版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進(jìn)行舉報(bào)或認(rèn)領(lǐng)
文檔簡介
1、<p> 畢業(yè)設(shè)計(jì)(論文)外文資料翻譯</p><p> 系 部: 機(jī)械工程系 </p><p> 專 業(yè): 機(jī)械工程及自動(dòng)化 </p><p> 姓 名: </p>
2、<p> 學(xué) 號: </p><p> 外文出處: Journal of Engieering Materials </p><p> and Technology </p><
3、;p> 附 件: 1.外文資料翻譯譯文;2.外文原文。 </p><p> 附件1:外文資料翻譯譯文</p><p> 加強(qiáng)使用粘結(jié)劑力軌跡和神經(jīng)網(wǎng)絡(luò)控制</p><p><b> 獲得一致的最小回彈</b></p><p> 制造鈑金屬片面臨的最大挑戰(zhàn)之一是要取得一致的回彈量?;貜?,當(dāng)工件加工撤掉
4、后,彈性材料恢復(fù),主要的原因是最后一部分的幾何變化的不一致。由于非線性效應(yīng),相互作用過程和材料參數(shù)的關(guān)系,要獲得一致可取的回彈量是極其困難。在本文中,回彈量在模擬通道的形成過程證明了神經(jīng)網(wǎng)絡(luò)系統(tǒng)的獨(dú)特的能力同時(shí)加上加強(qiáng)粘結(jié)劑力軌跡控制回彈角和最大主應(yīng)變量起的作用。當(dāng)面臨著很大差異甚至在材料性能,板度,摩擦條件下,我們的控制系統(tǒng)將會(huì)制作出零件最終的形狀。</p><p><b> 導(dǎo)言</b>
5、;</p><p> 在當(dāng)今競爭激烈的制造業(yè)中零件獲得一致準(zhǔn)確的尺寸是至關(guān)重要的。不一致的零件尺寸會(huì)延緩新產(chǎn)品的開發(fā),增加轉(zhuǎn)換時(shí)間,使后續(xù)開發(fā)生產(chǎn)困難,還需要額外的保證工序質(zhì)量,同時(shí)還降低客戶對終端產(chǎn)品的滿意度和信任度。在金屬板料成形過程中,回彈方面的材料是保證零件最終精確的一個(gè)主要因素?;貜検窃谄錆M載的條件下幾何的差額部分,例如,符合工件的幾何形狀,當(dāng)局部已卸下,自由的狀態(tài),幾何之間的差額。對于一個(gè)復(fù)雜的三維
6、零件,不必要的扭曲是另一種形式的回彈。薄板厚度方向中壓力分布不一致和沖壓加載放寬和卸載過程中,其結(jié)果是產(chǎn)生回彈。影響回彈的因素包括變化過程和材料參數(shù),如摩擦條件下,工件的幾何形狀,材料性能,板的厚度,以及模具溫度。由于在制造過程中控制所有這些變數(shù)是幾乎不可能的,反過來證明回彈是不能隨時(shí)控制?;貜検且粋€(gè)高度非線性效應(yīng),這個(gè)新增加的困難是事實(shí),因此,模擬和糾正的方法是復(fù)雜的。最近幾年出現(xiàn)了許多研究有關(guān)回彈,例如,汽車工程師學(xué)會(huì),NUMIFO
7、RM和NUMISHEET會(huì)議。</p><p> 通過修改形成過程可減少回彈。一些研究人員提出了使用加強(qiáng)粘結(jié)劑力軌跡以實(shí)現(xiàn)這一目標(biāo)(Ayres,1984;Hishida and Wagoner, 1993; Sunseri et al., 1996)。進(jìn)一步加強(qiáng)粘結(jié)劑力軌跡是一個(gè)瞬時(shí)跳躍從低粘結(jié)力(LBF)到高粘合劑力(HBF),在指定的百分比總額的沖壓位移(PD%)(參閱圖1)。Sunseri et al.
8、(1996),調(diào)查鋁通道形成過程的回彈,圖2所示。他們的工作是通過模擬實(shí)驗(yàn)明確形成過程和材料參數(shù)用處。在生產(chǎn)環(huán)境中,可是,由于不同的進(jìn)程回彈量將偏離所期望的水平。因此,控制系統(tǒng),可變化的工藝參數(shù)要求</p><p> 近年來,許多研究小組調(diào)查了利用人工神經(jīng)網(wǎng)絡(luò)系統(tǒng)來控制金屬板料成形過程。由于非線性效應(yīng)和相互作用的過程參數(shù)金屬成型是an ideal candidate,神經(jīng)網(wǎng)絡(luò)系統(tǒng)控制。(Cho et al. (
9、1997)采用了神經(jīng)網(wǎng)絡(luò)預(yù)測的力量在冷軋和,(Di and Thomson(1997)預(yù)測起皺限制正方米板材下角的緊拉應(yīng)力。除此之外,(Elkins, 1994 and Yang et al., 1992), Forcellese et al. (1997)采用了神經(jīng)網(wǎng)絡(luò)系統(tǒng)控制60度鋁V型沖壓空氣彎曲回彈的進(jìn)程。他們的系統(tǒng)利用實(shí)驗(yàn)獲得的例子組成的5個(gè)參數(shù)從力沖壓軌跡,離線測量板厚,神經(jīng)網(wǎng)絡(luò)目標(biāo)彎曲角度的投入和沖壓的移位的輸出。在另一項(xiàng)
10、研究項(xiàng)目,Ruffini and Cao(1998)建議采用神經(jīng)網(wǎng)絡(luò)控制回彈角的渠道的形成過程與軌跡沖壓力作為唯一的來源,確定工藝變化和調(diào)整的HBF,Sunseri et al. (1996)。初步結(jié)果表明這種方法是很有希望的。 </p><p> 本文中,鋁渠道的回彈是通過加強(qiáng)控制粘結(jié)劑力軌跡和神經(jīng)網(wǎng)絡(luò)控制。神經(jīng)網(wǎng)絡(luò)確定了HBF和PD%的粘結(jié)劑力軌跡。沖床力軌跡被確定為關(guān)鍵參數(shù)的變化,反映在材料性能,板的厚
11、度,還有摩擦系數(shù)。因此,四個(gè)多項(xiàng)式系數(shù)從曲線擬合的沖壓力軌跡被用來作為投入的神經(jīng)網(wǎng)絡(luò)系統(tǒng)。圖3顯示流程圖的控制系統(tǒng)提出這一點(diǎn)應(yīng)用。盡管材料性能有著很大差異,板厚度(T),和摩擦系數(shù)(m),回彈角(θ)之間保持了0.2和0.6度,(e)項(xiàng)限于8%和10%。最后,由Sunseri et al.(1996)提出一個(gè)比較閉環(huán)控制方法,是為了顯示我們控制方法的好處。雖然模擬結(jié)果在這里只有數(shù)值,然而控制系統(tǒng)將在未來改進(jìn)需求,以驗(yàn)證實(shí)際執(zhí)行情況。&l
12、t;/p><p><b> 通道的形成過程</b></p><p> 調(diào)查回彈在一個(gè)通道的形成過程是一個(gè)簡單的幾何學(xué)。因此,由受雇這里得Sunseri et al.(1996)提出使用鋁通道形成過程(參閱圖.2 ) 。 </p><p> 第一,連續(xù)的粘結(jié)劑力(CBF)對影響回彈進(jìn)行了評價(jià)。由于CBF增加,我們的模擬在所有其他進(jìn)程和材料參數(shù)進(jìn)
13、行了不斷實(shí)驗(yàn),回彈角(u)降低,圖形顯示圖4,物理圖5,然而,增加粘結(jié)劑力同時(shí)還將造成材料的最大的拉緊力增加,實(shí)線圖4,鋁通道的突出(Graf and Hosford, 1993)。通過利用粘結(jié)劑加強(qiáng)力軌跡,適度的增加側(cè)壁得到應(yīng)變力水平,同時(shí)減少回彈的過程中體現(xiàn)在表中的圖1。</p><p> 當(dāng)面臨著偏差的摩擦系數(shù),只是為了產(chǎn)生一個(gè)健全的進(jìn)程,Sunseri et al.(1996)實(shí)施閉環(huán)變量粘結(jié)劑力控制后
14、續(xù)零件的沖壓力軌跡從局部加強(qiáng)粘結(jié)劑的情況制定工藝條件。當(dāng)摩擦系數(shù)在不同的用途系數(shù)為0.1,甚至是到0.25,這種控制方法能產(chǎn)生回彈的水平一致。然而,這種方法是否可以承受其他參數(shù)的變化,如材料性能和板料厚還沒有確定的情況下。</p><p> 商業(yè)有限元分析套件(ABAQUS,1997年)被我們用于形成過程中的數(shù)值模擬通道,因?yàn)檫@個(gè)問題是接近平面變形應(yīng)力條件,并對稱,只有十六分之一的寬度和一半以上的長度在整個(gè)空白
15、(220毫米346毫米)為藍(lán)本。粘結(jié)劑,模具和沖壓是仿照作為三個(gè)獨(dú)立的硬質(zhì)表面。每個(gè)面是仿照采用四節(jié)點(diǎn)界面元素(ABAQUS類型IRS4)和庫侖摩擦法是假設(shè)。我們的空白網(wǎng)孔40四節(jié)點(diǎn)分配不均,減少了一體化殼元素(ABAQUS類型S4R)更致密的沖壓和彎道半徑粘結(jié)劑集中接觸在空白的網(wǎng)孔處。邊界條件指定創(chuàng)建一個(gè)平面應(yīng)變條件。該材料是各向同性的模板,彈塑性后,馮米塞斯屈服準(zhǔn)則和各向同性應(yīng)變硬化。彈性性能的楊氏模量中,E,,70千兆和泊松比,n
16、,0.3。塑料薄板材料的參數(shù)是模板,可使用一些教學(xué)關(guān)系(s5Ken))。我們的名義材料,指材料1,有材料強(qiáng)度系數(shù),K,528兆帕和應(yīng)變硬化指數(shù),n,0.265有限元模型</p><p><b> 擬議控制系統(tǒng)</b></p><p> 在一個(gè)通道的形成過程,回彈一般是極為敏感的材料的變化和成形參數(shù)。在這項(xiàng)工作中,我們制訂一種方法控制回彈,同時(shí)材料通過結(jié)合加強(qiáng)粘結(jié)劑
17、力軌跡和神經(jīng)網(wǎng)絡(luò)控制產(chǎn)生一個(gè)可以接受的最大應(yīng)變量。在加強(qiáng)局部粘結(jié)劑軌跡,兩個(gè)臨界值需要加以確定,巨大的HBF和總沖壓力的移位的PD%,這兩個(gè)參數(shù)的輸出從神經(jīng)網(wǎng)絡(luò)系統(tǒng)。從我們以往的研究經(jīng)驗(yàn),在金屬板料成形過程控制(Ruffiniand Cao, 1998;Kinsey and Cao, 1997; and Sunseri et al.,1996) ,沖壓力的軌跡被選定為參數(shù),提供有關(guān)當(dāng)前進(jìn)程。因此,多項(xiàng)式系數(shù)從曲線擬合的沖壓力軌跡被用來作
18、為輸入的神經(jīng)網(wǎng)絡(luò)。</p><p> 我們提出的控制系統(tǒng)流程圖如圖所示圖3,形成過程通常是將著手在連續(xù)使用的正常的粘結(jié)劑力,16千牛頓,深度為10毫米。雖然沖壓的移位會(huì)繼續(xù)下去,多項(xiàng)式系數(shù)從曲線擬合的沖壓力軌跡的計(jì)算和輸入和輸出的神經(jīng)網(wǎng)絡(luò),該HBF和PD%的粘結(jié)劑力加強(qiáng)軌跡,在指定的沖床的位置將得到及時(shí)作出HBF適當(dāng)?shù)恼{(diào)整,19毫米乘以PD%。</p><p> 其中系數(shù)10毫米的用處
19、從沖壓力軌跡的計(jì)算,選擇有兩個(gè)原因。在這個(gè)距離,沖壓力軌跡是很精確的,有足夠的數(shù)據(jù)點(diǎn)可以準(zhǔn)確擬合發(fā)生的曲線。圖6顯示的效果差異t和m對沖壓武力軌跡從0到10毫米。其次,10毫米距離允許足夠的CPU時(shí)間來控制計(jì)算方法。假設(shè)沖床速度50毫米/秒,并設(shè)定最低PD</p><p> ?。?,57.5%的總沖壓移位,10.925毫米,大約有18.5毫秒可用來計(jì)算輸入人工神經(jīng)網(wǎng)絡(luò)和預(yù)測階梯粘結(jié)劑力軌跡的HBF和PD%。曲線擬合
20、和神經(jīng)網(wǎng)絡(luò)的程序?qū)ξ覀兊谋简vII,233兆赫的電腦需要大約10毫秒來計(jì)算。在形成過程中如何接近函數(shù)將取決于粘結(jié)劑力速度不同進(jìn)程的機(jī)制。</p><p> 沖壓力的軌跡被分為三個(gè)區(qū)域,區(qū)域A,一個(gè)過渡區(qū)域,和B區(qū)(參照圖6)。薄板厚度的用途的改變與區(qū)域A的變化相類似,在這個(gè)區(qū)域中二次多項(xiàng)式被用于計(jì)算出正確的數(shù)據(jù)。盡管事實(shí)是二次多項(xiàng)式只有兩個(gè)輸入系數(shù)分別用于本區(qū)域的特點(diǎn),因?yàn)闆_壓力軌跡幾乎傳遞給了來源數(shù)據(jù)。在B區(qū),
21、造成不同的傾斜度的主要主要原因是摩擦系數(shù)的變化。網(wǎng)絡(luò)系統(tǒng)線性插值提供了兩個(gè)額外的輸入數(shù)據(jù)。因此,一共有四個(gè)多項(xiàng)式系數(shù)被用作輸入到我們的神經(jīng)網(wǎng)絡(luò)。</p><p> 值得注意的是沖壓力軌跡取自數(shù)值模擬,因此,它們是光滑的并由于t和m的變化而多樣性。在實(shí)際形成過程中,噪聲數(shù)據(jù)采集設(shè)備將產(chǎn)生的差額用這些曲線表達(dá)出來。通過使用多項(xiàng)式系數(shù)來表達(dá)曲線擬合的沖壓力軌跡,輸入數(shù)據(jù)到神經(jīng)網(wǎng)絡(luò)時(shí)的本質(zhì)是要過濾掉實(shí)踐時(shí)考慮的因素。&
22、lt;/p><p> 為了實(shí)施這控制系統(tǒng)在實(shí)際成形技術(shù),應(yīng)用的數(shù)據(jù)是需要我們整理出來。該資料可從嘗試通過改變沖壓模具工藝參數(shù)在實(shí)際生產(chǎn)中的得到。舉例來說,考慮到一批一批的材料變化,材料可以得到和建立數(shù)據(jù),它體現(xiàn)了金屬板厚度和材料性質(zhì)變化的規(guī)律。此外,潤滑狀態(tài)可能是多種多樣的。利用各種組合的這些工藝參數(shù),該HBF和PD%的價(jià)值,產(chǎn)生理想的回彈量,可確定。此外,數(shù)值模擬,校準(zhǔn)實(shí)驗(yàn)結(jié)果,以確保準(zhǔn)確性,可以用來迅速增加的訓(xùn)
23、練數(shù)據(jù)。一旦神經(jīng)網(wǎng)絡(luò)是訓(xùn)練有素的整個(gè)一系列潛在的工藝參數(shù)值,實(shí)際值的材料性能,板厚度,摩擦狀態(tài)并不需要,因?yàn)橥度肷窠?jīng)網(wǎng)絡(luò)的多項(xiàng)式系數(shù)從沖壓力軌跡曲線擬合。</p><p> 任何實(shí)施增加粘結(jié)劑力軌跡,可以減少回彈。然而,簡單的加強(qiáng)粘結(jié)劑力軌跡,只需要兩個(gè)輸入?yún)?shù),HBF和PD%。此外,現(xiàn)有在工業(yè)上的擠壓有能力產(chǎn)生局部加強(qiáng)粘結(jié)劑力軌跡。因此,局部加強(qiáng)粘結(jié)劑力軌跡是一個(gè)理想的選擇。</p><p
24、><b> 神經(jīng)網(wǎng)絡(luò)</b></p><p> 人工神經(jīng)網(wǎng)絡(luò)已研究多年,希望能模仿人類大腦的能力,解決那些模糊的需要大量處理的問題。人類大腦為了實(shí)現(xiàn)這些數(shù)據(jù)的處理,利用大規(guī)模并行處理能力,數(shù)以百萬計(jì)的神經(jīng)元的共同努力,來解決這些復(fù)雜的問題。同樣,人工神經(jīng)網(wǎng)絡(luò)模型包含許多計(jì)算單元,稱謂的“神經(jīng)元”,以符合其生物對位,并聯(lián)運(yùn)行,并連接與變權(quán)重組合預(yù)測方法。這些重量都在適應(yīng)訓(xùn)練過程,最常
25、見的通過反傳算法(Rumelhart and McClelland, 1986),提出了神經(jīng)網(wǎng)絡(luò),例如輸入產(chǎn)出對的關(guān)系網(wǎng)絡(luò)正在試圖了解。這個(gè)目標(biāo)是神經(jīng)網(wǎng)絡(luò)推論,或引用,模式給出了輸入輸出的例子。神經(jīng)網(wǎng)絡(luò)進(jìn)一步細(xì)節(jié),一般被發(fā)現(xiàn)(Widrow and Lehr (1990)。對于我們的特殊應(yīng)用,神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)被確定為四個(gè)輸入?yún)?shù),5個(gè)隱藏的神經(jīng)元,和兩個(gè)輸出。一個(gè)S形是激活功能是用于隱藏神經(jīng)元,同時(shí)利用線性產(chǎn)出。Kinsey(1998)提出
26、了,如何優(yōu)化結(jié)構(gòu)更詳細(xì)地結(jié)論。</p><p> 最初確定利用神經(jīng)網(wǎng)絡(luò)來控制和減少回彈的可行性,神經(jīng)網(wǎng)絡(luò)的能力是處理大量不同的t和m進(jìn)行研究。板材厚度值從0.8到1.4毫米(從0.1毫米遞增至0.8或1.2毫米),摩擦系數(shù)的水平從0.04至0.20(從0.01遞增到0.04和0.12之間,及以后遞增至0.02)進(jìn)行了考慮。這些變化在t和m值明顯大于那些在實(shí)際大規(guī)模生產(chǎn)形成過程會(huì)被視作的值,但用在這里的示范目的,
27、以顯示神經(jīng)網(wǎng)絡(luò)控制甚至相當(dāng)大的變化的系數(shù)的能力。實(shí)驗(yàn)數(shù)據(jù)是通過試驗(yàn)和錯(cuò)誤模擬104組合這兩個(gè)工藝參數(shù),以確定HBF和PD%的值,這些值產(chǎn)生了回彈的角度,θ,范圍在0.4至0.5度,最大應(yīng)力,e,范圍在8%至10%。這個(gè)極其狹窄的范圍內(nèi),回彈范圍有一定的接近真實(shí)的值預(yù)測HBF和PD%的值,實(shí)驗(yàn)網(wǎng)絡(luò)可能不會(huì)導(dǎo)致回彈數(shù)值在相同的窄幅波動(dòng)。在指定的范圍給出的HBF和PD%的組合不一定是唯一一個(gè)能夠提供u和E的值。但是,通過有選擇u和e的小范圍變
28、化,我們保證網(wǎng)絡(luò)將接受一個(gè)可能的HBF和PD%的值的狹窄的窗口實(shí)驗(yàn)數(shù)據(jù)。這些104模擬運(yùn)行提供了輸入輸出對,實(shí)驗(yàn)有四個(gè)數(shù)據(jù)曲線擬合多項(xiàng)式系數(shù)從沖力軌跡作為輸入,HBF和PD%占所期望的輸出。</p><p><b> 結(jié)果</b></p><p> 一旦神經(jīng)網(wǎng)絡(luò)有更好的實(shí)訓(xùn),在網(wǎng)絡(luò)預(yù)測中4t和m組合實(shí)訓(xùn)集排除以及另外4t和m的組合不包括在以前的一系列t和m值分別為
29、“fedforward”,HBF和PD%值為加強(qiáng)局部軌跡粘結(jié)劑,以下所述的程序圖3,先前章節(jié)。由此產(chǎn)生的回彈角,u和最大應(yīng)力,e,從這一進(jìn)程中,然后計(jì)算實(shí)驗(yàn)數(shù)據(jù)。表1顯示了出色的成績,得到了這8數(shù)據(jù)結(jié)果。所有的回彈角和最大應(yīng)力值分別在0.3至0.6度和8%至10%范圍之間。</p><p> 神經(jīng)網(wǎng)絡(luò)的能力,提供HBF和PD%的值使用,面臨著t和m很大的差異顯示在實(shí)際應(yīng)用神經(jīng)網(wǎng)絡(luò)的潛力。然而,過程中還有其他參數(shù)
30、的變化,對于最后部分的形狀產(chǎn)生類似的效果。材料特性,例如,已被證明造成嚴(yán)重的三維變化的金屬薄板沖壓件(Kinsey and Cao, 1997)。當(dāng)原先的實(shí)訓(xùn)網(wǎng)絡(luò)用來預(yù)測不同的材料特性HBF和PD%值,會(huì)產(chǎn)生不符合的回彈水平和最大應(yīng)力。這并不奇怪,因?yàn)榫W(wǎng)絡(luò)的實(shí)訓(xùn)并未適應(yīng)差異的材料性質(zhì)。因此,更多的實(shí)驗(yàn)結(jié)果數(shù)據(jù)包含材料性質(zhì)是必要的。偏差的真實(shí)應(yīng)力應(yīng)變曲線的名義材料,材料1,創(chuàng)造了不同的強(qiáng)度系數(shù),K,由+10%(材料2和3)和+20%(材料
31、第6和第7)和應(yīng)變硬化部分,n,約+16%(材料4和5)。7個(gè)組合(t=0.9/u=0.04,t=1.1/u=0.06,=1.4t/=0.08,t=1.0/u=0.10,t=0.8/m=0.12t=1.0/u=0.16,與t=1.2/u=0.20)被選為該實(shí)驗(yàn)的例子與新材料的創(chuàng)建。再次通過反復(fù)試驗(yàn),某一假定材料的36種實(shí)驗(yàn)范例,t,與u組合產(chǎn)生了相同的變化范圍,比如以前使用的θ和ε,分別為0.4到0.5度和8%至10%。其中6個(gè)新型材料
32、</p><p> 表2顯示的結(jié)果的,當(dāng)這結(jié)果從被排除在實(shí)訓(xùn)集的六個(gè)組合,和4個(gè)新材料厚度,t,和摩擦系數(shù),u,在新實(shí)驗(yàn)網(wǎng)絡(luò)組合實(shí)訓(xùn)。另外,神經(jīng)網(wǎng)絡(luò)能夠提供可接受的的HBF和PD%值,即產(chǎn)生可接受θ0.2至0.6度,e,8%至10%。然而,作為獲得的t和θ的結(jié)果,這些模擬產(chǎn)出的數(shù)據(jù)并不總是在0.3至0.6度這個(gè)狹窄的回彈角范圍內(nèi)。這表明,神經(jīng)網(wǎng)絡(luò)在材料中具有很多復(fù)雜的變化。如前所述,范圍狹窄對u在實(shí)驗(yàn)中故意造成
33、被允許的合理差異的前饋過程。此外,更多的培訓(xùn),集或一個(gè)更聰明的神經(jīng)網(wǎng)絡(luò)結(jié)構(gòu),如給予“提示”網(wǎng)絡(luò)有關(guān)的材料特性,能提高能力的神經(jīng)網(wǎng)絡(luò)處理變化的材料特性。</p><p><b> 閉環(huán)控制的比較結(jié)果</b></p><p> Sunseri et al. (1996)控制系統(tǒng)提出的,利用閉環(huán)控制的粘結(jié)劑局部按照沖壓力軌跡,在同一通道的形成過程以控制回彈。關(guān)于摩擦系數(shù)
34、的可能是最重要的工藝參數(shù),不同的摩擦系數(shù)在控制系統(tǒng)在進(jìn)行測試,與數(shù)據(jù)獲得。但是,他們的工作是調(diào)查不同的材料特性和板厚。因此,進(jìn)一步閉環(huán)控制模擬與這些變化進(jìn)行了這里,以形成一個(gè)比較神經(jīng)網(wǎng)絡(luò)控制系統(tǒng)。</p><p> 首先,從我們的軌跡案名義的沖壓局部力,已創(chuàng)建的材料1,1.0mm的t,和0.10的m,。加比例積分(PI)的控制器,用于調(diào)整粘結(jié)劑力,使下面的沖壓力軌跡,Sunseriet al. (1996)。最
35、大位移之間的沖壓調(diào)整的粘結(jié)劑力為0.02毫米。表3所列的比例增加kp,以及整體調(diào)整ki,所使用的控制器的3個(gè)事件進(jìn)行數(shù)據(jù)調(diào)查。圖7顯示如何更好的追蹤沖壓力軌跡沖壓軌跡這三個(gè)事件(材料4ki50.8)。圖8顯示了如何在不同的粘合劑力的沖壓的移位,使下面的沖壓力軌跡(材料4ki50.8)。請注意,粘結(jié)劑力軌跡趨于平穩(wěn)結(jié)束時(shí)的形成過程類似于粘結(jié)劑加強(qiáng)局部的軌道。</p><p> 從表3顯示從這些閉環(huán)控制實(shí)驗(yàn)得到的
36、θ和e的值同從神經(jīng)網(wǎng)絡(luò)控制系統(tǒng)的結(jié)果聯(lián)系。此表清楚地表明,回彈角的神經(jīng)網(wǎng)絡(luò)控制系統(tǒng),大大低于和接近原來的0.4至0.5度范圍這所有三種情況。即使如標(biāo)稱t和m值,1.0毫米和0.10,分別使用的材料5,神經(jīng)網(wǎng)絡(luò)控制系統(tǒng)超越了閉環(huán)控制系統(tǒng)。</p><p> 雖然神經(jīng)網(wǎng)絡(luò)系統(tǒng)需要生產(chǎn)足夠數(shù)量的實(shí)例實(shí)訓(xùn)網(wǎng)絡(luò)作為后續(xù)工作,但是帶來的好處是回彈量和最大應(yīng)變的控制。此外,還有額外的好處,神經(jīng)網(wǎng)絡(luò)控制系統(tǒng)在這個(gè)閉環(huán)控制策略,
37、的這個(gè)的價(jià)值基與進(jìn)行調(diào)整的材料4,表3相比,材料3與5的事件相比,為了使局部軌跡沖壓應(yīng)遵循準(zhǔn)確。這就是說,作為材料和工藝參數(shù)的變化,取得與后續(xù)力軌跡密切合作是必要的但是也可能改變的。圖7還顯示本來材料4的實(shí)際沖壓力軌跡,如果k值是材料3和5,2.0中被使用。因此,當(dāng)面臨著過程的很多變化和物質(zhì)需求,這個(gè)預(yù)先選定的閉環(huán)控制系統(tǒng)的實(shí)際收獲并不是很大。目前,行業(yè)中發(fā)現(xiàn)的應(yīng)力有隨時(shí)加強(qiáng)生產(chǎn)粘結(jié)劑力軌跡的能力,這是用于神經(jīng)網(wǎng)絡(luò)控制系統(tǒng),同時(shí)隨著一個(gè)
38、不斷變化軌跡沖壓力,我們將需要更加有力的控制系統(tǒng)。</p><p><b> 結(jié)論</b></p><p> 本文提出了一種神經(jīng)網(wǎng)絡(luò)系統(tǒng),同時(shí)加強(qiáng)局部粘結(jié)劑軌跡,在模擬通道鋁成形中提出了控制回彈和最大應(yīng)變。神經(jīng)網(wǎng)絡(luò)被選用,源于其有處理高度線性問題的能力,并能發(fā)現(xiàn)金屬成形過程中不同的材料和工藝參數(shù)發(fā)生的變化。沖壓力的軌跡被確定為工藝參數(shù),提供了各種材料和工藝的變數(shù)最
39、大的可供參考的偏差值。因此,多項(xiàng)式系數(shù)從曲線擬合的沖壓力軌跡被用來作為輸入的神經(jīng)網(wǎng)絡(luò)。結(jié)果表明,為加強(qiáng)局部粘結(jié)劑軌跡神經(jīng)網(wǎng)絡(luò)成功的提供了高粘結(jié)力(HBF)和沖壓位置(PD%)值,產(chǎn)生可接受的值回彈(ü),0.2至0.6度,最大應(yīng)變(e),8%至10%,得到最終產(chǎn)品時(shí)面臨著不同的材料應(yīng)力(k)+20%,+16%的應(yīng)變硬化指數(shù)(n),+25%的金屬片厚度(t),和+65%的摩擦系數(shù)(m)。又當(dāng)Sunseri et al.(1996
40、)提出的與閉環(huán)控制戰(zhàn)略在同一過程中,神經(jīng)網(wǎng)絡(luò)系統(tǒng)被證明在變化材料特性優(yōu)于閉環(huán)系統(tǒng)而進(jìn)一步減少回彈,,如果有適當(dāng)?shù)膶?shí)訓(xùn)數(shù)據(jù),將被作為一個(gè)更加要有效的制度來執(zhí)行。</p><p> 雖然這項(xiàng)工作進(jìn)行了使用模擬,制定的方法可以很容易地延伸到實(shí)際的形成過程或?qū)嶒?yàn),但是在未來將進(jìn)行更多的實(shí)驗(yàn)以驗(yàn)證我們的需求。唯一的硬件要求是將一個(gè)有運(yùn)算能力的CPU來測量沖壓力軌跡,并有能力改變一旦循環(huán)過程中生效的粘結(jié)劑。神經(jīng)網(wǎng)絡(luò)控制系統(tǒng)
41、在一個(gè)金屬板料成形的進(jìn)程中的材料工藝參數(shù)將有力的變化;因此,建立一致局部回彈量,對接下來的流程和客戶滿意度是至關(guān)重要的。</p><p><b> 致謝</b></p><p> 這項(xiàng)研究提供了部分經(jīng)費(fèi)由美國國家科學(xué)基金會(huì)贈(zèng)款#CMS-9622271和#DMI-9703249。</p><p><b> 參考文獻(xiàn):</b&
42、gt;</p><p> ABAQUS,1997,User’s Manual,Theory Manual,Version 5.6,Hibbit, Karlsson,and Sorensen,Providence,RI.</p><p> Ayres,R.A.,1984,“SHAPESET:A Process to Reduce Sidewall Curl Springback in H
43、igh-Strength Steel Rails,” Journal of Applied Metalworking, Vol.3,No.2,pp.127–34.</p><p> Cho,S.,Cho,Y.,and Yoon,S.,1997,“Reliable Roll Force Prediction in Cold Mill Using Multiple Neural Networks,”IEEE Tra
44、nsactions on Neural Networks,Vol.8,No.4,pp.874–82.</p><p> Di,S.,and Thomson,P.F.,1997,“Neural Network Approach for Prediction of Wrinkling Limit in Square Metal Sheet Under Diagonal Tension,”Journal of Tes
45、ting and Evaluation, Vol.25,pp.74–81.</p><p> Elkins,K.L.,1994,“On-Line Angle Control for Small Radius Air Bending,”Carnegie Mellon University,UMI Dissertation Services,Ph.D.thesis.</p><p> Fo
46、rcellese,A.,Gabrielli F.,and Ruffini,R.,1997,“Springback Controlin a Air Bending Process by Neural Network,”Submitted to Proc. III Convegno AITEM.</p><p> Graf,A.,and Hosford,W.,1993,“Plane-strain Tension T
47、est of AL 2008-T4Sheets,” Proceedings SAE Congress Symposium on Sheet Forming, Detroit,MI.</p><p> Hishida,Y.,andWagoner,R., 1993,“Experimental Analysis of Blank Holding Force Control in Sheet Forming,”Jour
48、nal of Materials and Manufacturing,Vol.2,pp.409–15.</p><p> Hornik,K.,Stinchcombe,M.,and White,H.,1989,“Multilayer Feedforward Networks are Universal Approximators,”Neural Networks,Vol. 2,pp.359–66.</p&g
49、t;<p> Kinsey,B.,1998,“Process Control in Sheet Metal Forming,”Master’sthesis,Department of Mechanical Engineering,Northwestern University.</p><p> Kinsey,B.,and Cao,J.,1997,“An Experimental Study t
50、o Determine the Feasibility of Implementing Process Control to Reduce Part Variation in a Stamping Plant,” Sheet Metal Stamping: Development Applications, SAE Paper 970713, SP-1221, pp. 107–12.</p><p> Ruff
51、ini, R., and Cao, J., 1998, “Using Neural Network for Springback Minimization in a Channel Forming Process,”Developments in Sheet Metal Stamping,SAE Paper 98M-154,SP-1322,pp.77–85.</p><p> Rumelhart,D.E.and
52、 McClelland,J. L.,1986,“Learning InternalRepresentation by Error Propagation,”Parallel Distributed Processing,MIT Press,Cambridge,MA,Vol.1,Ch. 8.</p><p> Sunseri,M.,Cao,J.,Karafillis,A.P.,and oyce,M.C.,1996
53、,“Accommodation of Springback Error in Channel Forming Using Active Binder Force Control:Numerical</p><p> Sunseri,M.,Cao,J.,Karafillis,A.P.,and oyce,M.C.,1996,“Accommodation of Springback Error in Channel
54、Forming Using Active Binder Force Control: Numerical Simulations and Results,” ASME JOURNAL OF ENGINEERING MATERIALS AND TECHNOLOGY, Vol.118,No.3,pp.426–35.</p><p> Widrow,B.and Lher,M.A.,1990,“30 Years of
55、Adaptive Neural Networks: Perceptron,Madaline and Backpropagation,”Proceedings of the IEEE,Vol.78,No.9,pp.1415–41.</p><p> Yang M.,Shima, S.,and Watanabe,T.,1992,“Development of Control System Using Neural
56、Network Combined with Deformation Model for an Intelligent V-Bending Process of Sheet Metals,” Proceedings of Japan/USA Symposium on Flexible Automation, ASME,Vol.2,pp.1485–90.</p><p> 件2:外文原文(復(fù)印件)</p>
57、;<p> Consistent and Minimal Springback Using a Stepped Binder Force Trajectory and Neural Network Control</p><p> One of the greatest challenges of manufacturing sheet metal parts is to obtain cons
58、istent part dimensions. Springback, the elastic material recovery when the tooling is removed,is the major cause of variations and inconsistencies in the final part geometry. Obtaining a consistent and desirable amount o
59、f springback is extremely difficult due to the nonlinear effects and interactions between process and material parameters. In this paper, the exceptional ability of a neural network along with a st</p><p>
60、Introduction</p><p> Obtaining consistent and accurate part dimensions is crucial in today’s competitive manufacturing industry.Inconsistencies in part dimensions slow new product launches, increase changeo
61、ver times, create difficulties in downstream processes,require extra quality assurance efforts,and decrease customer satisfaction and loyalty for the final product.In the sheet metal forming process,a major factor preven
62、ting accurate final part dimensions is springback in the material.Springback is the geometric d</p><p> Springback can be reduced through modifications to the forming process.Several researchers have propos
63、ed to use a stepped binder force trajectory to accomplish this objective (Ayres,1984;Hishida and Wagoner,1993;Sunseri et al.,1996).A stepped binder force trajectory is an instantaneous jump from a low binder force (LBF)
64、value to a high binder force (HBF)at a specified percentage of the total punch displacement (PD%)(see Fig. 1).Sunseri et al.(1996)investigated springback in the Aluminum channel </p><p> In recent years, ma
65、ny research groups have investigated the use of artificial neural networks to control sheet metal forming processes. Metal forming is an ideal candidate for neural network control due to the nonlinear effects and interac
66、tions of the process parameters. Cho et al. (1997) used a neural network to predict the force in cold rolling, and Di and Thomson (1997) predicted the wrinkling limit in square metal sheets under diagonal tension. Among
67、others (Elkins, 1994 and Yang et al., 199</p><p> In this paper, the springback of an aluminum channel is controlled via a stepped binder force trajectory and neural network control. The neural network dete
68、rmines the HBF and PD% of the stepped binder force trajectory. Punch force trajectory was identified as the key parameter that reflects variations in material properties, sheet thickness, and friction coefficient. Theref
69、ore, four polynomial coefficients from curve fitting the punch force trajectory were used as inputs into the neural network.</p><p><b> . </b></p><p> A simple geometry to investig
70、ate spr ingback is a channel forming process. Therefore, the same aluminum channel forming process used by Sunseri et al. (1996) is employed here (see Fig. 2). First, the effect of constant binder force (CBF) on springba
71、ck was evaluated. As the CBF was increased while all other process and material parameters were held constant in our simulations, the springback angle u, was reduced as shown graphically in Fig. 4 and physically in Fig.
72、5. However, the increased binde</p><p> To produce a robust process when faced with deviations in the friction coefficient, Sunseri et al. (1996) implemented closed-loop variable binder force control to fol
73、low the punch force trajectory obtained from the stepped binder force case with nominal process conditions. This control method was able to produce consistent springback levels when the friction coefficient was varied fr
74、om a value of 0.1, for the nominal case, to 0.25. However, whether this methodology could withstand variations of </p><p> Finite Element Model. </p><p> A commercial Finite Element Analysis p
75、ackage (ABAQUS, 1997) was used for our numerical simulations of the channel forming process. Since the problem is close to a plane-strain condition and is symmetric, only one sixteenth of the width and half of the length
76、 of the entire blank (220 mm 3 46 mm) was modelled. The binder, the die, and the punch were modelled as three separate rigid surfaces. Each surface was modelled using four-node interface elements (ABAQUS type IRS4), and
77、a Coulomb friction law</p><p> Proposed Control System</p><p> In a channel forming process, springback in general is extremely sensitive to variations in material and forming parameters. In t
78、his work, we develop a methodology for controlling springback while producing an acceptable amount of maximum strain in the material through a combination of a stepped binder force trajectory and neural network control.
79、In a stepped binder force trajectory, two critical values need to be determined, the magnitude of the HBF and the PD% of the total punch displacement. </p><p> A flowchart of our proposed control system is
80、shown in Fig. 3. The forming process would proceed normally at a constant initial binder force, 16 kN, to a depth of 10 mm. While the punch displacement continues, the polynomial coefficients from curvefitting of the pun
81、ch force trajectory are calculated and fed into a neural network. The outputs from the neural network, the HBF and PD% for the stepped binder force trajectory, would be obtained in time to make the appropriate HBF adjust
82、ment at the sp</p><p> The value of 10 mm, where the coefficients from the punch force trajectory are calculated, was chosen for two reasons. At this distance, the punch force trajectory is well defined and
83、 enough data points are available for accurate curve fitting to occur. Figure 6 shows the effect variations in t and m have on the punch force trajectory from 0 to 10 mm. Secondly, the distance of 10 mm allows adequate C
84、PU time for calculations to be conducted. Assuming a punch speed of 50 mm/sec and setting the min</p><p> The punch force trajectory was broken up into three regions, Region A, a transition region, and Regi
85、on B (see Fig. 6). The groups of lines in Region A correspond to varying values of sheet thickness, and a second-degree polynomial was used to fit the data in this region. Despite the fact that it was a second degree pol
86、ynomial, only two input coefficients were used to characterize this region since the punch force trajectory nearly passed through the origin. In Region B, the dissimilar slopes are</p><p> Note that the pun
87、ch force trajectories are obtained from numerical simulations; therefore, they are smooth and have consistent patterns due to variations in t and m. In an actual forming process, noise in the data acquisition equipment w
88、ould produce disparities in these curves. By using polynomial coefficients from curve fitting the punch force trajectory, the inputs into the neural network are in essence filtered to account for these practical consider
89、ations.</p><p> To implement this control system in an actual forming operation, training data would need to be produced. This information could be derived from actual production of stampings during die try
90、 out by varying process parameters. For instance, to account for batch to batch variation in the material, material could be obtained and formed which represents the full range of potential sheet thicknesses and material
91、 properties. Also, the lubrication state could be varied. Using various combinations of th</p><p> Any increasing binder force trajectory could be implemented to reduce springback. However, the simplicity o
92、f the stepped binder force trajectory requires only two input parameters, HBF and PD%. In addition, existing presses in industry have the ability to produce a stepped binder force trajectory. Therefore, a stepped binder
93、force trajectory is an ideal choice.</p><p> Neural Networks.</p><p> Artificial neural networks have been studiedfor many years in the hope of mimicking the human brain’s ability to solve pro
94、blems that are ambiguous and require a large amount of processing. Human brains accomplish this data processing by utilizing massive parallelism, with millions of neurons working together to solve complicated problems. S
95、imilarly, artificial neural network models consist of many computational elements,called “neurons” to correspond to their biological counterparts,operating in</p><p> To initially determine the feasibility
96、of using a neural network to control and minimize springback, the ability of the neural network to handle large variations in the t and m was investigated. Sheet thickness values from 0.8 to 1.4 mm (in increments of 0.1
97、mm between 0.8 and 1.2 mm) and friction coefficient levels from 0.04 to 0.20 (in increments of 0.01 between 0.04 and 0.12 and afterwards in increments of 0.02) were considered. These variations in t and m values are obvi
98、ously larger than wha</p><p><b> Results</b></p><p> Once the neural network was well trained, four t and m combinations left out of the training set as well as four additional t a
99、nd m combinations not included in the previous range of t and m values were “fedforward” in the network to predict HBF and PD% values for the stepped binder force trajectory following the procedure outlined in Fig. 3 and
溫馨提示
- 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
- 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
- 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會(huì)有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
- 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
- 5. 眾賞文庫僅提供信息存儲(chǔ)空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負(fù)責(zé)。
- 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請與我們聯(lián)系,我們立即糾正。
- 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時(shí)也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。
最新文檔
- 外文翻譯--加強(qiáng)使用粘結(jié)劑力軌跡和神經(jīng)網(wǎng)絡(luò)控制獲得一致的最小回彈.doc
- 外文翻譯--加強(qiáng)使用粘結(jié)劑力軌跡和神經(jīng)網(wǎng)絡(luò)控制獲得一致的最小回彈.doc
- 玻璃纖維棉氈粘結(jié)劑中柔性劑研究及神經(jīng)網(wǎng)絡(luò)優(yōu)化.pdf
- 粘結(jié)劑作用
- 基于神經(jīng)網(wǎng)絡(luò)的篇章一致性建模.pdf
- 粘結(jié)劑的分類
- 口腔粘結(jié)劑
- 涂料與粘結(jié)劑
- 人工神經(jīng)網(wǎng)絡(luò)外文翻譯
- 外文翻譯---神經(jīng)網(wǎng)絡(luò)概述
- 外文翻譯---人工神經(jīng)網(wǎng)絡(luò)
- 使用rbf神經(jīng)網(wǎng)絡(luò)進(jìn)行優(yōu)化冷藏庫的控制外文文獻(xiàn)翻譯
- 結(jié)合相位一致性和自適應(yīng)脈沖耦合神經(jīng)網(wǎng)絡(luò)的圖像檢索.pdf
- 使用RBF神經(jīng)網(wǎng)絡(luò)進(jìn)行優(yōu)化 外文文獻(xiàn)翻譯.doc
- 使用RBF神經(jīng)網(wǎng)絡(luò)進(jìn)行優(yōu)化 外文文獻(xiàn)翻譯.doc
- 使用RBF神經(jīng)網(wǎng)絡(luò)進(jìn)行優(yōu)化 外文文獻(xiàn)翻譯.doc
- 天然石材專用粘結(jié)劑
- 外文翻譯--RBF 神經(jīng)網(wǎng)絡(luò).doc
- 羊肉無磷保水劑和粘結(jié)劑的研究.pdf
- 外文翻譯--RBF 神經(jīng)網(wǎng)絡(luò).doc
評論
0/150
提交評論