應(yīng)用特征識別的神經(jīng)網(wǎng)絡(luò)模增強(qiáng)塑料部件

第第頁應(yīng)用特征識別的神經(jīng)網(wǎng)絡(luò)模增強(qiáng)塑料部件應(yīng)用特征識別的神經(jīng)網(wǎng)絡(luò)模加強(qiáng)塑料部件

Applicationofneuralnetworksinfeaturerecognitionofmouldreinforcedplasticparts

Marquez,M.1,3,4,5;Gill,R.2,6,7;White,A.2,6

Source:ConcurrentEngineeringResearchandApplications,v7,n2,p115-122,June1999;ISSN:1063293*;Publisher:TechnomicPublCoInc

Authoraffiliations:

1

2

3UniversidadNacionalE*perimentaldelTachira(UNET),SanCristobal,VenezuelaSchoolofEngineeringSystems,Middlese*University,London,UnitedKingdomMiddlese*University,SchoolofEngineeringSystems,BoundsGreenRoad,N112NQ,London,UnitedKingdom

4

5

6

7UniversidadNacionalE*perimentaldelTachira(UNET,Venezuela)Middlese*University,London,UnitedKingdomSchoolofEngineeringSystems,Middlese*UniversityCambridgeUniversityManufacturingGroup

Abstract:

Featurerecognitionisanapplicationdependanttask,whichhasbeenmostlyfocusedinproductionplanningofmachiningprocess.Itplaysafundamentalroleandusuallyisthefirststepindownstreamactivitiesconcerningproductdevelopmentprocesssuchasdesignformanufacturing,designforassemblyandprocessplanning.Thisreportpresentsamethodologytocarryoutrecognitionofdesignformanufacturingfeaturesofreinforcedplasticcomponents.Athree-layerneuralnetworksystemwascreatedandtrainedusingback-propagation-supervisedlearningtorecognisenineofthemostimportantdesignfeaturesrelatedtothismanufacturingprocess.Also,amethodologyforpre-processing3-Dsolidmodelssuchthatgeometricalandtopologicalinformationofthepartcouldbesuitableasnetworkinputispresented.Highperformanceofthenetsystemwasachievedontherecognitionofthetrainedfeaturesasitwasobservedinseveraltestparts.(7refs)

Mainheading:

Patternrecognition

Controlledterms:

Backpropagation-Computeraideddesign-Concurrentengineering-Learningsystems-Mathematicalmodels-Neuralnetworks-Plasticparts-Reinforcedplastics

Uncontrolledterms:

Designformanufacturing-Featurerecognition-Mouldreinforcedplasticparts

ClassificationCode:

723.4ArtificialIntelligence-723.5ComputerApplications-817.1PolymerProducts-913.6ProductDevelopment;ConcurrentEngineering-921.6NumericalMethods

Treatment:

Applications(APP)

應(yīng)用特征識別的神經(jīng)網(wǎng)絡(luò)模加強(qiáng)塑料部件

Database:

Compende*應(yīng)用特征識別的神經(jīng)網(wǎng)絡(luò)模加強(qiáng)塑料部件

馬爾克斯,M。1、3、4、5,吉爾。2、6、7,白色。2、6

來源:并行工程的討論和應(yīng)用,v7n2,p115-122年,1999年6月,臺北:1063293*;出版者:Technomic出版有限公司

社會兼職:

1高校試驗舉措讓德爾(UNET)、圣克里斯托瓦爾委內(nèi)瑞拉

2工程系統(tǒng),密德薩斯高校,倫敦,英國

3米德爾塞克斯高校工程學(xué)院的系統(tǒng)范圍內(nèi)綠色道路,N112向,倫敦,英國

4高校試驗舉措讓德爾(UNET,委內(nèi)瑞拉)

5米德爾塞克斯高校,倫敦,英國

6工程系統(tǒng),密德薩斯高校

7劍橋高校制造集團(tuán)

文摘:

特征識別是一個應(yīng)用程序依靠的任務(wù),一貫主要集中在加工過程的生產(chǎn)計劃。它起著基本的作用,通常是第一步在下游活動涉及產(chǎn)品開發(fā)過程如設(shè)計制造、組裝和設(shè)計流程規(guī)劃。這份報告提出了一個方法進(jìn)行識別的設(shè)計制造加強(qiáng)塑料組件的特性。創(chuàng)建一個三層的神經(jīng)網(wǎng)絡(luò)系統(tǒng)學(xué)習(xí)和訓(xùn)練運用back-propagation-supervised承認(rèn)九最重要的設(shè)計特點與此相關(guān)的制造過程。此外,預(yù)處理三維固體模型的方法,幾何和拓?fù)湫畔⒌囊徊糠挚赡苁呛线m的作為網(wǎng)絡(luò)的輸入。高性能網(wǎng)絡(luò)系統(tǒng)實現(xiàn)了識別的訓(xùn)練有素的特性是觀測在幾個測試部分。參(7)主標(biāo)題:模式識別

掌握:反向傳播——計算機(jī)幫助設(shè)計并行工程-學(xué)習(xí)系統(tǒng)數(shù)學(xué)模型-神經(jīng)網(wǎng)絡(luò)——塑料部件——加強(qiáng)塑料

掌握方面:設(shè)計制造-特性識別模具加強(qiáng)塑料部件

人工智能分類代碼:723.4-723.5計算機(jī)應(yīng)用817.1-817.1聚合物產(chǎn)品——產(chǎn)品開發(fā),并行工程-921.6數(shù)值方法

治療:應(yīng)用程序(應(yīng)用程序)

數(shù)據(jù)庫:核心期刊

應(yīng)用特征識別的神經(jīng)網(wǎng)絡(luò)模加強(qiáng)塑料部件

Applicationofneuralnetworksinfeaturerecognitionofmouldreinforcedplasticparts

Marquez,M.1,3,4,5;Gill,R.2,6,7;White,A.2,6

Source:ConcurrentEngineeringResearchandApplications,v7,n2,p115-122,June1999;ISSN:1063293*;Publisher:TechnomicPublCoInc

Authoraffiliations:

1

2

3UniversidadNacionalE*perimentaldelTachira(UNET),SanCristobal,VenezuelaSchoolofEngineeringSystems,Middlese*University,London,UnitedKingdomMiddlese*University,SchoolofEngineeringSystems,BoundsGreenRoad,N112NQ,London,UnitedKingdom

4

5

6

7UniversidadNacionalE*perimentaldelTachira(UNET,Venezuela)Middlese*University,London,UnitedKingdomSchoolofEngineeringSystems,Middlese*UniversityCambridgeUniversityManufacturingGroup

Abstract:

Featurerecognitionisanapplicationdependanttask,whichhasbeenmostlyfocusedinproductionplanningofmachiningprocess.Itplaysafundamentalroleandusuallyisthefirststepindownstreamactivitiesconcerningproductdevelopmentprocesssuchasdesignformanufacturing,designforassemblyandprocessplanning.Thisreportpresentsamethodologytocarryoutrecognitionofdesignformanufacturingfeaturesofreinforcedplasticcomponents.Athree-layerneuralnetworksystemwascreatedandtrainedusingback-propagation-supervisedlearningtorecognisenineofthemostimportantdesignfeaturesrelatedtothismanufacturingprocess.Also,amethodologyforpre-processing3-Dsolidmodelssuchthatgeometricalandtopologicalinformationofthepartcouldbesuitableasnetworkinputispresented.Highperformanceofthenetsystemwasachievedontherecognitionofthetrainedfeaturesasitwasobservedinseveraltestparts.(7refs)

Mainheading:

Patternrecognition

Controlledterms:

Backpropagation-Computeraideddesign-Concurrentengineering-Learningsystems-Mathematicalmodels-Neuralnetworks-Plasticparts-Reinforcedplastics

Uncontrolledterms:

Designformanufacturing-Featurerecognition