinese with English abstract)

[20] ZHAO L, HU Y M, ZHOU W, et al. Estimation methods for soil mercury content using hyperspectral remote sensing[J]. Sustainability, 2018, 10(7): 2474.

[21] 郭利進(jìn),喬志忠. 基于遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的糧食溫度預(yù)測(cè)研究[J]. 糧食與油脂,2023,36(1):34-37,51.

GUO Lijin, QIAO Zhizhong. Research on grain temperature prediction based on BP neural network optimization based on genetic algorithm[J]. Grain and Fat, 2023, 36(1): 34-37, 51. (in Chinese with English abstract)

[22] ISMAIL F S, BAKAR N A. Adaptive mechanism for GA-NN to enhance prediction model[C] //Proceedings of the 9th International Conference on Ubiquitous Information Management and Communication, New York, US: ACM, 2015: 1-5.

[23] 周明,孫樹棟. 遺傳算法原理及應(yīng)用[M]. 北京:國(guó)防工業(yè)出版社,1999.

[24] LIU R, LIU L. Predicting housing price in China based on long short-term memory incorporating modified genetic algorithm[J]. Soft Computing, 2019, 23(22): 11829-11838.

[25] DE C L N, VON Z F J. Learning and optimization using the clonal selection principle[J]. IEEE Transactions on Evolutionary Computation, 2002, 6(3): 239-251.

[26] 李陽,周步祥,林楠,等.修正克隆遺傳算法在分布式電源規(guī)劃中的應(yīng)用[J].電力系統(tǒng)及其自動(dòng)化學(xué)報(bào),2013,25(4):128-132.

LI Yang, ZHOU Buxiang, LIN Nan, et al. Correction of the application of cloning genetic algorithm in distributed power planning[J]. Proceedings of the Department of Power Systems and its Automation, 2013, 25(4): 128-132.(in Chinese with English abstract)

[27] 謝宏,李云峰,禹文科,等. 改進(jìn)免疫克隆選擇算法的多目標(biāo)軌跡優(yōu)化[J]. 電子測(cè)量與儀器學(xué)報(bào),2016,30(10):1534-1542.

XIE Hong, LI Yunfeng, YU Wenke, et al. Improved multi-objective trajectory optimization of immunocloning selection algorithm[J]. Journal of Electronic Measurement and Instrumentation, 2016, 30(10): 1534-1542. (in Chinese with English abstract)

[28] YAN J, XU Z, YU Y, et al. Application of a hybrid optimized BP network model to estimate water quality parameters of Beihai Lake in Beijing[J]. Applied Sciences, 2019, 9(9): 1863.

[29] 陳嘯,王紅英,孔丹丹,等.基于粒子群參數(shù)優(yōu)化和BP神經(jīng)網(wǎng)絡(luò)的顆粒飼料質(zhì)量預(yù)測(cè)模型[J]. 農(nóng)業(yè)工程學(xué)報(bào),2016,32(14):306-314.

CHEN Xiao, WANG Hongying, KONG Dandan, et al. Quality prediction model of pellet feed basing on BP network using PSO parameters optimization method[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2016, 32(14): 306-314. (in Chinese with English abstract)

Prediction model of the aeration oxygen supply for aerobic composting using CGA-BP neural network

DING Guochao1, SHI Xueling1, HU Jun2※

(1.,,163319,;2.,,163319,)

Aerobic compost has been commonly used to efficiently dispose of resources recycling and environmental protection in modern agriculture. Among them, aeration can be one of the most important environmental factors to affect composting fermentation. It is necessary for a feasible network model to accurately control the oxygen supply of aeration. This study aims to improve the aeration efficiency and prediction accuracy of aerobic composting aeration. In-depth learning was selected to train a network model, in order to predict the oxygen supply of aeration during aerobic composting fermentation in this experiment. Raw materials were taken as cow dung, cow dung biogas residue, chicken manure, and corn straw in the Haihua Biogas Plant in Miyun District, Beijing, China. The corn straw was crushed by 1-2 cm in grain size. The cow dung, cow dung biogas residue, and chicken manure were uniformly mixed with the crushed corn straw for composting and fermentation. The sensor was used in the composting fermentation tank to collect the parameter data during aerobic fermentation. 268 groups of data were selected as the sample data, 218 groups of data were randomly selected as the input data, and 50 groups of data were selected as the test data. Clonal genetic algorithm (CGA) was used to predict the standard back propagation (BP) neural network model for the aeration oxygen supply, whereas, the 6-14-1 three-layer network structure was used as the basic structure of the prediction model. The input parameters were the temperature, humidity, oxygen concentration, room temperature, pH value, and electrical conductivity (EC). The mean square error (MSE) of the number of hidden layer nodes was determined to be 14 after training and calculation. The output data was aeration. This article establishes BP neural network model for predicting aeration oxygen supply. Then the genetic algorithm (GA) and clonal selection algorithm were used to improve the prediction accuracy of the model. The experiment shows that the CGA-BP neural network model has the best prediction effect on aeration oxygen supply. 1) The CGA-BP neural network model accelerated the obtaining of the optimal solution, with an efficiency improvement of 75.36% and 51.30% compared to the BP model and GA-BP model, respectively. 2) In the prediction model of aeration oxygen supply, the CGA-BP model had a more accurate prediction effect, with a prediction accuracy of 99.65%. The prediction accuracy of aeration oxygen supply was 96.99% and the prediction accuracy of the GA-BP neural network model reached 99.26%. A comparison was made to evaluate the errors of BP, GA-BP and CGA-BP neural models. The model evaluation showed that the best performance was found in the CGA-BP neural network model with the smallest error, as shown by the mean absolute error (MAE), mean absolute percentage error (MAPE) and mean square error (MSE). 3) The improved CGA-BP neural network model can be predicted the aeration oxygen supply of aerobic composting, increasing the aeration control efficiency by 3.22%. The improved model can be expected to accurately predict the aeration oxygen supply of aerobic compost. The finding can provide a strong reference for accurate data for the next aeration.

model; trial; genetic algorithms; aerobic composting; aerated oxygen supply; BP neural networks; CGA-BP neural networks

2022-11-09

2023-02-06

國(guó)家重點(diǎn)研發(fā)計(jì)劃:農(nóng)業(yè)廢棄物好氧發(fā)酵技術(shù)與智能控制設(shè)備研發(fā)(2016YFD0800600)

丁國(guó)超,博士,副教授,研究方向?yàn)樯镄畔⑻幚怼mail:dgcer@163.com

胡軍,博士,教授,研究方向?yàn)檗r(nóng)業(yè)機(jī)械化工程。Email:gcxykj@126.com

10.11975/j.issn.1002-6819.202211088

TP183; TP389.1; S210.6

A

1002-6819(2023)-07-0211-07

丁國(guó)超,施雪玲,胡軍. 基于CGA-BP神經(jīng)網(wǎng)絡(luò)的好氧堆肥曝氣供氧量預(yù)測(cè)模型[J]. 農(nóng)業(yè)工程學(xué)報(bào),2023,39(7):211-217. doi:10.11975/j.issn.1002-6819.202211088 http://www.tcsae.org

DING Guochao, SHI Xueling, HU Jun. Prediction model of the aeration oxygen supply for aerobic composting using CGA-BP neural network[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2023, 39(7): 211-217. (in Chinese with English abstract) doi:10.11975/j.issn.1002-6819.202211088 http://www.tcsae.org

猜你喜歡
氧量遺傳算法神經(jīng)網(wǎng)絡(luò)
一種催化器儲(chǔ)氧量的測(cè)試及計(jì)算方法
神經(jīng)網(wǎng)絡(luò)抑制無線通信干擾探究
電子制作(2019年19期)2019-11-23 08:42:00
基于自適應(yīng)遺傳算法的CSAMT一維反演
一種基于遺傳算法的聚類分析方法在DNA序列比較中的應(yīng)用
基于遺傳算法和LS-SVM的財(cái)務(wù)危機(jī)預(yù)測(cè)
基于神經(jīng)網(wǎng)絡(luò)的拉矯機(jī)控制模型建立
基于改進(jìn)的遺傳算法的模糊聚類算法
復(fù)數(shù)神經(jīng)網(wǎng)絡(luò)在基于WiFi的室內(nèi)LBS應(yīng)用
基于GALM算法的循環(huán)流化床鍋爐氧量軟測(cè)量
河南科技(2015年3期)2015-02-27 14:20:52
基于支持向量機(jī)回歸和RBF神經(jīng)網(wǎng)絡(luò)的PID整定
财经| 绍兴市| 江门市| 临沂市| 昌黎县| 阳新县| 保定市| 肃北| 灵川县| 民勤县| 阜阳市| 南雄市| 万源市| 营山县| 桂阳县| 钟祥市| 宜都市| 淮安市| 怀柔区| 扶风县| 屯昌县| 霍山县| 吉木萨尔县| 松溪县| 买车| 南平市| 昌都县| 平陆县| 彭山县| 天气| 红桥区| 凌海市| 防城港市| 通河县| 华池县| 宜君县| 霸州市| 西乌珠穆沁旗| 镇赉县| 思茅市| 湖州市|