，一個控制(zhi)規則(ze)表(biao)中(zhong)會出現空項,這是不能滿足實際(ji)控制(zhi)要求的(de)。為了取得更(geng)滿意(yi)的(de)控制(zhi)效(xiao)果，可以(yi)對(dui)原始的(de)控制(zhi)規則(ze)進行改進。這時，應以(yi)粗糙的(de)控制(zhi)規則(ze)為基礎，通過(guo)仿真實驗和系統調試加以(yi)完(wan)善。

4.模糊智能充電系統的工(gong)作原(yuan)理及結構

智能充電(dian)(dian)(dian)(dian)系統主(zhu)要(yao)(yao)由(you)充電(dian)(dian)(dian)(dian)電(dian)(dian)(dian)(dian)源(yuan)和(he)(he)單片機控(kong)(kong)(kong)制(zhi)(zhi)電(dian)(dian)(dian)(dian)路(lu)兩部分(fen)組成。220V的(de)(de)交(jiao)流(liu)市電(dian)(dian)(dian)(dian)經(jing)整流(liu)濾波(bo)電(dian)(dian)(dian)(dian)路(lu)變為(wei)脈動的(de)(de)310V高壓(ya)直(zhi)(zhi)流(liu)。然后經(jing)DC-DC變換電(dian)(dian)(dian)(dian)路(lu)（脈沖功率變壓(ya)器(qi)）變為(wei)充電(dian)(dian)(dian)(dian)所需的(de)(de)60V直(zhi)(zhi)流(liu)電(dian)(dian)(dian)(dian)壓(ya)。為(wei)了保證輸出電(dian)(dian)(dian)(dian)壓(ya)的(de)(de)穩定性(xing)，采用了UC3842對60V直(zhi)(zhi)流(liu)電(dian)(dian)(dian)(dian)壓(ya)進行穩壓(ya)。二次斬(zhan)(zhan)波(bo)電(dian)(dian)(dian)(dian)路(lu)主(zhu)要(yao)(yao)由(you)MOSFET管、電(dian)(dian)(dian)(dian)感、電(dian)(dian)(dian)(dian)容和(he)(he)二極管組成，輸出24-36V的(de)(de)充電(dian)(dian)(dian)(dian)電(dian)(dian)(dian)(dian)壓(ya)。控(kong)(kong)(kong)制(zhi)(zhi)部分(fen)采用C504單片機，通過對蓄電(dian)(dian)(dian)(dian)池端(duan)電(dian)(dian)(dian)(dian)壓(ya)信號的(de)(de)采集、分(fen)析處理(li)、模(mo)糊推理(li)[8]、模(mo)糊決策等，控(kong)(kong)(kong)制(zhi)(zhi)二次斬(zhan)(zhan)波(bo)電(dian)(dian)(dian)(dian)路(lu)中的(de)(de)MOSFET管的(de)(de)通斷時(shi)間來控(kong)(kong)(kong)制(zhi)(zhi)充電(dian)(dian)(dian)(dian)電(dian)(dian)(dian)(dian)壓(ya)。控(kong)(kong)(kong)制(zhi)(zhi)部分(fen)還包括對電(dian)(dian)(dian)(dian)流(liu)和(he)(he)溫(wen)度的(de)(de)采集以及電(dian)(dian)(dian)(dian)壓(ya)和(he)(he)電(dian)(dian)(dian)(dian)流(liu)的(de)(de)顯示。總(zong)體結構如圖(tu)3所示。

圖(tu)3 智能充電系統總體結構框圖(tu)

5 蓄(xu)電(dian)(dian)(dian)(dian)池(chi)的(de)(de)充(chong)放電(dian)(dian)(dian)(dian)過(guo)程是(shi)(shi)一個(ge)復雜的(de)(de)過(guo)程，要用精確數學模(mo)型(xing)對蓄(xu)電(dian)(dian)(dian)(dian)池(chi)充(chong)電(dian)(dian)(dian)(dian)的(de)(de)控制則有相(xiang)當的(de)(de)難度(du)(du)。蓄(xu)電(dian)(dian)(dian)(dian)池(chi)的(de)(de)充(chong)電(dian)(dian)(dian)(dian)控制系(xi)統是(shi)(shi)個(ge)非線性(xing)的(de)(de)、時變(bian)的(de)(de)、有干擾(rao)的(de)(de)、具有純滯后的(de)(de)控制系(xi)統，在充(chong)放電(dian)(dian)(dian)(dian)過(guo)程中涉及到很多參數，如充(chong)電(dian)(dian)(dian)(dian)率、最大(da)允許充(chong)電(dian)(dian)(dian)(dian)電(dian)(dian)(dian)(dian)流、內(nei)阻(zu)、出氣點電(dian)(dian)(dian)(dian)壓(ya)、溫度(du)(du)、壽命等。

   作者創(chuang)新(xin)點為：

(1) 隸(li)(li)屬函(han)數(shu) 的形狀，對(dui)控(kong)(kong)制效果影響較大(da)。窄型隸(li)(li)屬函(han)數(shu)，反映模糊集(ji)(ji)合具(ju)(ju)有(you)高(gao)(gao)(gao)分辨(bian)特性(xing)。如果系統(tong)誤差，采用(yong)高(gao)(gao)(gao)分辨(bian)率模糊集(ji)(ji)合，則誤差控(kong)(kong)制的靈(ling)敏(min)度就會提(ti)高(gao)(gao)(gao)。在系統(tong)誤差較大(da)的范圍內，采用(yong)具(ju)(ju)有(you)低分辨(bian)率隸(li)(li)屬函(han)數(shu)的模糊集(ji)(ji)合；而在系統(tong)誤差較小(xiao)，或接近于(yu)零時(shi)，宜(yi)采用(yong)具(ju)(ju)有(you)高(gao)(gao)(gao)分辨(bian)率隸(li)(li)屬函(han)數(shu)的模糊集(ji)(ji)合。

(2)在(zai)定義某一語(yu)言(yan)變(bian)量(liang)，如(ru)誤差、誤差變(bian)化率(lv)和控制量(liang)變(bian)化的(de)全部集(ji)合(he)(he)時，要考慮其對論(lun)域[-n,+n]的(de)覆(fu)蓋程度，語(yu)言(yan)變(bian)量(liang)的(de)全部模(mo)糊(hu)集(ji)合(he)(he)所包含的(de)非零隸屬度對應的(de)論(lun)域元素個數(shu)，應是模(mo)糊(hu)集(ji)合(he)(he)總數(shu)的(de)3-4倍(bei)。

，將(jiang)模糊(hu)控制(zhi)表(biao)格中的數據存儲于微(wei)控制(zhi)器外部存儲空(kong)間(jian)中，基本(ben)上克服了這個(ge)缺點。芯(xin)片電可擦除X5045）查表(biao)法(fa)(fa)作為(wei)模糊(hu)控制(zhi)算法(fa)(fa)有表(biao)格結構單(dan)一，修改(gai)繁瑣，缺乏靈活性的缺點。針對使(shi)用查表(biao)法(fa)(fa)作為(wei)模糊(hu)控制(zhi)算法(fa)(fa)暴露的缺點，在硬件(jian)設計(ji)中與以補償，加入了一片