BP神经网络对船舶远程监控系统的改进分析_英文_.pdf

　第 32 卷 第 2 期 中　国　航 　海 Vol . 32 No . 2 　 　2009 年 6 月 N AV I GA TION O F C H IN A J un . 2009 　 文章编号 :1000 - 4653 (2009) 02 - 00 14 - 06 A Shipboard Remote Surveillance System Improved by BP Neural Net works Z H A N G Yuew en , 　CUI Wenbi n , 　W U Guit ao , 　S UN Peit i ng ( College of Marine Engineering , Dalian Maritime U niver sit y , Dalian 116026 , China) Abstract : So a s to sati sfy t he requirement s for remot ely monitoring oceangoing ship s by more effective met ho ds , B P neural network s were app lied to imp rove t he exi sting monitoring sy st em by outp utting p rewarning signal s in st ead of failure signal s. Then t he alarm app arat u s on boar d ship s could be act uat ed , and t he relevant identification number s were ready to be sent to t he comp anies a shore . Sub sequently , wit h t he alarm app arat u s act uat ed and t he identifica tion number s obt ained , bot h t he ship s in t he ocean and t he comp anies ashore could gra sp t he best chance to solve t he pot ential p roblem s. The exhau st t emp erat ure alt ering t endency of a sixcylinder diesel engine wa s p ropo sed to ana lyze t he app lication of B P neural network s to t he ship boar d remot e surveillance sy st em . Due to it s learning abilit y , B P neural network s earned t he t endency ,and t he identification number of gradually ri sing exhau st t emp erat ure wa s di stingui shed . The result indicat ed t hat B P neural network s were p racticable for