，一個控制(zhi)規(gui)則表中會(hui)出(chu)現空項,這(zhe)是不能(neng)滿(man)足實際控制(zhi)要求的。為了取得更滿(man)意的控制(zhi)效(xiao)果，可以對(dui)原始的控制(zhi)規(gui)則進行改進。這(zhe)時，應以粗糙的控制(zhi)規(gui)則為基礎，通過仿真實驗和系統調試加以完善。

4.模糊智能充電系統的工(gong)作(zuo)原理及結構

智能充(chong)(chong)(chong)電(dian)(dian)(dian)系(xi)統主(zhu)要由(you)充(chong)(chong)(chong)電(dian)(dian)(dian)電(dian)(dian)(dian)源和(he)(he)單片(pian)機控制電(dian)(dian)(dian)路兩部分(fen)組成。220V的(de)(de)交流(liu)(liu)市(shi)電(dian)(dian)(dian)經(jing)整(zheng)流(liu)(liu)濾(lv)波電(dian)(dian)(dian)路變(bian)為脈(mo)動的(de)(de)310V高壓(ya)直(zhi)(zhi)流(liu)(liu)。然后經(jing)DC-DC變(bian)換電(dian)(dian)(dian)路（脈(mo)沖功率(lv)變(bian)壓(ya)器）變(bian)為充(chong)(chong)(chong)電(dian)(dian)(dian)所需的(de)(de)60V直(zhi)(zhi)流(liu)(liu)電(dian)(dian)(dian)壓(ya)。為了保證輸出電(dian)(dian)(dian)壓(ya)的(de)(de)穩(wen)定性，采(cai)(cai)用(yong)(yong)了UC3842對(dui)(dui)60V直(zhi)(zhi)流(liu)(liu)電(dian)(dian)(dian)壓(ya)進行穩(wen)壓(ya)。二次斬波電(dian)(dian)(dian)路主(zhu)要由(you)MOSFET管(guan)(guan)、電(dian)(dian)(dian)感(gan)、電(dian)(dian)(dian)容和(he)(he)二極(ji)管(guan)(guan)組成，輸出24-36V的(de)(de)充(chong)(chong)(chong)電(dian)(dian)(dian)電(dian)(dian)(dian)壓(ya)。控制部分(fen)采(cai)(cai)用(yong)(yong)C504單片(pian)機，通過對(dui)(dui)蓄電(dian)(dian)(dian)池端電(dian)(dian)(dian)壓(ya)信(xin)號的(de)(de)采(cai)(cai)集(ji)、分(fen)析(xi)處(chu)理、模糊推理[8]、模糊決(jue)策(ce)等(deng)，控制二次斬波電(dian)(dian)(dian)路中的(de)(de)MOSFET管(guan)(guan)的(de)(de)通斷時間來控制充(chong)(chong)(chong)電(dian)(dian)(dian)電(dian)(dian)(dian)壓(ya)。控制部分(fen)還包括對(dui)(dui)電(dian)(dian)(dian)流(liu)(liu)和(he)(he)溫度的(de)(de)采(cai)(cai)集(ji)以及電(dian)(dian)(dian)壓(ya)和(he)(he)電(dian)(dian)(dian)流(liu)(liu)的(de)(de)顯示(shi)。總體結構如(ru)圖3所示(shi)。

圖(tu)3 智能充(chong)電系統總體(ti)結構框圖(tu)

5 蓄電(dian)(dian)池的充(chong)放電(dian)(dian)過程是(shi)一個(ge)復雜的過程，要用精確數學模型對蓄電(dian)(dian)池充(chong)電(dian)(dian)的控(kong)制(zhi)(zhi)則有(you)相當的難(nan)度(du)。蓄電(dian)(dian)池的充(chong)電(dian)(dian)控(kong)制(zhi)(zhi)系統(tong)是(shi)個(ge)非(fei)線性(xing)的、時變的、有(you)干擾的、具有(you)純(chun)滯后的控(kong)制(zhi)(zhi)系統(tong)，在(zai)充(chong)放電(dian)(dian)過程中(zhong)涉及到很多參數，如充(chong)電(dian)(dian)率、最大允許充(chong)電(dian)(dian)電(dian)(dian)流、內阻、出(chu)氣點(dian)電(dian)(dian)壓、溫(wen)度(du)、壽(shou)命(ming)等。

   作(zuo)者創新點(dian)為：

(1) 隸(li)(li)屬(shu)函數 的(de)形狀，對控(kong)(kong)制效果影響(xiang)較大。窄型隸(li)(li)屬(shu)函數，反映模糊集(ji)合具(ju)有(you)高(gao)分(fen)辨(bian)特性。如(ru)果系(xi)統誤差，采用(yong)高(gao)分(fen)辨(bian)率(lv)(lv)模糊集(ji)合，則誤差控(kong)(kong)制的(de)靈敏度就會(hui)提高(gao)。在系(xi)統誤差較大的(de)范圍內，采用(yong)具(ju)有(you)低分(fen)辨(bian)率(lv)(lv)隸(li)(li)屬(shu)函數的(de)模糊集(ji)合；而(er)在系(xi)統誤差較小，或接近于零時，宜采用(yong)具(ju)有(you)高(gao)分(fen)辨(bian)率(lv)(lv)隸(li)(li)屬(shu)函數的(de)模糊集(ji)合。

(2)在定義(yi)某一(yi)語言變量(liang)(liang)，如(ru)誤(wu)差、誤(wu)差變化率(lv)和(he)控制量(liang)(liang)變化的全(quan)部集合(he)時，要考慮其對論(lun)域[-n,+n]的覆蓋(gai)程度，語言變量(liang)(liang)的全(quan)部模(mo)糊(hu)集合(he)所包含的非零(ling)隸(li)屬度對應(ying)的論(lun)域元素個數，應(ying)是模(mo)糊(hu)集合(he)總數的3-4倍。

，將模(mo)糊(hu)控(kong)制(zhi)表(biao)格中的(de)(de)數據存(cun)儲于微控(kong)制(zhi)器外(wai)部存(cun)儲空間中，基(ji)本上克服了這個缺點。芯片(pian)電(dian)可(ke)擦除X5045）查(cha)表(biao)法作為(wei)模(mo)糊(hu)控(kong)制(zhi)算法有表(biao)格結構單(dan)一，修改(gai)繁瑣，缺乏(fa)靈活性的(de)(de)缺點。針對(dui)使用查(cha)表(biao)法作為(wei)模(mo)糊(hu)控(kong)制(zhi)算法暴露的(de)(de)缺點，在硬件(jian)設(she)計中與以補償，加入了一片(pian)