1.前言

    蓄電(dian)池的(de)(de)充(chong)(chong)電(dian)控(kong)制系統(tong)是個非線性[1]的(de)(de)、時變的(de)(de)、有干擾的(de)(de)、具有純(chun)滯后的(de)(de)控(kong)制系統(tong)[2]。在(zai)充(chong)(chong)放(fang)電(dian)過程中涉及(ji)到很多(duo)參數，如充(chong)(chong)電(dian)率、最大允許(xu)充(chong)(chong)電(dian)電(dian)流、內(nei)阻、出氣點電(dian)壓(ya)、溫度、壽命(ming)等。

   傳統(tong)(tong)的控(kong)制系統(tong)(tong)是建立在被控(kong)對象(xiang)精確(que)數(shu)(shu)學模(mo)型(xing)基(ji)礎上的，如果被控(kong)對象(xiang)的數(shu)(shu)學模(mo)型(xing)很(hen)復雜或者(zhe)數(shu)(shu)學模(mo)型(xing)無(wu)法建立，控(kong)制系統(tong)(tong)就較(jiao)難(nan)實現(xian)。蓄電(dian)(dian)池充(chong)電(dian)(dian)正是屬于這種情況，由于蓄電(dian)(dian)池的充(chong)電(dian)(dian)過程有自己獨特的物(wu)理化學規(gui)律，因此考慮采用模(mo)糊控(kong)制技術來進行蓄電(dian)(dian)池的充(chong)電(dian)(dian)控(kong)制 [3]。

   模糊控(kong)制[4]是(shi)以模糊集(ji)合(he)理(li)論(lun)為基礎(chu)的(de)(de)控(kong)制手(shou)段，它是(shi)模糊系(xi)統理(li)論(lun)、模糊技(ji)術與自(zi)動控(kong)制技(ji)術相結合(he)的(de)(de)產物,出發點是(shi)操作(zuo)人員(yuan)的(de)(de)控(kong)制經驗或相關專家的(de)(de)知識，在設(she)計中不需要建立(li)被控(kong)對(dui)象的(de)(de)精確數學模型[5]，因而使得(de)控(kong)制機理(li)和策略易于(yu)(yu)接受與理(li)解，設(she)計簡(jian)單，便于(yu)(yu)應用。

2.模糊控制器的結構及算(suan)法

   基于微(wei)控(kong)制器(qi)組(zu)成的模糊控(kong)制器(qi)包括模糊化(hua)接口、決策邏(luo)輯、知識(shi)庫和反模糊化(hua)接口四個(ge)部分，如(ru)圖1所示。

圖1模糊控制器的結構框圖

Fig.1 The structure diagram of fuzzy controller

   在(zai)進行模糊控(kong)制(zhi)(zhi)(zhi)算法(fa)(fa)設計(ji)時，須(xu)首先(xian)考慮智(zhi)能(neng)(neng)充(chong)電系統的(de)(de)(de)技(ji)術(shu)要求，體現(xian)智(zhi)能(neng)(neng)充(chong)電控(kong)制(zhi)(zhi)(zhi)技(ji)術(shu)的(de)(de)(de)優勢，解決長期困擾蓄電池裝備中的(de)(de)(de)效率(lv)和壽(shou)命問題(ti)，所以(yi)應(ying)當提高(gao)智(zhi)能(neng)(neng)充(chong)電控(kong)制(zhi)(zhi)(zhi)中對(dui)控(kong)制(zhi)(zhi)(zhi)量(liang)的(de)(de)(de)精度；其次，模糊控(kong)制(zhi)(zhi)(zhi)算法(fa)(fa)必須(xu)高(gao)速可靠，對(dui)外(wai)部檢測得(de)到的(de)(de)(de)物理(li)量(liang)要有非常快的(de)(de)(de)反應(ying)速度。采用新型大(da)存(cun)(cun)儲容量(liang)的(de)(de)(de)微控(kong)制(zhi)(zhi)(zhi)器解決使用查表法(fa)(fa)所帶來的(de)(de)(de)需(xu)要大(da)容量(liang)內存(cun)(cun)的(de)(de)(de)問題(ti)。

3 模糊控制器(qi)的(de)設計

3.1雙輸(shu)(shu)入單輸(shu)(shu)出模(mo)糊控制(zhi)器

    常見的模糊控制器[6]根據輸(shu)(shu)(shu)入與輸(shu)(shu)(shu)出(chu)個數分為單(dan)輸(shu)(shu)(shu)入單(dan)輸(shu)(shu)(shu)出(chu)、雙輸(shu)(shu)(shu)入單(dan)輸(shu)(shu)(shu)出(chu)、多輸(shu)(shu)(shu)入單(dan)輸(shu)(shu)(shu)出(chu)和(he)雙輸(shu)(shu)(shu)入多輸(shu)(shu)(shu)出(chu)等(deng)幾種(zhong)。平時(shi)應用較多的是雙輸(shu)(shu)(shu)入單(dan)輸(shu)(shu)(shu)出(chu)模糊控制器。

圖2 雙輸入單輸出模糊控制器(qi)方(fang)框(kuang)圖

 Fig.2 The rectangular diagram of fuzzy cONtroller witch has two input units and one output unit

    圖2是雙輸(shu)入單輸(shu)出模(mo)(mo)(mo)糊(hu)(hu)控(kong)制器(qi)的(de)方(fang)框圖。其中屬于論(lun)域X的(de)模(mo)(mo)(mo)糊(hu)(hu)集(ji)合(he) 取自(zi)系(xi)統誤(wu)差(cha)e的(de)模(mo)(mo)(mo)糊(hu)(hu)化，屬于論(lun)域Y的(de)模(mo)(mo)(mo)糊(hu)(hu)集(ji)合(he) 取自(zi)系(xi)統誤(wu)差(cha)變化率 的(de)模(mo)(mo)(mo)糊(hu)(hu)化，二(er)者構成(cheng)模(mo)(mo)(mo)糊(hu)(hu)控(kong)制器(qi)的(de)二(er)維輸(shu)入；屬于論(lun)域Z的(de)模(mo)(mo)(mo)糊(hu)(hu)集(ji)合(he) 是反映控(kong)制量變化的(de)模(mo)(mo)(mo)糊(hu)(hu)控(kong)制器(qi)的(de)一維輸(shu)出，模(mo)(mo)(mo)糊(hu)(hu)控(kong)制器(qi)的(de)控(kong)制規則[7]通常由模(mo)(mo)(mo)糊(hu)(hu)條件語句if    and    then  來(lai)表達。

3.2精確量的模糊化

   模(mo)糊控(kong)(kong)(kong)制(zhi)系(xi)統中的(de)(de)被控(kong)(kong)(kong)對象狀態變化(hua)是(shi)(shi)(shi)(shi)連續(xu)(xu)的(de)(de)，系(xi)統給(gei)定值(zhi)也是(shi)(shi)(shi)(shi)連續(xu)(xu)的(de)(de),反(fan)映到模(mo)糊控(kong)(kong)(kong)制(zhi)器(qi)輸入端(duan)的(de)(de)輸入量(liang)也必然是(shi)(shi)(shi)(shi)連續(xu)(xu)的(de)(de)。但(dan)模(mo)糊控(kong)(kong)(kong)制(zhi)器(qi)由(you)計算機構成，它只能執(zhi)行離散處理[7]。因此模(mo)糊控(kong)(kong)(kong)制(zhi)器(qi)要求輸入量(liang)是(shi)(shi)(shi)(shi)離散模(mo)糊量(liang)，即(ji)論(lun)域是(shi)(shi)(shi)(shi)離散的(de)(de)。對連續(xu)(xu)論(lun)域要進行離散化(hua)。連續(xu)(xu)論(lun)域[8]經過量(liang)化(hua)后就成為一個離散論(lun)域[8]。

設有連(lian)續論(lun)域[a，b]，而量(liang)化之后的(de)離散論(lun)域為 ，亦(yi)即(ji)將連(lian)續論(lun)域分為2n段，則存在系(xi)數

                                                                   （1）

   隨后(hou)，在求出(chu)每條規則(ze)的強(qiang)度之后(hou)，把相互矛(mao)盾的規則(ze)中(zhong)強(qiang)度較小的舍(she)去；把相同規則(ze)合(he)成一條規則(ze)，得到最(zui)后(hou)控制規則(ze)基。

3.3.2根據(ju)學(xue)習算法(fa)生成控(kong)制規則

   對(dui)被控(kong)(kong)(kong)(kong)對(dui)象執行手動控(kong)(kong)(kong)(kong)制所得(de)到的控(kong)(kong)(kong)(kong)制規(gui)則是較粗(cu)糙(cao)的，有時還可能會出(chu)現(xian)控(kong)(kong)(kong)(kong)制死區，一個控(kong)(kong)(kong)(kong)制規(gui)則表(biao)中會出(chu)現(xian)空項,這(zhe)(zhe)是不能滿足實際控(kong)(kong)(kong)(kong)制要(yao)求的。為了(le)取(qu)得(de)更滿意(yi)的控(kong)(kong)(kong)(kong)制效果，可以對(dui)原始(shi)的控(kong)(kong)(kong)(kong)制規(gui)則進行改進。這(zhe)(zhe)時，應(ying)以粗(cu)糙(cao)的控(kong)(kong)(kong)(kong)制規(gui)則為基(ji)礎(chu)，通過仿真(zhen)實驗(yan)和系統調試加以完善。

4.模糊(hu)智能(neng)充(chong)電系(xi)統的工作(zuo)原(yuan)理及結(jie)構

   智(zhi)能充(chong)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)系統主要由(you)充(chong)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)源和單片(pian)機(ji)控(kong)制(zhi)(zhi)(zhi)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)路(lu)兩部分組成。220V的(de)交流(liu)(liu)市電(dian)(dian)(dian)(dian)(dian)(dian)(dian)經整流(liu)(liu)濾波電(dian)(dian)(dian)(dian)(dian)(dian)(dian)路(lu)變(bian)(bian)為脈(mo)(mo)動的(de)310V高壓(ya)直流(liu)(liu)。然后(hou)經DC-DC變(bian)(bian)換電(dian)(dian)(dian)(dian)(dian)(dian)(dian)路(lu)（脈(mo)(mo)沖(chong)功率變(bian)(bian)壓(ya)器）變(bian)(bian)為充(chong)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)所需(xu)的(de)60V直流(liu)(liu)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)壓(ya)。為了保證輸出電(dian)(dian)(dian)(dian)(dian)(dian)(dian)壓(ya)的(de)穩(wen)定性，采(cai)用了UC3842對(dui)60V直流(liu)(liu)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)壓(ya)進行穩(wen)壓(ya)。二(er)次斬(zhan)波電(dian)(dian)(dian)(dian)(dian)(dian)(dian)路(lu)主要由(you)MOSFET管、電(dian)(dian)(dian)(dian)(dian)(dian)(dian)感、電(dian)(dian)(dian)(dian)(dian)(dian)(dian)容和二(er)極管組成，輸出24-36V的(de)充(chong)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)壓(ya)。控(kong)制(zhi)(zhi)(zhi)部分采(cai)用C504單片(pian)機(ji)，通過對(dui)蓄電(dian)(dian)(dian)(dian)(dian)(dian)(dian)池端電(dian)(dian)(dian)(dian)(dian)(dian)(dian)壓(ya)信號的(de)采(cai)集(ji)、分析處理、模糊(hu)推(tui)理[8]、模糊(hu)決策等，控(kong)制(zhi)(zhi)(zhi)二(er)次斬(zhan)波電(dian)(dian)(dian)(dian)(dian)(dian)(dian)路(lu)中的(de)MOSFET管的(de)通斷(duan)時間來控(kong)制(zhi)(zhi)(zhi)充(chong)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)壓(ya)。控(kong)制(zhi)(zhi)(zhi)部分還包(bao)括對(dui)電(dian)(dian)(dian)(dian)(dian)(dian)(dian)流(liu)(liu)和溫(wen)度的(de)采(cai)集(ji)以(yi)及電(dian)(dian)(dian)(dian)(dian)(dian)(dian)壓(ya)和電(dian)(dian)(dian)(dian)(dian)(dian)(dian)流(liu)(liu)的(de)顯(xian)示(shi)。總體(ti)結構如圖3所示(shi)。

圖3 智能充電系統總體結構框圖

5結論

  蓄(xu)電(dian)池(chi)(chi)的(de)充(chong)(chong)(chong)放(fang)電(dian)過(guo)程(cheng)是一個(ge)(ge)復雜(za)的(de)過(guo)程(cheng)，要用精確(que)數學(xue)模型對蓄(xu)電(dian)池(chi)(chi)充(chong)(chong)(chong)電(dian)的(de)控制則有相(xiang)當的(de)難度(du)(du)。蓄(xu)電(dian)池(chi)(chi)的(de)充(chong)(chong)(chong)電(dian)控制系(xi)統(tong)是個(ge)(ge)非(fei)線性的(de)、時變的(de)、有干擾(rao)的(de)、具有純滯后的(de)控制系(xi)統(tong)，在(zai)充(chong)(chong)(chong)放(fang)電(dian)過(guo)程(cheng)中涉及到(dao)很(hen)多參(can)數，如充(chong)(chong)(chong)電(dian)率、最(zui)大允許充(chong)(chong)(chong)電(dian)電(dian)流、內阻(zu)、出氣點電(dian)壓、溫度(du)(du)、壽(shou)命等。

   作者創(chuang)新點為(wei)：

(1) 隸(li)(li)屬(shu)(shu)(shu)函數 的(de)(de)(de)形狀(zhuang)，對控制效(xiao)果影響較大。窄(zhai)型(xing)隸(li)(li)屬(shu)(shu)(shu)函數，反映模糊集(ji)合(he)具有高(gao)(gao)分辨特性(xing)。如果系統(tong)誤(wu)差(cha)，采(cai)用(yong)高(gao)(gao)分辨率(lv)模糊集(ji)合(he)，則(ze)誤(wu)差(cha)控制的(de)(de)(de)靈敏度就會提高(gao)(gao)。在系統(tong)誤(wu)差(cha)較大的(de)(de)(de)范圍內，采(cai)用(yong)具有低分辨率(lv)隸(li)(li)屬(shu)(shu)(shu)函數的(de)(de)(de)模糊集(ji)合(he)；而在系統(tong)誤(wu)差(cha)較小，或接(jie)近于(yu)零(ling)時，宜采(cai)用(yong)具有高(gao)(gao)分辨率(lv)隸(li)(li)屬(shu)(shu)(shu)函數的(de)(de)(de)模糊集(ji)合(he)。

(2)在定義某(mou)一(yi)語(yu)言變量(liang)，如誤差、誤差變化(hua)率和控制量(liang)變化(hua)的(de)全(quan)部集(ji)合(he)時，要考慮其對(dui)論域[-n,+n]的(de)覆蓋(gai)程(cheng)度，語(yu)言變量(liang)的(de)全(quan)部模(mo)糊(hu)集(ji)合(he)所包含的(de)非零隸屬度對(dui)應(ying)的(de)論域元素個數，應(ying)是(shi)模(mo)糊(hu)集(ji)合(he)總數的(de)3-4倍(bei)。

（3）查表法作為模糊控制算法有表格結構單一，修改繁瑣，缺乏靈活性的缺點。針對使用查表法作為模糊控制算法暴露的缺點，在硬件設計中與以補償，加入了一片X5045電可擦除芯片，將模糊控制表格中的數據存儲于微控制器外部存儲空間中，基本上克服了這個缺點。