外文翻譯--智能沖壓工藝規(guī)劃系統(tǒng)的研究

1、<p>　　智能沖壓工藝規(guī)劃系統(tǒng)的研究</p><p>　　摘要：本文對(duì)建立一個(gè)智能沖壓工藝設(shè)計(jì)知識(shí)為基礎(chǔ)的系統(tǒng)給出了一個(gè)簡(jiǎn)單的介紹。研究該系統(tǒng)的框架，對(duì)模型和知識(shí)推理模式進(jìn)行了介紹。對(duì)有些關(guān)鍵技術(shù)如沖壓工藝的可行性、排樣的最佳算法、智能地帶的布局和內(nèi)力計(jì)算進(jìn)行了研究。該系統(tǒng)可以改善工藝規(guī)劃效率。</p><p>　　關(guān)鍵詞： 排樣 KBS 知識(shí)模型 帶狀排樣法&l

2、t;/p><p>　　1簡(jiǎn)介 沖壓工藝規(guī)劃是沖壓產(chǎn)品開發(fā)的一個(gè)核心項(xiàng)目。它是金屬成型應(yīng)用的一個(gè)重要組成部分，它與生產(chǎn)質(zhì)量、成本、生產(chǎn)率和工具壽命有直接的影響?，F(xiàn)代制造業(yè)的快速發(fā)展對(duì)沖壓提出了更高的要求，尤其是在沖壓工藝方面。多年來(lái)，相關(guān)研究已就如何在創(chuàng)新的環(huán)境加強(qiáng)工藝規(guī)劃的集成化和智能化程度進(jìn)行研究。近年來(lái)，通過(guò)生產(chǎn)金屬成形智能設(shè)計(jì)系統(tǒng)、自動(dòng)化技術(shù)，整和了工藝規(guī)劃的原則。智能工藝規(guī)劃方法可以有效地提高設(shè)計(jì)效率

3、與質(zhì)量、創(chuàng)新設(shè)計(jì)能力。</p><p>　　[1].對(duì)于冷鍛序列的設(shè)計(jì)開發(fā)了一種基于PC的專家系統(tǒng)，該系統(tǒng)采用基于塑性理論和實(shí)際考慮的規(guī)則。在美國(guó)俄亥俄州立大學(xué)一個(gè)稱作FORMEX的規(guī)則系統(tǒng)被Altan和他的同事們寫入多級(jí)冷鍛的工藝規(guī)劃程序語(yǔ)言中。[2].它依靠冷鍛零件各種形狀的廣泛分類。[3] 實(shí)施以知識(shí)為本的冷成形序列設(shè)計(jì)系統(tǒng)，采用設(shè)計(jì)規(guī)則確定建立一個(gè)可行的序列，然后使用有限元分析優(yōu)化這個(gè)序列。一個(gè)以知識(shí)為基

4、礎(chǔ)的模具設(shè)計(jì)自動(dòng)化系統(tǒng)被Cheok和他的同事精心設(shè)計(jì)出來(lái)。[4] 在新加坡國(guó)立大學(xué)。一些零件表象技術(shù)、沖壓零件識(shí)別和模具構(gòu)成也存在于這項(xiàng)工作中。在中國(guó)，華中科技大學(xué)的科學(xué)技術(shù)研究者們也開發(fā)出了基于知識(shí)系統(tǒng)的用于對(duì)小型金屬件沖壓級(jí)進(jìn)模的程序包。[5].使用特點(diǎn)，用戶可以在3D立體構(gòu)架下設(shè)計(jì)產(chǎn)品。在手工設(shè)置排樣后，用戶可以使用交互命令來(lái)開發(fā)帶裝布局設(shè)計(jì)。來(lái)自利物浦大學(xué)工業(yè)研究部門的研究者們也在研究沖壓工藝和沖裁模的專用系統(tǒng)。[6].他們的研

5、究集中在分解較小的橋型廢料的形狀編碼和識(shí)別技術(shù)。[7]在上海沖壓模具和工具技術(shù)研究所的研究者們也開發(fā)出了級(jí)進(jìn)模的CAD/CAM系統(tǒng)。他們研究的該系統(tǒng)依靠特殊的相關(guān)數(shù)據(jù)來(lái)描繪工件和模具結(jié)構(gòu)。</p><p>　　上述研究的研究工作的目的是為了促進(jìn)金屬成形的發(fā)展。從金屬智能成型的回顧和分析中，使用智能設(shè)計(jì)的理論和方法來(lái)研究沖壓工藝規(guī)劃的步驟。在本文中介紹了應(yīng)用于沖壓工藝規(guī)劃的智能的系統(tǒng)。該智能系統(tǒng)在處理一些復(fù)雜的設(shè)計(jì)

6、問(wèn)題時(shí)是種強(qiáng)有力的工具。由專門知識(shí)構(gòu)成的智能系統(tǒng)可以用一種交互的方式協(xié)助用戶解決各種各樣的問(wèn)題或疑問(wèn)。[8].智能系統(tǒng)是一種計(jì)算機(jī)系統(tǒng)，它試圖代表人類知識(shí)和專業(yè)知識(shí), 以一種實(shí)際和有效的途徑提供快捷、方便的知識(shí)。智能系統(tǒng)能夠完成一般需要專家才能完成的任務(wù)。它能自動(dòng)化實(shí)時(shí)利用現(xiàn)有的專業(yè)知識(shí),并解釋它的推理過(guò)程。沖壓工藝規(guī)劃是一個(gè)含有豐富知識(shí)的復(fù)雜設(shè)計(jì)過(guò)程。整合在沖壓工藝規(guī)劃設(shè)計(jì)中智能系統(tǒng)的關(guān)鍵技術(shù)是至關(guān)重要的。使用智能理論的沖壓工藝規(guī)劃智

7、能系統(tǒng)被提出來(lái)。對(duì)一些關(guān)鍵技術(shù),如集成產(chǎn)品知識(shí)建模和戰(zhàn)略規(guī)劃的綜合沖壓成形過(guò)程進(jìn)行了研究。在沖壓設(shè)計(jì)中包括各種各樣的知識(shí)，如專業(yè)領(lǐng)域知識(shí)、多任務(wù)知識(shí)、非標(biāo)準(zhǔn)知識(shí)。每一種知識(shí)都需要集成到該系統(tǒng)中。沖壓模具的核心是沖壓工藝。必須考慮到多種因素,如幾何形狀、技術(shù)要求、材料性能、沖壓件的可行性、工作程序安排、模具工具的結(jié)構(gòu)。沖壓工藝規(guī)劃是一種基于專家知識(shí)的創(chuàng)造</p><p><b>　　2系統(tǒng)構(gòu)架和框架<

8、;/b></p><p>　　智能系統(tǒng)的關(guān)鍵技術(shù)是建立和應(yīng)用的信息化模型制作。。該產(chǎn)品信息模型,包括三個(gè)階段：一種基于幾何的模型、一種基于特征的模型、一種基于智能的模型?；趲缀蔚哪Ｐ兔枋隽慵膸缀瓮?fù)湫畔ⅰＳ捎诹慵臄?shù)據(jù)信息不能被完整的描述、數(shù)據(jù)分離水平太低，幾何模型被特征模型取代。這個(gè)信息模型包括一組幾何實(shí)體。依靠此模型的工程語(yǔ)義模型，許多與設(shè)計(jì)相關(guān)的功能可以被實(shí)現(xiàn)。隨著人工智能的發(fā)展，智能模型開始被應(yīng)

9、用。專業(yè)知識(shí)、設(shè)計(jì)過(guò)程的知識(shí),和相關(guān)的知識(shí)都包含在知識(shí)模型中[9、10]。智能模型支持表達(dá)和傳遞有用的信息。</p><p>　　本文主要概括了一種沖壓工藝規(guī)劃的智能系統(tǒng)。該智能系統(tǒng)對(duì)產(chǎn)品的定義有效且完整。它幾何了不同模型的優(yōu)點(diǎn)且能滿足幾何設(shè)計(jì)和推理過(guò)程。面向?qū)ο蠹夹g(shù)應(yīng)用到整合各種各樣的知識(shí)。此集成的知識(shí)系統(tǒng)模型可被共享和用于智能設(shè)計(jì)和產(chǎn)品信息溝通。</p><p>　　這個(gè)關(guān)于沖壓模具工

10、藝規(guī)劃的智能系統(tǒng)構(gòu)架已經(jīng)被設(shè)計(jì)出來(lái)。這個(gè)零件的結(jié)構(gòu)設(shè)計(jì),包括一個(gè)圖形用戶界面,一個(gè)應(yīng)用程序系統(tǒng)、設(shè)計(jì)資源、知識(shí)工具,混合推理機(jī)制、基礎(chǔ)模型。在這個(gè)構(gòu)架中知識(shí)模型有不同的分類。知識(shí)模型從設(shè)計(jì)資源中獲取有用的信息，支持知識(shí)獲取和知識(shí)表達(dá)的程序。這個(gè)模型把有用信息轉(zhuǎn)移到知識(shí)庫(kù)。知識(shí)庫(kù)由CAD軟件支持。設(shè)計(jì)結(jié)果以3D模型、圖畫和資料庫(kù)的形式保存在知識(shí)庫(kù)中，它對(duì)在知識(shí)庫(kù)中不同零件的知識(shí)傳遞來(lái)說(shuō)非常的重要。</p><p>

11、<b>　　3 實(shí)施方法和應(yīng)用</b></p><p>　　3.1沖壓智能模型的可行性論證</p><p>　　智能系統(tǒng)對(duì)沖壓工件的質(zhì)量、成本、模具壽命進(jìn)行評(píng)價(jià)。該評(píng)價(jià)基于成熟的智能模型。此模型集成了規(guī)則庫(kù)、零件信息和結(jié)論庫(kù)。系數(shù)根據(jù)知識(shí)規(guī)則推理在知識(shí)庫(kù)得出。沖壓成型可行性可以從信息庫(kù)中零件信息和相關(guān)系數(shù)推出。在設(shè)計(jì)過(guò)程中被新結(jié)論擴(kuò)大的結(jié)果保存在結(jié)論庫(kù)中。</p

12、><p>　　模型的智能推理過(guò)程和零件的規(guī)格相比有一定限度范圍的工藝參數(shù)。此規(guī)格</p><p>　　包括輸入輸出半徑、孔徑、孔板、孔網(wǎng)、槽、槽網(wǎng)。結(jié)果來(lái)證實(shí)零件的形狀是否符合模具工具加工。智能推理用于自動(dòng)和交互的方式。這樣做的目的是來(lái)研究沖壓該產(chǎn)品的可行性。智能推理的關(guān)鍵是確定基于零件厚度和相關(guān)系數(shù)的加工極限值。圖二所示為產(chǎn)品可行性論證模型的流程圖。</p><p>

13、　　知識(shí)規(guī)則和設(shè)計(jì)結(jié)果保存在機(jī)械推理的數(shù)據(jù)庫(kù)中。零件的形狀可以在知識(shí)模型中修改。</p><p>　　由知識(shí)模型決定的沖壓工藝規(guī)劃是非常重要的一步，它同時(shí)也提供了選擇一個(gè)單步工序刀具或是復(fù)合工具或是一個(gè)改進(jìn)工具的方法。各種不同領(lǐng)域的知識(shí)、經(jīng)驗(yàn)和專業(yè)知識(shí)都被保存在工藝規(guī)劃專業(yè)系統(tǒng)中。</p><p>　　知識(shí)庫(kù)的發(fā)展是基于規(guī)則表達(dá)的共同原則。這一步的目的是集成專業(yè)經(jīng)驗(yàn)和零件</p>

14、;<p><b>　　的形狀</b></p><p>　　3.2基于優(yōu)化算法的智能排樣模型</p><p>　　為了達(dá)到較高的材料利用率，空白的知識(shí)模型被建立，保存在知識(shí)庫(kù)中的結(jié)果是其他模塊建立的基礎(chǔ)。</p><p>　　在知識(shí)庫(kù)中有四種排樣類型：</p><p><b>　　一排列布局模式&l

15、t;/b></p><p><b>　　與一排列相對(duì)的模式</b></p><p><b>　　兩排列布局模式</b></p><p>　　與兩排列布局相對(duì)的模式</p><p>　　建立這個(gè)知識(shí)模型的目的是改善材料的利用。由知識(shí)庫(kù)提供的限制情況可以由人類專家來(lái)選擇。這個(gè)知識(shí)模型控制著整個(gè)排樣

16、的設(shè)計(jì)過(guò)程。</p><p>　　圖三所示為平面布局的等級(jí)體系結(jié)構(gòu)</p><p>　　第一種模式的作用是選擇粗略數(shù)值和計(jì)算工作區(qū)域的總體輪廓。此模式提供了原始參數(shù)。粗略數(shù)值的全部信息都由此得到，不管這個(gè)數(shù)字是否被概略畫出或是被選中。</p><p>　　第二種模式用來(lái)確定布局類型、角度范圍、布局大小和條帶區(qū)的寬度。</p><p>　　第三

17、種模式中應(yīng)用了優(yōu)化算法。設(shè)計(jì)結(jié)果包括材料利用率、材料寬度和每步間隙都被保存在此模式中，不同布局的繪圖也同時(shí)生成。</p><p>　　在第四種模式中可以修改布局規(guī)劃的結(jié)果。最終參數(shù)包括每步間隙、材料寬度、各類網(wǎng)格和轉(zhuǎn)換能力。當(dāng)參數(shù)有所改變時(shí)，布局規(guī)劃圖可以被更新。</p><p>　　該知識(shí)的主要作用是布局規(guī)劃的算法優(yōu)化。該算法共有六步。</p><p>　　1.在

18、圖形周圍最適合的矩形第一次生成。復(fù)制件和原件之間的距離是包含在接洽網(wǎng)中的。圖四說(shuō)明了此種算法。</p><p>　　2.在兩個(gè)環(huán)形中間的值是經(jīng)過(guò)計(jì)算的。這兩個(gè)環(huán)形分解成線和圓弧的單元。每對(duì)元素中間的距離需要重新補(bǔ)償。然后就可以找到最短的距離。</p><p>　　3.計(jì)算出的最小值和所要求的值之間的差異就是誤差。當(dāng)誤差小于允許值時(shí)，排樣規(guī)劃就可以完成。另外，布局圖形需要沿著視野的方向移動(dòng)。

19、</p><p>　　4.材料利用率可以以布局規(guī)劃的角度上被計(jì)算出來(lái)。</p><p>　　5.排樣圖形旋轉(zhuǎn)一定的角度。旋轉(zhuǎn)中心是矩形中心點(diǎn)附近的粗略數(shù)值。材料利用率在當(dāng)前角度下被計(jì)算出來(lái)。</p><p>　　6.排樣圖形旋轉(zhuǎn)到另外一個(gè)角度。重復(fù)第三部的的步驟，直到角度達(dá)到180度。 3.3帶狀布局的開發(fā)</p><p>　　帶狀布局的工

20、序規(guī)則被集成于知識(shí)基礎(chǔ)級(jí)進(jìn)刀具設(shè)計(jì)。該智能模型的功能是：選擇零件位置，設(shè)計(jì)方位和安排帶狀工步距離。為了解決運(yùn)行程序，該規(guī)則應(yīng)該被制定的合理和有效。</p><p>　　自動(dòng)設(shè)計(jì)模塊是智能模型中最重要的模塊。人工智能技術(shù)被應(yīng)用于此模塊中。此模型中的預(yù)處理模塊，包括定位產(chǎn)品模塊和從產(chǎn)品模塊中提取精確的信息。為了在修改模塊中生成一個(gè)模型，最初的設(shè)計(jì)工程被修改[11]。被修改的模塊代替了處理模塊。</p>

21、<p>　　3.3.1 自動(dòng)帶狀布局設(shè)計(jì)的預(yù)處理</p><p>　　1）確定零件的位置和排列。用戶可以用界面來(lái)確定預(yù)處理模塊中的一些參數(shù)。確定位置的過(guò)程可以和其他元素一起來(lái)做，例如：零件形狀、尺寸精度、和用戶要求。</p><p>　　零件的形狀也在智能模型中定義，結(jié)果被保存在知識(shí)庫(kù)中。</p><p>　　2）獲取零件精確信息。此精確信息應(yīng)該在帶狀布局

22、知識(shí)庫(kù)中得到。有用的信息包括沖孔的精確信息和相對(duì)位置信息。由此種類型信息組成的知識(shí)模型將會(huì)決定零件的沖壓順序。這個(gè)設(shè)計(jì)過(guò)程的主要要求是為位置精度開發(fā)一種知識(shí)模型[12]。首先，零件的形狀被分成封閉的輪廓。輪廓的數(shù)目為n</p><p>　　K = {k1, k2, . . ., ki, . . ., kn} (1)</p><p>　　這里 ki 表示零件的第i個(gè)輪廓。

23、所有輪廓間的相對(duì)關(guān)系包含在關(guān)系P中。如果在輪廓ki 和kj 之間要求精準(zhǔn)，這里存在(ki , kj) ∈ p。</p><p>　　p = {. . ., (ki , kj), . . .} ki , kj ∈ K, 1 ≤ i, j ≤ n(i _= j). (2)</p><p>　　每種類型的精確信息通過(guò)相關(guān)矩陣被保存在知識(shí)模型中。</p><p>　　

24、3.3.2帶狀布局自動(dòng)設(shè)計(jì)</p><p>　　帶狀布局的自動(dòng)設(shè)計(jì)模塊在知識(shí)模型中是最重要的一個(gè)。在知識(shí)模型中包含很多重要的規(guī)則，例如在一次單沖程中沖壓所有內(nèi)輪廓比較好。在下一個(gè)階段這個(gè)部分被切斷。有時(shí)候，如果沖壓點(diǎn)之間的距離非常小，一些內(nèi)輪廓就要被搬到下一階段進(jìn)行加工。如果沖壓點(diǎn)離分餾點(diǎn)太近的話，分餾點(diǎn)就需要被更改到下一階段。如果這里仍然有不合適的尺寸，一些點(diǎn)可以被移動(dòng)到下一階段。重復(fù)整個(gè)過(guò)程直到矩陣點(diǎn)間的每個(gè)

25、尺寸都可以被接受。布局智能設(shè)計(jì)的核心是開發(fā)干涉點(diǎn)的智能模型[13]。</p><p>　　零件坯料被分成許多點(diǎn)的形式。這些點(diǎn)的名字是k1, k2, . . ., kn. 這里dij </p><p>　　是ki和 kj 之間最小的距離。矩陣的臨界值是S。如果dij<S ，ki和 kj</p><p>　　不能在相同的步驟中得出。這種情況是智能模型中兩個(gè)點(diǎn)的沖突

26、。開發(fā)干涉點(diǎn)的智能模型的目的是確定沖突點(diǎn)的存在。此矩陣是一個(gè)系統(tǒng)矩陣。為了使設(shè)計(jì)過(guò)程更方便，可以把矩陣中的上半部分元素置零。</p><p>　　此處，μij 是關(guān)聯(lián)系數(shù)，它表示了每對(duì)點(diǎn)之間的不同關(guān)系。如果兩個(gè)點(diǎn)之間有沖突，它們中的一個(gè)則要被移到下一步。在每一步中重復(fù)上述步驟直到?jīng)_突點(diǎn)消失。最后矩陣M成為空矩陣。</p><p>　　3.3.3對(duì)帶狀布局結(jié)果的處理</p>&

27、lt;p>　　帶狀布局的子處理知識(shí)模型中有兩部分：修改結(jié)果和創(chuàng)建布局圖形。從帶狀布局自動(dòng)設(shè)計(jì)模型中得出的結(jié)果是慣用的。它們可能滿足不了用戶的所有要求。依靠知識(shí)模型的數(shù)據(jù)結(jié)構(gòu)，通過(guò)移動(dòng)點(diǎn)和改變步驟，增加空步和刪除空步的目的可以被實(shí)現(xiàn)。我們能夠通過(guò)處理步驟的數(shù)據(jù)結(jié)果來(lái)修改帶狀布局的設(shè)計(jì)結(jié)果。工步改變可以通過(guò)交換兩個(gè)位置的編碼來(lái)實(shí)現(xiàn)，工步增加或減少可以通過(guò)插入或移除編碼的操作來(lái)完成。當(dāng)我們想移動(dòng)一些點(diǎn)時(shí)，我們可以從第一步到最后一步轉(zhuǎn)移鏈

28、表中相當(dāng)?shù)狞c(diǎn)。</p><p>　　3.4確定沖壓中心和力計(jì)算的智能模型。</p><p>　　沖壓中心設(shè)計(jì)模型的目的是建立組合力的工作點(diǎn)[11]。模具工具中心和沖壓中心的一致非常重要，只有那樣沖壓工具才能在一起正常的工作。沖壓中心從知識(shí)模型的每一個(gè)輪廓位置的計(jì)算中得出。設(shè)計(jì)的第一部是得到工具的工作區(qū)域。CAD平臺(tái)上的零件圖形的輪廓提供了零件的外矩形。依靠沖壓中心和外矩形之間的關(guān)系可以生成

29、工作區(qū)域。因?yàn)椴黄胶饬Φ慕Y(jié)果的可能性，同時(shí)也提供了沖壓中心的再生成。再生成的步驟由人機(jī)接口軟件來(lái)完成。圖八所示為復(fù)合模打孔機(jī)工 作區(qū)域的設(shè)計(jì)結(jié)果。</p><p>　　保存在知識(shí)庫(kù)中的內(nèi)容包括模具工具的每種類型、零件落料、廢料移除等等。不同情況下的力計(jì)算的方法是不同的。力方程是由知識(shí)規(guī)則庫(kù)的推理得到的。首先，加工力和切削力是基于零件的輪廓長(zhǎng)度和知識(shí)庫(kù)中的知識(shí)規(guī)則得到的。然后，通過(guò)設(shè)計(jì)結(jié)果和合零件情況，可以得到

30、脫離力、阻力和推件力?？偟牧Π凑罩R(shí)庫(kù)中的導(dǎo)向一步一步計(jì)算。</p><p>　　4 結(jié)論和進(jìn)一步工作</p><p>　　計(jì)算機(jī)輔助設(shè)計(jì)工具的應(yīng)用在金屬成型中的應(yīng)用，節(jié)省了大量的時(shí)間和金錢。由于復(fù)雜零件沖壓工藝設(shè)計(jì)的復(fù)雜性，開發(fā)一種自動(dòng)生成工藝步驟的系統(tǒng)非常重要。這個(gè)研究開發(fā)了一個(gè)集成的CAD系統(tǒng)，該系統(tǒng)開發(fā)了一種工藝規(guī)劃系統(tǒng)使對(duì)不規(guī)則零件在高速下進(jìn)精密加工得以實(shí)現(xiàn)。該系統(tǒng)有一下特點(diǎn)：&

31、lt;/p><p>　　在設(shè)計(jì)過(guò)程中不斷改變的數(shù)據(jù)以不同的方式保存，包括數(shù)字形式和圖片形式的。用戶在設(shè)計(jì)過(guò)程中可以自由使用它們作為參考。</p><p>　　加工可行性檢查模型檢查沖壓的可行性，同時(shí)能對(duì)復(fù)雜零件的沖壓工藝規(guī)劃提供一些建議。</p><p>　　排樣模塊生成最佳排樣圖以到達(dá)材料的最大利用率。產(chǎn)品成本的減少取決于排樣最優(yōu)化計(jì)算。不僅最佳規(guī)劃而且每個(gè)合理的規(guī)劃

32、被保存在知識(shí)庫(kù)中。用戶可以選擇任意一個(gè)作為它們的最終設(shè)計(jì)結(jié)果。</p><p>　　帶狀排樣模塊生成自動(dòng)工藝規(guī)劃圖。根據(jù)用戶的要求帶狀排樣的結(jié)果可以在設(shè)計(jì)過(guò)程的任意時(shí)期修改。</p><p>　　在工藝規(guī)劃中協(xié)助設(shè)計(jì)者的此系統(tǒng)將會(huì)是一種有用的工具。它將會(huì)足夠的靈活允許設(shè)計(jì)者具有創(chuàng)造性，同時(shí)用計(jì)算機(jī)來(lái)執(zhí)行幾何計(jì)算和自動(dòng)得到設(shè)計(jì)結(jié)果。它提供了一個(gè)非常靈活的設(shè)計(jì)環(huán)境，用戶可以完全掌握即使是復(fù)雜零

33、件的沖壓工藝規(guī)劃設(shè)計(jì)。該系統(tǒng)擁有圖形交互界面，用戶可以在設(shè)計(jì)過(guò)程中交互式地改變各種設(shè)計(jì)參數(shù)。</p><p>　　進(jìn)一步的工作將會(huì)集中在排樣優(yōu)化的效率改善上，優(yōu)化用時(shí)將會(huì)減少。為排樣規(guī)劃，更多的設(shè)計(jì)方案的類型應(yīng)該被添加到知識(shí)模型中。根據(jù)沖壓工藝規(guī)劃的結(jié)果，沖壓模具設(shè)計(jì)應(yīng)用也將會(huì)在進(jìn)一步的工作中被研究。</p><p>　　Intelligent stamping process plann

34、ing system research </p><p>　　Abstract: this paper to build a intelligent stamping process design knowledge based system gives a brief introduction. Study the system framework of model, and knowledge reasoni

35、ng model are introduced. For some key technologies such as the feasibility of stamping process and exhaust kind of best algorithm, intelligent zone layout and internal force calculation is studied. This system can improv

36、e process planning efficiency. </p><p>　　Keywords: arrangement a KBS knowledge model ribbon arrangement method </p><p>　　introduction </p><p>　　Stamping process planning is one of t

37、he core punching product development project. It is the metal molding application an important component part of it and the production quality, cost, productivity and tool life have direct influence. The rapid developmen

38、t of modern manufacturing for stamping put forward higher request, especially in stamping process. For many years, the related research has the environment in innovation to strengthen the integration and process planning

39、 study intelligentize d</p><p>　　[1]. For cold forging sequence of designing and developing a kind of expert system based on PC, this system based on practical considerations plasticity theory and the rules.

40、 At the Ohio state university, a called FORMEX rules system is Altan and his colleagues write multistage cold forging process planning and programming language. [2]. It depends on a cold forging parts of various shapes w

41、idely classification. [3] implement knowledge based cold forming sequence design system, adopting the desi</p><p>　　The findings of the research work purpose is to promote the development of metal forming pr

42、ocess. From metal intelligence review and analysis of the forming of intelligent design, use the theory and method to study the stamping process planning steps. In this paper introduces applied in stamping process planni

43、ng of intelligent system. This intelligent system in dealing with some complex design problem is a powerful tool. By special knowledge construction intelligent systems can use an interactiv</p><p>　　2 system

44、 frame and the frame </p><p>　　Intelligent system key technology is built and application of information model making. This product information model, including three stages: a kind of the model, based on ge

45、ometric model based on features based on intelligence, the model. Based on geometric model describing the geometric topology information parts. Because parts of data message cannot be fully described, data separation lev

46、el is too low, geometric model was characteristic model replaced. This information model includes a set </p><p>　　This paper mainly summarizes a stamping process planning of intelligent system. This intellig

47、ent system for product definition effective and complete. It has the advantages of different geometrical model and can satisfy the geometric designs and reasoning process. Object-oriented technology is applied to integra

48、te all kinds of knowledge. This integrated knowledge system model can be Shared and used in intelligent design and product information communication. Figure 1 shows the stamping process p</p><p>　　This about

49、 stamping mould process planning of intelligent system frame has been designed. The components of the structure design, including a graphical user interface, an application system, design resources, knowledge tool, mixed

50、 reasoning mechanism, basic model. In this architecture knowledge model have different classification. Knowledge model from the design resource to extract useful information, support knowledge acquisition and knowledge e

51、xpression program. This model is useful informatio</p><p>　　3 implementation method and application </p><p>　　3.1 stamping feasibility of intelligent model </p><p>　　Intelligent sys

52、tem for stamping workpiece quality, cost, die life is evaluated. This evaluation based on mature intelligent model. This model has integrated rule library, parts information and conclusion library. Coefficient of knowled

53、ge rule reasoning in knowledge according to that. Stamping forming feasibility can from a database of information and related coefficient parts launch. In the design process of the extension of the new conclusion preserv

54、ed in conclusion library. </p><p>　　Model of intelligent reasoning process and parts specification limits range compared with the technological parameters. This specification </p><p>　　Including

55、 input/output radius, aperture, orifice plate, hole nets, chamfer, trough nets. Results to confirm whether accord with the shape of mould parts processing tools. Intelligent reasoning is used in the automatic and interac

56、tive way. It's purpose is to study the feasibility of pressing the product. Intelligent reasoning based on the key is to determine the thickness and the correlation coefficient parts processing limit. Figure 2 shows

57、the feasibility of the model for product flow chart. </p><p>　　Figure 2 shows the feasibility of the model for product flow chart. </p><p>　　Knowledge rules and design results stored in the data

58、base of mechanical reasoning. Parts in knowledge model shape can modify. </p><p>　　The decision by the knowledge model stamping process planning is very important step, it also provides to choose a single st

59、ep process tool or composite tools or a method of improvement tools. All sorts of different domain knowledge, experience and expertise are kept in the process planning of professional system. </p><p>　　Based

60、 on the development of knowledge base is the common principle rules expression. The purpose of this step is to integrate professional experience and parts </p><p><b>　　shape </b></p><p

61、>　　3.2 based on optimization algorithm of intelligent strip layout model </p><p>　　In order to achieve higher material utilization, blank knowledge model was established, the results are stored in knowled

62、ge base established basis other modules. </p><p>　　In the knowledge base there are four arrangement type: </p><p>　　Arranged layout pattern determined </p><p>　　With an array of Was

63、hington relative pattern </p><p>　　Second-ranking arranged layout mode two </p><p>　　With two second-ranking arranged layout relative mode </p><p>　　The purpose of establishing the

64、knowledge model is to improve the material utilization. The restrictions by knowledge can provide human experts to choose from. This knowledge model control over the whole arrangement design process. </p><p>

65、;　　Figure 3 shows the layout rating system structure </p><p>　　The first kind of mode selection function is roughly calculated the numerical and working area general outline. This model provides the original

66、 parameters. All the information is roughly value resulting from them, no matter the figures are outlined draw or selected. </p><p>　　The second mode used to determine the layout type, Angle range, layout si

67、ze and strip the width. </p><p>　　The third kind of mode applied optimization algorithm. Design results include material utilization, material width and every step clearance are kept in this mode, the differ

68、ent layout drawing also generate. </p><p>　　In the fourth mode can modify layout results. Eventually parameters include clearance, material each step of grid, and the width, the ability to switch. When the p

69、arameters change, layout plans can be updated. </p><p>　　This knowledge is the main purpose of the algorithm to optimize the layout planning. This algorithm six steps. </p><p>　　1. The most suit

70、able around in graphics rectangular first generation. The original copy and the distance between the approach is included in the net. Figure 4 shows the algorithm. </p><p>　　2. The value of the two ring is a

71、mong a computation. The two ring is decomposed into line and arc units. The distance between each element needs to compensation. And then you can find the shortest distance. </p><p>　　3. The minimum value an

72、d calculated the value of the required the difference between is error. When the error less than value, arrangement planning can be completed. In addition, graphic layout to follow the direction of the view movement. <

73、;/p><p>　　Graph 4: arrangement algorithm. A primitive Angle graphics. B rotation Angle of graphics after </p><p>　　4. Material utilization in layout's point of view is calculated. </p>&

74、lt;p>　　5. Arrangement graphics rotating certain Angle. Rotating center near the center is the rectangular roughly value. Material utilization in the current Angle was calculated. </p><p>　　6. Arrangement

75、graphics rotated to another Angle. Repeat the steps of the third part, until Angle reached 180 degrees. </p><p>　　Figure 5 shows is the arrangement design results. </p><p>　　Graph 5: row kind of

76、 intelligent design results. </p><p>　　The development of 3.3 ribbon layout </p><p>　　The layout of the shingles rule was integration process in knowledge base level into tool design. This intel

77、ligent model function is: select parts location, design azimuth and arrange ribbon work step distance. In order to solve the operational procedures, and the rules should be reasonable and effective formulated. </p>

78、<p>　　Automatic design module is intelligent model in the most important modules. Artificial intelligence technology has been applied in this module. This model, including the pretreatment module orientating produ

79、cts module and extracted from the product modularization accurate information. In order to modify module generates a model, initial design engineering is modified [11]. The modified module instead of processing module. F

80、igure 6 shows the layout of the model and the algorithm for shingles. </p><p>　　3.3.1 automatic ribbon layout design preprocessing </p><p>　　1) determine the position and permutations. Parts The

81、 user can use interface to determine some of the preconditioning module parameter. The process can determine the position and other elements, such as: to do together shape and size accuracy, parts and user requirements.

82、</p><p>　　Parts in the shape of the intelligent model definition, the results are stored in a knowledge base. </p><p>　　2) get parts precise information. The precise information should get in ri

83、bbon layout knowledge base. Useful information including punching accurate information and relative location information. This type of knowledge model of information will decide parts stamping sequence. The design proces

84、s is the main requirements for the position precision to develop a knowledge model [12]. First, the shape of the parts were divided into closed contour. Outline the number of n </p><p>　　K = {k1, k2,..., ki,

85、..., kn} (1) </p><p>　　Here the first I ki says parts an outline. The relative relations between all contour contained in the relationship of P. If in contour kj ki and precision, there exists between requir

86、ements ki, kj) ∈ (p. </p><p>　　P = {...., (ki, kj),...} ki, kj ∈ K, 1 acuities were I, j acuities n (I _ = j). (2) </p><p>　　The position precision from P get relevant matrix is: </p><

87、;p><b>　　(3) </b></p><p>　　Each type of accurate information through the correlation matrices is preserved in knowledge model. </p><p>　　3.3.2 ribbon layout automatic design </p

88、><p>　　Ribbon automatic placement design module in the knowledge model is the most important one. In the knowledge model contains a lot of important rules, for example in a single stroke in stamping all the inn

89、er contour is better. In the next stage of this section was cut off. Sometimes, if the distance between the stamping point within is very small, some will be moved to outline a stage for machining. If stamping point too

90、close from fractionation point fractionation point, if will need to be change</p><p>　　Parts were divided into many point blank form. These point name is k1, k2,..., dij kn. Here </p><p>　　Kj ki

91、 and is the smallest distance between. Matrix critical value is S. If dij < S, ki and kj </p><p>　　Can't draw in the same steps. This kind of situation is intelligent model of the conflict between two

92、 points. Development of the intelligent model interference point ChongTuDian aim was to determine the existence. This matrix is a system matrix. In order to make the design process is more convenient, can put the matrix

93、of half part element zeros. </p><p>　　Here, mu ij correlation coefficient, which it says is for every to point between different relations. If there is a contradiction between two points, one of them is to b

94、e moved to the next step. In every step of ChongTuDian disappear until repeat this step. Finally matrix M become empty matrix. </p><p>　　The shingles layout of 3.3.3 trace process </p><p>　　The

95、layout of the shingles son two of handling knowledge model: change results and create part graphic layout. Automatic design model from ribbon layout draw result is staple. They may can't satisfy the users of all the

96、requirements. Depended on the knowledge model data structures, by moving point and change steps, increase steps and deletes empty empty the purpose can be realized step. We can pass the data processing steps to revise th

97、e ribbon of the layout design results. Step change can work </p><p>　　Figure 7: ribbon automatic placement design result </p><p>　　3.4 sure stamping center and force calculation of intelligent m

98、odel. </p><p>　　Press center design model establishment is the purpose of the group operates [11] together. Mould tool center and stamping center's consistent is very important, only as stamping tools ca

99、n together the normal work. Press center from the knowledge model calculation of every contour position in conclusion. Design is to get the first part of the tools work area. CAD platform of parts provided parts of graph

100、ic outline the rectangle. Rely on stamping center and the relationship between the rectangula</p><p>　　Stored in the repository includes every type of mould tool parts blanking, waste removal, etc. Different

101、 situations force calculation method is different. By force equations is obtained knowledge rules, the reasoning of the library. First, processing power and cutting force is based on parts of the contour length and knowl