智(zhi)能(neng)控(kong)制是(shi)(shi)自動(dong)控(kong)制發展的(de)(de)(de)(de)高(gao)級階段，是(shi)(shi)人工智(zhi)能(neng)、控(kong)制論(lun)、系(xi)統(tong)論(lun)和信息論(lun)等(deng)多種學科的(de)(de)(de)(de)高(gao)度綜合與集成，它(ta)主要(yao)包(bao)括模糊控(kong)制、神(shen)經網絡控(kong)制、學習控(kong)制和專(zhuan)家控(kong)制等(deng)。智(zhi)能(neng)控(kong)制在各種非穩(wen)定的(de)(de)(de)(de)動(dong)態工程系(xi)統(tong)中的(de)(de)(de)(de)應用日益廣泛與深入，特(te)別是(shi)(shi)近(jin)年來取得(de)的(de)(de)(de)(de)研究與應用成果(guo)更(geng)受矚(zhu)目。由于(yu)污(wu)水(shui)處(chu)理(li)(li)的(de)(de)(de)(de)運行費用是(shi)(shi)龐大的(de)(de)(de)(de)、長期的(de)(de)(de)(de)，如果(guo)通過有效的(de)(de)(de)(de)控(kong)制能(neng)將(jiang)城市污(wu)水(shui)處(chu)理(li)(li)廠的(de)(de)(de)(de)運行費用節(jie)省1%，也是(shi)(shi)個(ge)天文數字。由此可(ke)見，加強城市污(wu)水(shui)處(chu)理(li)(li)系(xi)統(tong)智(zhi)能(neng)控(kong)制的(de)(de)(de)(de)研究非常必要(yao)。

1、國內外自(zi)控(kong)技術現狀分析

發達(da)國(guo)(guo)(guo)家(jia)在二級處(chu)(chu)(chu)理普及(ji)以后投(tou)入大量資(zi)金和科研力量加強污(wu)(wu)水(shui)處(chu)(chu)(chu)理設施的(de)(de)監測、運行(xing)和管理，實(shi)(shi)現了(le)計算機控制(zhi)、報警(jing)、計算和瞬時記錄。美國(guo)(guo)(guo)在20世紀70年(nian)代(dai)中期開始實(shi)(shi)現污(wu)(wu)水(shui)處(chu)(chu)(chu)理廠(chang)的(de)(de)自(zi)(zi)動(dong)(dong)控制(zhi)，目前主(zhu)要(yao)污(wu)(wu)水(shui)處(chu)(chu)(chu)理廠(chang)已實(shi)(shi)現了(le)工(gong)藝流(liu)程中主(zhu)要(yao)參數(shu)(shu)的(de)(de)自(zi)(zi)動(dong)(dong)測試和控制(zhi)。80年(nian)代(dai)以來在美國(guo)(guo)(guo)召開了(le)兩次(ci)水(shui)處(chu)(chu)(chu)理儀器和自(zi)(zi)動(dong)(dong)化(hua)的(de)(de)國(guo)(guo)(guo)際學術會議，會上發表的(de)(de)數(shu)(shu)百篇(pian)論(lun)文反(fan)映出水(shui)處(chu)(chu)(chu)理自(zi)(zi)動(dong)(dong)化(hua)已發展到實(shi)(shi)用水(shui)平(ping)。與國(guo)(guo)(guo)外相比，我國(guo)(guo)(guo)污(wu)(wu)水(shui)處(chu)(chu)(chu)理自(zi)(zi)動(dong)(dong)化(hua)控制(zhi)起步較晚，進入90年(nian)代(dai)以后污(wu)(wu)水(shui)處(chu)(chu)(chu)理廠(chang)才開始引入自(zi)(zi)動(dong)(dong)控制(zhi)系統(tong)，但(dan)多(duo)是(shi)直接引進國(guo)(guo)(guo)外成套自(zi)(zi)控設備，國(guo)(guo)(guo)產自(zi)(zi)動(dong)(dong)控制(zhi)系統(tong)在污(wu)(wu)水(shui)處(chu)(chu)(chu)理廠(chang)應用很少。

近年來，國內外均有學者對污水處(chu)理(li)自(zi)動(dong)控(kong)(kong)制(zhi)工(gong)藝進行研究，以尋(xun)求更精(jing)確、更可(ke)靠的方法實施(shi)自(zi)動(dong)控(kong)(kong)制(zhi)。Zipper等(deng)開發了適用于小型污水處(chu)理(li)廠(chang)的自(zi)動(dong)控(kong)(kong)制(zhi)系(xi)統，該(gai)系(xi)統采用基于氧(yang)化(hua)還原電位(ORP)的控(kong)(kong)制(zhi)器。這個控(kong)(kong)制(zhi)器自(zi)動(dong)工(gong)作，并可(ke)以在(zai)(zai)硝化(hua)和反硝化(hua)之間進行優(you)化(hua)，從而減(jian)少能耗，他們在(zai)(zai)中試中發現，污水處(chu)理(li)廠(chang)的實際負荷(he)與ORP曲線變化(hua)具有很強(qiang)的相(xiang)關性。采用兩點ORP控(kong)(kong)制(zhi)保證了在(zai)(zai)增加(jia)(jia)負荷(he)時(shi)硝化(hua)時(shi)間占運(yun)行時(shi)間的比率也隨之增加(jia)(jia)，這些都(dou)為開發小型污水處(chu)理(li)廠(chang)控(kong)(kong)制(zhi)規則奠定了基礎。

John等采用(yong)(yong)(yong)兩(liang)種SBR反(fan)(fan)應(ying)(ying)(ying)(ying)器(qi)(qi)(qi)對家禽生產廢(fei)水(shui)(shui)(shui)(shui)進(jin)行(xing)處(chu)理(li)(li)(li)，并且(qie)評價了它(ta)們的(de)(de)(de)(de)處(chu)理(li)(li)(li)效率，同時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)也考察了脫氮與反(fan)(fan)應(ying)(ying)(ying)(ying)器(qi)(qi)(qi)ORP之間(jian)(jian)(jian)(jian)的(de)(de)(de)(de)關系，并且(qie)使(shi)用(yong)(yong)(yong)了用(yong)(yong)(yong)于(yu)(yu)實(shi)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)pH、ORP和(he)(he)DO監控(kong)(kong)(kong)(kong)(kong)的(de)(de)(de)(de)先(xian)進(jin)儀器(qi)(qi)(qi)設備和(he)(he)基(ji)于(yu)(yu)ORP設定(ding)值控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)曝(pu)氣時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)間(jian)(jian)(jian)(jian)的(de)(de)(de)(de)過(guo)程(cheng)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)。當處(chu)理(li)(li)(li)變組分廢(fei)水(shui)(shui)(shui)(shui)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)，該研(yan)(yan)究不(bu)僅獲得了穩(wen)定(ding)的(de)(de)(de)(de)出水(shui)(shui)(shui)(shui)水(shui)(shui)(shui)(shui)質(zhi)，而(er)(er)且(qie)依靠ORP控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)曝(pu)氣時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)間(jian)(jian)(jian)(jian)，減少了空壓機(ji)(ji)(ji)的(de)(de)(de)(de)運(yun)行(xing)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)間(jian)(jian)(jian)(jian)。Yu等設計(ji)(ji)(ji)研(yan)(yan)究了一(yi)套帶有(you)實(shi)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)ORP和(he)(he)pH控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)系統的(de)(de)(de)(de)連續進(jin)水(shui)(shui)(shui)(shui)SBR反(fan)(fan)應(ying)(ying)(ying)(ying)器(qi)(qi)(qi)。該實(shi)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)監測和(he)(he)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)系統由傳感器(qi)(qi)(qi)、計(ji)(ji)(ji)算(suan)機(ji)(ji)(ji)、人機(ji)(ji)(ji)對話界(jie)面和(he)(he)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)部件組成(cheng)。SBR反(fan)(fan)應(ying)(ying)(ying)(ying)器(qi)(qi)(qi)中安(an)裝了4個(ge)(ge)(ge)帶有(you)Ag/AgCl電(dian)極(ji)的(de)(de)(de)(de)ORP儀表(biao)(biao)、1個(ge)(ge)(ge)DO表(biao)(biao)和(he)(he)1個(ge)(ge)(ge)pH儀表(biao)(biao)，傳感器(qi)(qi)(qi)的(de)(de)(de)(de)模擬(ni)信(xin)號通過(guo)AD/DA轉換(huan)器(qi)(qi)(qi)轉換(huan)成(cheng)數(shu)字信(xin)號，并且(qie)依靠計(ji)(ji)(ji)算(suan)機(ji)(ji)(ji)每秒鐘采集(ji)一(yi)次信(xin)號。計(ji)(ji)(ji)算(suan)機(ji)(ji)(ji)對采集(ji)來的(de)(de)(de)(de)數(shu)據進(jin)行(xing)分析后，通過(guo)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)線(xian)路傳遞到繼電(dian)器(qi)(qi)(qi)，由它(ta)開/關攪(jiao)拌器(qi)(qi)(qi)、潷水(shui)(shui)(shui)(shui)器(qi)(qi)(qi)和(he)(he)鼓風機(ji)(ji)(ji)。試驗(yan)結(jie)果顯示，采用(yong)(yong)(yong)實(shi)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)的(de)(de)(de)(de)SBR反(fan)(fan)應(ying)(ying)(ying)(ying)器(qi)(qi)(qi)在(zai)(zai)底物(wu)去(qu)除效率和(he)(he)降(jiang)低(di)能耗方面均優于(yu)(yu)采用(yong)(yong)(yong)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)序(xu)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)的(de)(de)(de)(de)SBR反(fan)(fan)應(ying)(ying)(ying)(ying)器(qi)(qi)(qi)。Puznava等在(zai)(zai)同步硝化(hua)/反(fan)(fan)硝化(hua)的(de)(de)(de)(de)生物(wu)濾池中引入了實(shi)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)曝(pu)氣控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)，建立(li)了基(ji)于(yu)(yu)DO在(zai)(zai)線(xian)監測的(de)(de)(de)(de)反(fan)(fan)饋(kui)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)和(he)(he)基(ji)于(yu)(yu)氨氮和(he)(he)DO在(zai)(zai)線(xian)監測的(de)(de)(de)(de)串聯(lian)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)。與傳統硝化(hua)—反(fan)(fan)硝化(hua)生物(wu)曝(pu)氣濾池(BAF)相(xiang)比，采用(yong)(yong)(yong)實(shi)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)曝(pu)氣控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)的(de)(de)(de)(de)生物(wu)濾池在(zai)(zai)達(da)(da)到相(xiang)同處(chu)理(li)(li)(li)效果(出水(shui)(shui)(shui)(shui)TN＜20mg/L)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)，曝(pu)氣量(liang)低(di)于(yu)(yu)傳統方法的(de)(de)(de)(de)50%。王淑瑩在(zai)(zai)國外已有(you)的(de)(de)(de)(de)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)間(jian)(jian)(jian)(jian)和(he)(he)流量(liang)程(cheng)序(xu)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)的(de)(de)(de)(de)基(ji)礎上，提(ti)出一(yi)種SBR法有(you)機(ji)(ji)(ji)物(wu)濃(nong)(nong)度(du)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)，使(shi)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)過(guo)程(cheng)更(geng)(geng)定(ding)量(liang)化(hua)和(he)(he)精密化(hua)。工業廢(fei)水(shui)(shui)(shui)(shui)的(de)(de)(de)(de)水(shui)(shui)(shui)(shui)質(zhi)變化(hua)很大，當進(jin)水(shui)(shui)(shui)(shui)有(you)機(ji)(ji)(ji)物(wu)濃(nong)(nong)度(du)高時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)，為使(shi)出水(shui)(shui)(shui)(shui)水(shui)(shui)(shui)(shui)質(zhi)達(da)(da)標，應(ying)(ying)(ying)(ying)適當增加反(fan)(fan)應(ying)(ying)(ying)(ying)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)間(jian)(jian)(jian)(jian)使(shi)運(yun)行(xing)更(geng)(geng)可(ke)靠；而(er)(er)當進(jin)水(shui)(shui)(shui)(shui)有(you)機(ji)(ji)(ji)物(wu)濃(nong)(nong)度(du)低(di)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)可(ke)以減少反(fan)(fan)應(ying)(ying)(ying)(ying)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)間(jian)(jian)(jian)(jian)以節省(sheng)運(yun)行(xing)費用(yong)(yong)(yong)。彭永臻等將(jiang)ORP作(zuo)(zuo)為SBR反(fan)(fan)應(ying)(ying)(ying)(ying)器(qi)(qi)(qi)有(you)機(ji)(ji)(ji)物(wu)降(jiang)解程(cheng)度(du)間(jian)(jian)(jian)(jian)接(jie)指(zhi)標的(de)(de)(de)(de)研(yan)(yan)究結(jie)果表(biao)(biao)明，無論是(shi)在(zai)(zai)很大范圍內(nei)改變曝(pu)氣量(liang)或者(zhe)改變MLSS濃(nong)(nong)度(du)，還是(shi)使(shi)反(fan)(fan)應(ying)(ying)(ying)(ying)初(chu)始COD在(zai)(zai)230～2180mg/L之間(jian)(jian)(jian)(jian)逐漸變化(hua)或突然變化(hua)，當COD達(da)(da)到難降(jiang)解濃(nong)(nong)度(du)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)，ORP都(dou)迅速(su)、大幅度(du)地(di)升高，隨后又很快趨于(yu)(yu)平穩(wen)，并在(zai)(zai)某一(yi)特定(ding)范圍內(nei)穩(wen)定(ding)下來。因此(ci)，可(ke)以用(yong)(yong)(yong)ORP作(zuo)(zuo)為SBR法反(fan)(fan)應(ying)(ying)(ying)(ying)時(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)(shi)間(jian)(jian)(jian)(jian)的(de)(de)(de)(de)計(ji)(ji)(ji)算(suan)機(ji)(ji)(ji)控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)參數(shu)，實(shi)現計(ji)(ji)(ji)算(suan)機(ji)(ji)(ji)在(zai)(zai)線(xian)自動控(kong)(kong)(kong)(kong)(kong)制(zhi)(zhi)(zhi)(zhi)。

通過以上分析，目前污水處理自動控制中存(cun)在以下問(wen)題：

① 傳(chuan)統污水(shui)處(chu)理(li)(li)自動(dong)控(kong)制系統要求建(jian)立精確(que)的(de)數學模(mo)型，并(bing)且提出(chu)必須(xu)遵(zun)循一(yi)些比較苛刻的(de)線性化(hua)假(jia)(jia)設，然(ran)而(er)實(shi)際污水(shui)處(chu)理(li)(li)系統由于存在復雜性、非線性、時變性、不(bu)確(que)定性和不(bu)完(wan)全性等，一(yi)般無法獲得精確(que)的(de)數學模(mo)型和與實(shi)際相(xiang)符的(de)假(jia)(jia)設，因此采用傳(chuan)統控(kong)制理(li)(li)論(lun)建(jian)立的(de)污水(shui)處(chu)理(li)(li)自動(dong)控(kong)制系統在實(shi)際工程應用上(shang)存在出(chu)水(shui)水(shui)質波動(dong)較大等問(wen)題。

② 污(wu)水(shui)處理(li)自動(dong)控制(zhi)系統中所采用的一些自動(dong)化檢(jian)測(ce)設(she)備(bei)(bei)、儀表的功能目(mu)(mu)前(qian)還很不完善，在實際(ji)檢(jian)測(ce)中達不到預(yu)期效果、誤差很大(da)，因此依靠這(zhe)些檢(jian)測(ce)設(she)備(bei)(bei)判(pan)斷污(wu)水(shui)處理(li)情(qing)況并實施自動(dong)控制(zhi)，往往很難(nan)達到處理(li)水(shui)質達標排放和節(jie)約(yue)能源的目(mu)(mu)的。

③ 國內外許(xu)(xu)(xu)多學者(zhe)為(wei)提高污(wu)水(shui)(shui)處(chu)(chu)理(li)廠(chang)的處(chu)(chu)理(li)效(xiao)率和(he)降(jiang)低能(neng)耗開展(zhan)了許(xu)(xu)(xu)多實時控制(zhi)研究，如(ru)采用ORP、DO和(he)pH值作為(wei)控制(zhi)參數(shu)來控制(zhi)出水(shui)(shui)水(shui)(shui)質和(he)減小曝(pu)氣量，但這(zhe)些方法也存在一些問題，例如(ru)控制(zhi)污(wu)水(shui)(shui)處(chu)(chu)理(li)廠(chang)硝化—反硝化過程所使用的ORP就(jiu)很(hen)難判定，因此絕大(da)多數(shu)基于ORP控制(zhi)的污(wu)水(shui)(shui)處(chu)(chu)理(li)廠(chang)也執行時間(jian)控制(zhi)，作為(wei)當(dang)控制(zhi)器無法找(zhao)到ORP特征點(dian)時的應(ying)急控制(zhi)，這(zhe)樣(yang)就(jiu)導致許(xu)(xu)(xu)多污(wu)水(shui)(shui)處(chu)(chu)理(li)系統(tong)實際上仍然(ran)采用的是(shi)按時間(jian)控制(zhi)整個處(chu)(chu)理(li)過程。

④ 污(wu)水處理自動(dong)控制(zhi)有(you)別于其他控制(zhi)系統，它需要(yao)(yao)對大(da)量閥(fa)門、泵、鼓風機和(he)吸(刮)泥(ni)(ni)機、曝氣池(chi)和(he)污(wu)泥(ni)(ni)消(xiao)化(hua)池(chi)內的(de)攪拌(ban)器等機械設備(bei)及沉淀池(chi)和(he)消(xiao)化(hua)池(chi)進(jin)、排泥(ni)(ni)量進(jin)行控制(zhi)，因(yin)此污(wu)水處理廠(chang)(chang)需要(yao)(yao)自動(dong)控制(zhi)的(de)開關量多，它們常(chang)常(chang)要(yao)(yao)根據(ju)一定(ding)時(shi)間或邏(luo)輯順序定(ding)時(shi)開/停，然而(er)目前(qian)我(wo)國(guo)生產(chan)的(de)閥(fa)門質量存在一些問(wen)題，使(shi)用(yong)壽命較短，如(ru)果從國(guo)外進(jin)口價(jia)格(ge)又很(hen)昂貴，一般污(wu)水處理廠(chang)(chang)很(hen)難承(cheng)受，因(yin)此筆者認為制(zhi)約我(wo)國(guo)污(wu)水處理自動(dong)控制(zhi)發展的(de)主要(yao)(yao)原因(yin)不是生產(chan)工藝問(wen)題，而(er)是設備(bei)問(wen)題。

2、智能控制(zhi)技術應用與(yu)發展

作為(wei)智能(neng)控(kong)(kong)制(zhi)重要(yao)分支的(de)(de)(de)(de)模(mo)糊(hu)控(kong)(kong)制(zhi)、神經(jing)網絡控(kong)(kong)制(zhi)、專家控(kong)(kong)制(zhi)和自學(xue)習控(kong)(kong)制(zhi)等除了應(ying)用(yong)(yong)到工業過程控(kong)(kong)制(zhi)以(yi)外，已經(jing)擴大到軍事、醫(yi)學(xue)、高科技領域。由于(yu)智能(neng)控(kong)(kong)制(zhi)系(xi)統具有(you)(you)自學(xue)習、自適(shi)應(ying)和自組織功(gong)能(neng)，特別適(shi)用(yong)(yong)于(yu)復雜的(de)(de)(de)(de)污(wu)水處(chu)理(li)動(dong)態(tai)過程的(de)(de)(de)(de)控(kong)(kong)制(zhi)，因此近年來智能(neng)控(kong)(kong)制(zhi)在(zai)美國、歐洲、日本的(de)(de)(de)(de)給水處(chu)理(li)、污(wu)水生物(wu)處(chu)理(li)、污(wu)水的(de)(de)(de)(de)物(wu)理(li)化學(xue)處(chu)理(li)中都有(you)(you)典型的(de)(de)(de)(de)成(cheng)功(gong)應(ying)用(yong)(yong)，正(zheng)在(zai)研究(jiu)與開發的(de)(de)(de)(de)更是不勝枚舉(ju)。從現在(zai)可以(yi)檢(jian)索(suo)到的(de)(de)(de)(de)有(you)(you)關污(wu)水處(chu)理(li)自動(dong)控(kong)(kong)制(zhi)的(de)(de)(de)(de)研究(jiu)論文來看，有(you)(you)近1/3的(de)(de)(de)(de)論文涉及到智能(neng)控(kong)(kong)制(zhi)，可見智能(neng)控(kong)(kong)制(zhi)已成(cheng)為(wei)該領域的(de)(de)(de)(de)一個研究(jiu)熱點與前沿課題，顯示出極為(wei)廣闊(kuo)的(de)(de)(de)(de)應(ying)用(yong)(yong)前景。

2.1 模糊控制

模(mo)糊(hu)控(kong)(kong)(kong)制(zhi)(zhi)(zhi)(Fuzzy Control)能將操作者(zhe)或專家的(de)控(kong)(kong)(kong)制(zhi)(zhi)(zhi)經驗和(he)知識表示成(cheng)語(yu)言變量(liang)描述的(de)控(kong)(kong)(kong)制(zhi)(zhi)(zhi)規(gui)(gui)則(ze)，然后用這(zhe)些規(gui)(gui)則(ze)去控(kong)(kong)(kong)制(zhi)(zhi)(zhi)系(xi)統(tong)。因此，模(mo)糊(hu)控(kong)(kong)(kong)制(zhi)(zhi)(zhi)特別適(shi)用于數學模(mo)型未(wei)知的(de)、復雜的(de)非線(xian)性系(xi)統(tong)的(de)控(kong)(kong)(kong)制(zhi)(zhi)(zhi)。正(zheng)是基于模(mo)糊(hu)控(kong)(kong)(kong)制(zhi)(zhi)(zhi)這(zhe)些特點(dian)，近年來(lai)它已(yi)成(cheng)為(wei)污水(shui)處理系(xi)統(tong)的(de)研究(jiu)熱點(dian)。

1980年Tong等(deng)(deng)首次將模糊(hu)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)應(ying)用到(dao)污(wu)水處(chu)理中，將出(chu)(chu)水BOD、SS、曝(pu)氣(qi)(qi)池MLSS、DO及(ji)出(chu)(chu)水氨氮濃(nong)度(du)(du)(du)、回流污(wu)泥(ni)(ni)量(liang)等(deng)(deng)監(jian)測數據(ju)(ju)作(zuo)為輸(shu)入(ru)變(bian)量(liang)輸(shu)入(ru)該系統，“模糊(hu)化”以后(hou)再與“規則(ze)集”進行(xing)匹配(pei)，隨后(hou)確(que)定(ding)(ding)相(xiang)應(ying)的(de)(de)(de)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)手(shou)段，最后(hou)通過反模糊(hu)化得到(dao)量(liang)化的(de)(de)(de)具(ju)體信號(hao)來實(shi)施(shi)(shi)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)。Flanagan利用Olsson等(deng)(deng)提出(chu)(chu)的(de)(de)(de)曝(pu)氣(qi)(qi)池DO控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)技術，以沿池長的(de)(de)(de)DO濃(nong)度(du)(du)(du)變(bian)化曲(qu)(qu)線來估(gu)計曝(pu)氣(qi)(qi)池中底物利用效率和(he)微(wei)生(sheng)物活(huo)性(xing)。他的(de)(de)(de)知(zhi)(zhi)識(shi)(shi)庫中的(de)(de)(de)知(zhi)(zhi)識(shi)(shi)不(bu)僅有根據(ju)(ju)工(gong)藝狀態(tai)(tai)確(que)定(ding)(ding)采用何種控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)措施(shi)(shi)這一類(lei)啟(qi)發性(xing)規則(ze)，而(er)且還有DO曲(qu)(qu)線特(te)征及(ji)相(xiang)關工(gong)藝狀態(tai)(tai)方面的(de)(de)(de)知(zhi)(zhi)識(shi)(shi)。Barnett把這些知(zhi)(zhi)識(shi)(shi)稱為“匯(hui)編(bian)知(zhi)(zhi)識(shi)(shi)”(compiled knowledge)，“匯(hui)編(bian)知(zhi)(zhi)識(shi)(shi)”作(zuo)為啟(qi)發性(xing)知(zhi)(zhi)識(shi)(shi)的(de)(de)(de)補充，提高了(le)系統解決(jue)問題的(de)(de)(de)深度(du)(du)(du)和(he)廣度(du)(du)(du)[13]。由于(yu)活(huo)性(xing)污(wu)泥(ni)(ni)法出(chu)(chu)水BOD或COD濃(nong)度(du)(du)(du)通常隨出(chu)(chu)水懸浮物濃(nong)度(du)(du)(du)增加而(er)增大，因此Tsai等(deng)(deng)人建立了(le)對出(chu)(chu)水懸浮物濃(nong)度(du)(du)(du)進行(xing)預測和(he)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)的(de)(de)(de)動態(tai)(tai)活(huo)性(xing)污(wu)泥(ni)(ni)法模糊(hu)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)，他們所(suo)提出(chu)(chu)的(de)(de)(de)模糊(hu)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)(zhi)策略能有效地降低出(chu)(chu)水SS濃(nong)度(du)(du)(du)，從而(er)使處(chu)理系統的(de)(de)(de)運行(xing)穩定(ding)(ding)可(ke)靠。

與常(chang)(chang)規(gui)活性(xing)污泥法相比，高純氧活性(xing)污泥法(High Purity Oxygen Activated Sludge，HPO—AS)對控制(zhi)(zhi)的要求更(geng)加嚴格(ge)。由于過程滯(zhi)后和噪聲(sheng)干擾，此(ci)系統(tong)兩種(zhong)常(chang)(chang)規(gui)反(fan)饋(kui)(kui)控制(zhi)(zhi)在(zai)控制(zhi)(zhi)過程中經(jing)常(chang)(chang)出(chu)現問題。為此(ci)Yin等人(ren)研(yan)究了四種(zhong)模(mo)(mo)糊邏(luo)輯(ji)控制(zhi)(zhi)系統(tong)，結果(guo)表明(ming)在(zai)正常(chang)(chang)條(tiao)件下，模(mo)(mo)糊控制(zhi)(zhi)比常(chang)(chang)規(gui)的反(fan)饋(kui)(kui)控制(zhi)(zhi)更(geng)加節約能源、減少DO波(bo)動、穩(wen)定進水流量和出(chu)氣流速。

Manesis等人對一個(ge)前置反(fan)硝化污水處理(li)廠(chang)進行(xing)了模糊控(kong)制(zhi)系統(tong)研(yan)究。他(ta)們(men)以反(fan)應(ying)器中氨(an)氮(dan)、硝態氮(dan)、DO、溫度、MLSS和(he)二沉池進出水BOD的(de)(de)差值作(zuo)為(wei)模糊控(kong)制(zhi)系統(tong)的(de)(de)輸入變量，以曝氣區供氧速(su)率(lv)、好氧區向缺氧區的(de)(de)回(hui)流速(su)率(lv)以及二沉池向反(fan)應(ying)器的(de)(de)污泥回(hui)流速(su)率(lv)作(zuo)為(wei)輸出變量，以處理(li)廠(chang)操(cao)作(zuo)人員的(de)(de)經驗建立模糊控(kong)制(zhi)規則，并在希臘Patras污水處理(li)廠(chang)進行(xing)了仿(fang)真，取得了較好的(de)(de)結果。

與(yu)國(guo)外(wai)相(xiang)比，國(guo)內從事污水(shui)處(chu)理模(mo)(mo)糊(hu)(hu)控制(zhi)的(de)研究人員較少。彭永臻等對(dui)硝態(tai)氮污染水(shui)脫氮處(chu)理的(de)新方(fang)法——生(sheng)物電極法采用模(mo)(mo)糊(hu)(hu)控制(zhi)，也取(qu)得了較好的(de)控制(zhi)效(xiao)果。這種在線模(mo)(mo)糊(hu)(hu)控制(zhi)器(qi)具有(you)(you)(you)構造簡(jian)單、可行性(xing)好、可靠性(xing)高、穩定性(xing)好和對(dui)進水(shui)硝態(tai)氮負荷變化(hua)(hua)的(de)適應(ying)性(xing)強等優點，有(you)(you)(you)利于避免過量地投加有(you)(you)(you)機(ji)物并盡(jin)可能節省運行費(fei)用。彭永臻、曾薇(wei)等[19]采用SBR法處(chu)理石油化(hua)(hua)工廢(fei)水(shui)，根據反應(ying)器(qi)內有(you)(you)(you)機(ji)物的(de)去除(chu)與(yu)DO濃(nong)度的(de)相(xiang)關性(xing)，提出以DO作為SBR法的(de)模(mo)(mo)糊(hu)(hu)控制(zhi)參(can)數。通過大量試(shi)驗，認為可根據初始階(jie)段DO濃(nong)度及變化(hua)(hua)情(qing)況預測進水(shui)有(you)(you)(you)機(ji)物濃(nong)度，進而實現對(dui)曝氣量的(de)模(mo)(mo)糊(hu)(hu)控制(zhi)。

2.2 神(shen)經網絡(luo)控制

基于人(ren)工(gong)神(shen)經(jing)(jing)網(wang)(wang)絡的控制(zhi)(zhi)(ANN—based Control)簡稱神(shen)經(jing)(jing)控制(zhi)(zhi)(Neural Control)。神(shen)經(jing)(jing)網(wang)(wang)絡是(shi)由大量人(ren)工(gong)神(shen)經(jing)(jing)元廣(guang)泛聯結(jie)而(er)成的網(wang)(wang)絡，它具有很(hen)強的自適(shi)應(ying)性(xing)和學習能(neng)(neng)力(li)、非線(xian)性(xing)映射(she)能(neng)(neng)力(li)和容錯能(neng)(neng)力(li)。神(shen)經(jing)(jing)網(wang)(wang)絡因具備(bei)上述特點(dian)，近年來越來越受(shou)到國內(nei)外污水(shui)處理專(zhuan)家的重(zhong)視，并(bing)在(zai)污水(shui)處理自動控制(zhi)(zhi)系統(tong)中(zhong)開展人(ren)工(gong)神(shen)經(jing)(jing)網(wang)(wang)絡控制(zhi)(zhi)研究，取(qu)得了許多具有推廣(guang)應(ying)用價(jia)值的成果。

Zhu等研究開(kai)發了一(yi)(yi)種(zhong)基于(yu)時間延遲神(shen)經(jing)網(wang)(wang)絡(luo)模(mo)(mo)型的(de)(de)在線廢水(shui)(shui)(shui)水(shui)(shui)(shui)質(zhi)(zhi)預測系統。他們首先(xian)提出采用多層感知器(MLP)網(wang)(wang)絡(luo)模(mo)(mo)型對所建立的(de)(de)時間延遲神(shen)經(jing)網(wang)(wang)絡(luo)(TDNN)的(de)(de)輸入節點進行篩(shai)選(xuan)，最后得(de)到一(yi)(yi)個(ge)10輸入TDNN模(mo)(mo)型，網(wang)(wang)絡(luo)經(jing)過訓練以后，其對廢水(shui)(shui)(shui)處理(li)預測精(jing)度均優于(yu)標(biao)準MLP模(mo)(mo)型。Gontarski等應(ying)用BP算法人工(gong)(gong)神(shen)經(jing)網(wang)(wang)絡(luo)預測一(yi)(yi)個(ge)工(gong)(gong)業廢水(shui)(shui)(shui)處理(li)廠(chang)的(de)(de)出水(shui)(shui)(shui)水(shui)(shui)(shui)質(zhi)(zhi)，試(shi)驗中共(gong)使用了7個(ge)神(shen)經(jing)網(wang)(wang)絡(luo)，每一(yi)(yi)個(ge)反應(ying)器用一(yi)(yi)個(ge)神(shen)經(jing)網(wang)(wang)絡(luo)，最后一(yi)(yi)個(ge)神(shen)經(jing)網(wang)(wang)絡(luo)用來預測出水(shui)(shui)(shui)TOC的(de)(de)變化。試(shi)驗結果表(biao)明，廢水(shui)(shui)(shui)的(de)(de)流(liu)量和(he)進水(shui)(shui)(shui)pH值是(shi)廢水(shui)(shui)(shui)處理(li)廠(chang)重要(yao)的(de)(de)控制(zhi)參數。

在(zai)(zai)活性(xing)污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)法(fa)污(wu)(wu)(wu)(wu)(wu)水(shui)(shui)處(chu)理(li)廠中，污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)是引起運行不正常(chang)的(de)(de)(de)(de)(de)一(yi)個(ge)嚴重問題(ti)，它(ta)直接(jie)影(ying)響(xiang)(xiang)污(wu)(wu)(wu)(wu)(wu)水(shui)(shui)處(chu)理(li)廠的(de)(de)(de)(de)(de)處(chu)理(li)效率，因此許多學者從活性(xing)污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)法(fa)的(de)(de)(de)(de)(de)運行機理(li)上(shang)對(dui)污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)現(xian)象進(jin)行了(le)廣泛的(de)(de)(de)(de)(de)研究，但至(zhi)今尚未獲得克服(fu)污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)的(de)(de)(de)(de)(de)經濟而有效的(de)(de)(de)(de)(de)方(fang)法(fa)。近年來，國外一(yi)些(xie)(xie)(xie)學者采(cai)用人(ren)工神(shen)經網絡技術建立模(mo)型(xing)(xing)來預(yu)(yu)測(ce)和(he)(he)防止污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)現(xian)象的(de)(de)(de)(de)(de)發生。Capodaglio 等在(zai)(zai)分析活性(xing)污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)系統輸(shu)入和(he)(he)輸(shu)出的(de)(de)(de)(de)(de)基礎(chu)(chu)上(shang)，應用污(wu)(wu)(wu)(wu)(wu)水(shui)(shui)處(chu)理(li)廠的(de)(de)(de)(de)(de)數據建立了(le)人(ren)工神(shen)經網絡模(mo)型(xing)(xing)，隨后(hou)用這(zhe)種模(mo)型(xing)(xing)預(yu)(yu)測(ce)未來污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)的(de)(de)(de)(de)(de)發生。為使所構(gou)建的(de)(de)(de)(de)(de)模(mo)型(xing)(xing)能更好(hao)地反映活性(xing)污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)法(fa)的(de)(de)(de)(de)(de)實際狀況，他們為輸(shu)入參數選擇(ze)了(le)一(yi)個(ge)時(shi)間滯后(hou)輸(shu)入方(fang)案。從模(mo)型(xing)(xing)預(yu)(yu)測(ce)結(jie)(jie)果可(ke)以看出，這(zhe)種模(mo)型(xing)(xing)的(de)(de)(de)(de)(de)預(yu)(yu)測(ce)精度遠遠超過其他傳統預(yu)(yu)測(ce)方(fang)法(fa)。Belanche等在(zai)(zai)Capodaglio建立的(de)(de)(de)(de)(de)模(mo)型(xing)(xing)基礎(chu)(chu)上(shang)引入定性(xing)信(xin)息(xi)，這(zhe)些(xie)(xie)(xie)定性(xing)信(xin)息(xi)主(zhu)要來源于(yu)顯微鏡對(dui)細菌和(he)(he)微型(xing)(xing)動物的(de)(de)(de)(de)(de)觀察和(he)(he)一(yi)些(xie)(xie)(xie)主(zhu)觀經驗，并利用該模(mo)型(xing)(xing)對(dui)廢水(shui)(shui)處(chu)理(li)廠污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)現(xian)象進(jin)行預(yu)(yu)測(ce)。試驗結(jie)(jie)果顯示，定性(xing)信(xin)息(xi)對(dui)處(chu)理(li)廠污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)現(xian)象影(ying)響(xiang)(xiang)很(hen)大，模(mo)型(xing)(xing)對(dui)污(wu)(wu)(wu)(wu)(wu)泥(ni)(ni)膨(peng)(peng)脹(zhang)(zhang)(zhang)的(de)(de)(de)(de)(de)預(yu)(yu)測(ce)同污(wu)(wu)(wu)(wu)(wu)水(shui)(shui)處(chu)理(li)廠專家的(de)(de)(de)(de)(de)評價判(pan)斷吻合得很(hen)好(hao)。

Tay等人(ren)在一(yi)個(ge)神經(jing)模糊(hu)模型(xing)的基(ji)礎上，為污水厭氧處理系(xi)統(tong)(tong)開發(fa)出一(yi)個(ge)快速(su)預測神經(jing)模糊(hu)模型(xing)來預測高(gao)速(su)率厭氧系(xi)統(tong)(tong)對干擾(rao)的響(xiang)應，此系(xi)統(tong)(tong)可以提前1h對不同系(xi)統(tong)(tong)的干擾(rao)進行(xing)預測。因此，該系(xi)統(tong)(tong)在實時控制上有很(hen)大(da)的應用潛力。

Wen等人研(yan)究了(le)一(yi)種曝(pu)氣(qi)池神(shen)經(jing)(jing)網(wang)絡模型。該(gai)曝(pu)氣(qi)池神(shen)經(jing)(jing)網(wang)絡模型的(de)(de)(de)(de)(de)數(shu)據(ju)由一(yi)個(ge)專家(jia)(jia)系(xi)統(tong)(tong)來(lai)提(ti)供，專家(jia)(jia)系(xi)統(tong)(tong)又從神(shen)經(jing)(jing)網(wang)絡模型獲取其所要的(de)(de)(de)(de)(de)數(shu)據(ju)，從而對整個(ge)污(wu)水處理廠實施(shi)智(zhi)能控(kong)制。專家(jia)(jia)系(xi)統(tong)(tong)從各(ge)種傳(chuan)感器(qi)和檢測(ce)(ce)器(qi)獲得信號后(hou)檢查(cha)系(xi)統(tong)(tong)的(de)(de)(de)(de)(de)狀(zhuang)(zhuang)態(tai)(tai)，推斷(duan)(duan)出(chu)一(yi)個(ge)污(wu)泥回流(liu)(liu)比(bi)(bi)(bi)。然后(hou)，專家(jia)(jia)系(xi)統(tong)(tong)把這個(ge)值(zhi)(zhi)(zhi)送給神(shen)經(jing)(jing)網(wang)絡，神(shen)經(jing)(jing)網(wang)絡把從專家(jia)(jia)系(xi)統(tong)(tong)獲得的(de)(de)(de)(de)(de)當前(qian)狀(zhuang)(zhuang)態(tai)(tai)值(zhi)(zhi)(zhi)與(yu)通(tong)過網(wang)絡預(yu)測(ce)(ce)得到的(de)(de)(de)(de)(de)值(zhi)(zhi)(zhi)進(jin)行(xing)(xing)比(bi)(bi)(bi)較，分析(xi)該(gai)值(zhi)(zhi)(zhi)是(shi)(shi)增加還是(shi)(shi)減小(xiao)或者是(shi)(shi)維持不變。專家(jia)(jia)系(xi)統(tong)(tong)根據(ju)當前(qian)BOD和MLSS的(de)(de)(de)(de)(de)值(zhi)(zhi)(zhi)以及神(shen)經(jing)(jing)網(wang)絡預(yu)測(ce)(ce)的(de)(de)(de)(de)(de)曝(pu)氣(qi)池狀(zhuang)(zhuang)態(tai)(tai)判斷(duan)(duan)是(shi)(shi)否采用這個(ge)污(wu)泥回流(liu)(liu)比(bi)(bi)(bi)。如果預(yu)測(ce)(ce)狀(zhuang)(zhuang)態(tai)(tai)不是(shi)(shi)所期望的(de)(de)(de)(de)(de)，那么專家(jia)(jia)系(xi)統(tong)(tong)將再給出(chu)一(yi)個(ge)污(wu)泥回流(liu)(liu)比(bi)(bi)(bi)，重新進(jin)行(xing)(xing)一(yi)次測(ce)(ce)試，直到找到合理的(de)(de)(de)(de)(de)污(wu)泥回流(liu)(liu)比(bi)(bi)(bi)。若(ruo)專家(jia)(jia)系(xi)統(tong)(tong)想(xiang)增加曝(pu)氣(qi)池內的(de)(de)(de)(de)(de)BOD濃(nong)度(du)，它在向神(shen)經(jing)(jing)網(wang)絡模型傳(chuan)輸這組數(shu)據(ju)時，可(ke)以在當前(qian)BOD濃(nong)度(du)上加一(yi)個(ge)小(xiao)小(xiao)的(de)(de)(de)(de)(de)增量(例如0.05)作為目(mu)標值(zhi)(zhi)(zhi)，神(shen)經(jing)(jing)網(wang)絡模型就(jiu)以此值(zhi)(zhi)(zhi)預(yu)測(ce)(ce)一(yi)個(ge)污(wu)泥回流(liu)(liu)比(bi)(bi)(bi)，并把它反饋給專家(jia)(jia)系(xi)統(tong)(tong)。

2.3 專家控制

專(zhuan)家(jia)控(kong)(kong)制(zhi)(Expert Control)是(shi)智(zhi)能(neng)(neng)(neng)控(kong)(kong)制(zhi)的(de)一個重要分支(zhi)，又(you)稱專(zhuan)家(jia)智(zhi)能(neng)(neng)(neng)控(kong)(kong)制(zhi)。所謂專(zhuan)家(jia)控(kong)(kong)制(zhi)，是(shi)把(ba)專(zhuan)家(jia)系統的(de)理論和技術(shu)同控(kong)(kong)制(zhi)理論、方法與(yu)技術(shu)相結(jie)合，在未知環境(jing)下仿效專(zhuan)家(jia)的(de)智(zhi)能(neng)(neng)(neng)，實現(xian)對系統的(de)控(kong)(kong)制(zhi)。20世紀90年代國外就有(you)學者開始研究(jiu)采(cai)用專(zhuan)家(jia)系統智(zhi)能(neng)(neng)(neng)控(kong)(kong)制(zhi)技術(shu)來實現(xian)污水處理的(de)自(zi)動(dong)控(kong)(kong)制(zhi)，并取得了有(you)效成果。

Barnett建立了(le)一(yi)(yi)個(ge)基于規則的(de)(de)專(zhuan)家系(xi)統(tong)，用于污(wu)泥厭(yan)氧(yang)消(xiao)化(hua)(hua)的(de)(de)故障診斷。整個(ge)過程由(you)(you)計(ji)算機進行模擬(ni)，過程變量(liang)包括消(xiao)化(hua)(hua)池(chi)的(de)(de)輸入輸出及表征池(chi)內狀態(tai)(tai)的(de)(de)9個(ge)參數，控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)變量(liang)是進泥量(liang)、回流污(wu)泥量(liang)、稀釋水量(liang)和(he)調(diao)節(jie)pH值的(de)(de)酸堿投量(liang)。另外，研究者(zhe)(zhe)為專(zhuan)家系(xi)統(tong)界(jie)定5類(lei)消(xiao)化(hua)(hua)工藝運行不(bu)正(zheng)常(chang)狀態(tai)(tai)，每類(lei)狀態(tai)(tai)又細(xi)分為注意、警告、危急(ji)和(he)恢復(fu)(fu)正(zheng)常(chang)等(deng)幾類(lei)亞狀態(tai)(tai)。這些狀態(tai)(tai)和(he)亞狀態(tai)(tai)再與相關的(de)(de)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)措(cuo)施(shi)相對(dui)應(ying)，即(ji)不(bu)正(zheng)常(chang)狀態(tai)(tai)的(de)(de)類(lei)型和(he)程度決定了(le)該采取什么(me)樣(yang)的(de)(de)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)手段，以便使消(xiao)化(hua)(hua)恢復(fu)(fu)正(zheng)常(chang)。Flores 等(deng)設計(ji)了(le)一(yi)(yi)個(ge)智(zhi)能(neng)化(hua)(hua)系(xi)統(tong)來運行和(he)管理多級厭(yan)氧(yang)系(xi)統(tong)，這個(ge)厭(yan)氧(yang)系(xi)統(tong)由(you)(you)各自的(de)(de)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)器(qi)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)，而(er)這些控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)器(qi)又通過寬帶網與遠處的(de)(de)中(zhong)(zhong)央(yang)管理器(qi)相聯。中(zhong)(zhong)央(yang)管理器(qi)采集、分析(xi)、解釋和(he)存(cun)儲由(you)(you)各生(sheng)物(wu)反應(ying)器(qi)控(kong)(kong)制(zhi)(zhi)(zhi)(zhi)系(xi)統(tong)傳來的(de)(de)數據(ju)，并采用圖形界(jie)面的(de)(de)形式使操作者(zhe)(zhe)能(neng)清楚地看到這些信息。

Sung等采用(yong)在線綜合(he)控制(zhi)系統(tong)對水(shui)質、水(shui)量變化較大的(de)食品廢水(shui)進行控制(zhi)。控制(zhi)目標是使出水(shui)COD濃(nong)(nong)度較地(di)方標準低50%，并且盡量減少曝氣費用(yong)。此(ci)控制(zhi)系統(tong)由(you)兩(liang)層(ceng)(ceng)組成，即(ji)管理層(ceng)(ceng)和過程控制(zhi)層(ceng)(ceng)。在管理層(ceng)(ceng)中應用(yong)基于(yu)規則的(de)專家系統(tong)為過程控制(zhi)層(ceng)(ceng)提供(gong)最優控制(zhi)點。此(ci)外，為避(bi)免鼓風機超負(fu)荷(he)運行，還設計了基于(yu)規則的(de)負(fu)荷(he)分配系統(tong)。此(ci)控制(zhi)系統(tong)已經成功地(di)運行了2年，與不實行控制(zhi)之前相比，出水(shui)COD濃(nong)(nong)度降低大約(yue)50%，節(jie)能約(yue)50%。

通過(guo)以(yi)上分析可知，智(zhi)能控(kong)(kong)(kong)(kong)制(zhi)技術在(zai)(zai)污(wu)(wu)水處(chu)(chu)理(li)(li)中應用較晚(只是近20年才逐(zhu)漸得到應用)，而且大多(duo)數(shu)(shu)仍(reng)停(ting)留在(zai)(zai)實驗室研(yan)究(jiu)階(jie)段(duan)，很多(duo)地(di)方(fang)還很不完(wan)善。以(yi)神經網(wang)絡控(kong)(kong)(kong)(kong)制(zhi)為(wei)例(li)，目前研(yan)究(jiu)較多(duo)的(de)模型(xing)(xing)屬于靜態模型(xing)(xing)，在(zai)(zai)一定程度(du)上不太適合污(wu)(wu)水處(chu)(chu)理(li)(li)在(zai)(zai)線控(kong)(kong)(kong)(kong)制(zhi)，因為(wei)活性污(wu)(wu)泥法污(wu)(wu)水處(chu)(chu)理(li)(li)隨時(shi)間(jian)變化(hua)較大而且具(ju)有(you)較大滯后性。因此(ci)，建議從事污(wu)(wu)水處(chu)(chu)理(li)(li)智(zhi)能控(kong)(kong)(kong)(kong)制(zhi)的(de)科(ke)研(yan)人(ren)員(yuan)以(yi)實際污(wu)(wu)水處(chu)(chu)理(li)(li)廠(chang)為(wei)研(yan)究(jiu)目標(biao)，找出各種控(kong)(kong)(kong)(kong)制(zhi)參數(shu)(shu)隨時(shi)間(jian)的(de)變化(hua)規(gui)律，運用動態模型(xing)(xing)建立污(wu)(wu)水處(chu)(chu)理(li)(li)智(zhi)能控(kong)(kong)(kong)(kong)制(zhi)系(xi)統(tong)。

3、結語

① 雖然智能控(kong)(kong)制(zhi)已成為污水(shui)處(chu)理的研究(jiu)與應(ying)用中的前沿與熱點，但國(guo)內外仍處(chu)于廣泛應(ying)用的初級階段。從文獻來看，我國(guo)從事(shi)這方面研究(jiu)的人員(yuan)太少，這也(ye)是制(zhi)約我國(guo)污水(shui)處(chu)理自(zi)動控(kong)(kong)制(zhi)和智能控(kong)(kong)制(zhi)發展的主要因素。

② 與(yu)發達(da)國家(jia)相(xiang)比，我(wo)國在(zai)污(wu)(wu)水(shui)處理的(de)基本理論、工藝流程和(he)工程設計等方面(mian)并(bing)不(bu)明顯(xian)落后，但是(shi)在(zai)運(yun)行管(guan)理與(yu)自動控制(zhi)方面(mian)卻(que)存在(zai)著較大的(de)差距。目前，我(wo)國城市(shi)污(wu)(wu)水(shui)處理廠(chang)的(de)噸水(shui)耗電量是(shi)發達(da)國家(jia)的(de)近兩倍，而運(yun)行管(guan)理人員數又是(shi)其若干倍，因此加強我(wo)國污(wu)(wu)水(shui)處理系(xi)統智能控制(zhi)的(de)研究與(yu)應用具有重(zhong)要的(de)科學意義與(yu)應用價(jia)值。

③ 由于智能控制的優越性及其研究與應用的迅速發展，目前國外許多城市污水和工業廢水處理廠正在通過技術改造向實現智能控制方向過渡。我國應當在有條件的情況下，在污水處理廠的規劃、設計與建設初期就盡可能采用或部分采用智能控制。
 

  使用微信“掃一掃”功能添加“谷騰環保網”

相關文章