基于rbf神经网络的电化学co气体传感器的温度补偿传感技术学报.pdfVIP

第 22 卷 　第 1 期 传 感 技 术 学 报 Vol . 22 　N o . 1 2009 年 1 月 CHIN ES E J OU RNAL O F S EN SORS AND AC TUA TORS J an . 2009 Temperature Compensation of the CO Electrochemical Gas Sensor Based on RBF Neural Net work Z H A N G X i aoj un , Z H A N G M i ngl u , L I X i aoh ui ( ) S chool Of M echanical Eng i neeri ng , H ebei Uni vers ity Of Technology , Ti anj i n 300 130 , Chi na Abstract :The outp ut p reci sion of t he CO elect rochemical ga s sen sor i s ea sy to be influenced by t he ambient t emp erat ure . In view of t hi s short co ming , t he p ap er p rop o se s a new t emp erat ure co mp en sation met ho d ba sed on RB F ( Rae sine Ba sic Function) neural net wor k , an d ha s done t he exp eriment al st u dy by u sing t he de signed ga s gat hering sy st em . The exp eriment re sult s indicat e t hat t he bigge st error of t he sen sor outp ut s may be up to 20 . 0 % wit ho ut carrying on any t emp erat ure co mp en sation , and i s 1. 44 % by adop tin g t he t emp erat ure co mp en sation met ho d ba sed on B P neural net wor k ,even dow n to 0 . 12 % aft er u sin g t he met h o d ba sed on RB F neural net wor k . Therefore t hi s met ho d may be effectively u sed in t emp erat ure co mp en sa tion for t he CO elect rochemical ga s sen sor and make t he sen sor p o sse ss t he higher mea surin g accuracy and t he t emp erat ure st abilit y . Key words :RB F neural net wor k ; CO elect rochemical ga s sen sor ; t emp erat ure co mp en sation EEACC :7230 基于 RBF 神经网络的电化学