冰露礦泉水瓶瓶蓋的注塑模具設(shè)計(jì)-螺紋伺服脫模含14張CAD圖,礦泉水瓶,瓶蓋,注塑,模具設(shè)計(jì),螺紋,羅紋,伺服,脫模,14,cad 1．外文資料翻譯譯文（約3000漢字）： 注塑件模擬焊縫成形的實(shí)驗(yàn)驗(yàn)證 J.G. Kovács*, B. Sikló 布達(dá)佩斯技術(shù)經(jīng)濟(jì)大學(xué)高分子工程系 摘要：近幾年來，由于對(duì)注塑件性能要求的不斷提高，人們對(duì)注塑件的焊縫分析越來越感興趣。當(dāng)兩個(gè)熔化前沿相互接觸時(shí)形成焊縫。如果不修改零件的幾何結(jié)構(gòu)，就不可能完全消除焊縫，但可以將其對(duì)零件性能和外觀的負(fù)面影響降到最低。這可以通過試錯(cuò)實(shí)驗(yàn)或模型預(yù)測來實(shí)現(xiàn)。后者的成本和時(shí)間效率使其成為焊縫分析的首選方法。注射成形計(jì)算機(jī)模擬軟件包能夠準(zhǔn)確預(yù)測焊縫位置，但現(xiàn)有的軟件包都不能定量預(yù)測焊縫接觸角和力學(xué)性能。本文對(duì)焊縫成形過程進(jìn)行了分析，提出了改進(jìn)有限元網(wǎng)格的方法，以獲得較好的效果。 關(guān)鍵詞：焊縫；注塑成型；模擬；有限元網(wǎng)格 1、介紹 注塑成型是用于成型塑料零件的最有效的工藝之一[1–6]。該方法的有效性取決于產(chǎn)品的質(zhì)量，這可能會(huì)受到工藝設(shè)置不足或模具結(jié)構(gòu)造成各種缺陷的阻礙。許多缺陷如焊縫、翹曲、噴射或凹陷等都會(huì)降低注塑件的質(zhì)量，降低生產(chǎn)效率。在注塑件的設(shè)計(jì)中，焊縫的產(chǎn)生是一個(gè)重要的美學(xué)和機(jī)械問題。當(dāng)兩個(gè)熔化前沿相互接觸時(shí)形成焊縫。在具有多個(gè)澆口的零件中，在模具填充過程中，可變壁厚、孔或型芯形成單獨(dú)的熔體前沿，而分離的熔體前沿形成焊縫，從而在零件中造成許多故障[7,8]。它不僅惡化了局部的機(jī)械性能，而且會(huì)產(chǎn)生光學(xué)缺陷，特別是在使用高光澤材料時(shí)。Chen等人在ABS拉伸鋼筋上研究了感應(yīng)加熱在表面溫度控制中的應(yīng)用，消除了焊縫表面的痕跡 [9]。 的確許多參數(shù)對(duì)焊縫的性能有影響，這些因素已經(jīng)從多個(gè)方面進(jìn)行了研究。在力學(xué)性能方面，對(duì)焊縫強(qiáng)度和模量進(jìn)行了分析，結(jié)果表明焊縫對(duì)拉伸模量沒有顯著影響[10,11]。一些研究人員[12-15]使用焊縫系數(shù)（WL factor），定義為：有焊縫的試樣強(qiáng)度/沒有焊縫的試樣強(qiáng)度，來評(píng)估他們的實(shí)驗(yàn)。采用高熔體溫度、高保壓壓力和低結(jié)晶器溫度對(duì)未填充材料的影響系數(shù)最高。利用激光引伸計(jì)和聲發(fā)射對(duì)焊縫進(jìn)行了研究，得出的結(jié)論是焊縫不是材料中的簡單不連續(xù)，而是應(yīng)力應(yīng)變分布的局部擴(kuò)展擾動(dòng)[16]。 近年來，由于對(duì)注塑件性能要求的不斷提高，人們對(duì)注塑件焊縫分析的興趣大大增加。如果不修改零件的幾何結(jié)構(gòu)，就不可能完全消除焊縫，但可以將其對(duì)零件性能和外觀的負(fù)面影響降到最低。這可以通過試錯(cuò)實(shí)驗(yàn)或模型預(yù)測來實(shí)現(xiàn)。后者的成本和時(shí)間效率使其成為焊縫分析的首選方法。注射成型計(jì)算機(jī)模擬軟件包能夠準(zhǔn)確預(yù)測焊縫位置，但現(xiàn)有的軟件包都不能定量預(yù)測焊縫性能。這主要是因?yàn)槠駷橹?，焊縫特性的數(shù)學(xué)模型不可用[17]。 在他們的文章中，周和李[17]提出了一個(gè)基于人工神經(jīng)網(wǎng)絡(luò)方法（ANN）的焊縫強(qiáng)度評(píng)估模型。對(duì)于網(wǎng)絡(luò)的輸入，選擇了影響焊縫性能的因素，即材料的取向系數(shù)、相遇角和熔體流動(dòng)歷史系數(shù)。與試驗(yàn)結(jié)果的比較表明，該模型能夠定量地預(yù)測焊縫性能，為工程設(shè)計(jì)提供了依據(jù)。Zhou等人 [18] 研究了熔體溫度和保壓對(duì)焊縫試件力學(xué)性能的影響，發(fā)現(xiàn)隨著保壓和熔體溫度的升高，焊縫試件的屈服強(qiáng)度和疲勞強(qiáng)度增加。他們解釋了觀察到的不同性質(zhì)的皮膚核心形態(tài)，這是受熔體溫度和保持壓力的影響。 Au [19]使用幾何方法來生成塑料部件的填充圖案，并確定可能的焊接線的大致位置。Fathi和Behravesh[20]用可視化技術(shù)研究了焊縫成形過程中的流動(dòng)動(dòng)力學(xué)行為，而Zhou和Li[17]開發(fā)了一個(gè)人工網(wǎng)絡(luò)來預(yù)測焊縫性能。為了確定網(wǎng)絡(luò)的輸入?yún)?shù)，對(duì)影響因素進(jìn)行了詳細(xì)的分析。對(duì)不可避免的焊縫非關(guān)鍵區(qū)域的形成和定位進(jìn)行了仿真分析。多澆口零件焊縫定位的流量控制采用流道尺寸調(diào)整方法[21]。Mezghani[22]將模擬的焊縫位置結(jié)果與注塑件的實(shí)際位置進(jìn)行了比較。Zhou和Li[23]提出了一種基于初始相遇節(jié)點(diǎn)特征的焊縫檢測算法。Chen[24]在零件仿真模型中應(yīng)用模糊理論，通過改變壁厚和澆口位置來控制焊縫位置。Chun[25]通過模擬研究了壁厚和澆口位置對(duì)焊縫形成和位置的影響。 2、實(shí)驗(yàn) 實(shí)驗(yàn)在Arburg Allrounder 320C 600-250注塑機(jī)上使用雙腔注射模進(jìn)行。這種特殊的模具有可更換的插入件，可以用不同的澆口類型（標(biāo)準(zhǔn)、薄膜、特殊薄膜等）具有不同的模具表面光潔度（拋光、細(xì)腐蝕、粗腐蝕），并注入不同厚度的試樣（0.5–4 mm）。樣品的厚度是由一個(gè)移動(dòng)的部分來設(shè)置的，以定位空腔的深度。注射模的頂出系統(tǒng)不同于傳統(tǒng)的頂出系統(tǒng)，它不包括頂出銷，而是在整個(gè)零件表面積上工作，從而消除了試樣的變形。澆口類型可以隨插入型腔之間的嵌件的變化而變化，而無需從注塑機(jī)上拆下模具。實(shí)驗(yàn)中，采用了精細(xì)的腐蝕表面光潔度，并在模具中設(shè)置了雙標(biāo)準(zhǔn)澆口鑲塊。 每個(gè)零件的標(biāo)稱尺寸為80 mm×80mm×2mm，從兩點(diǎn)注射成型。兩個(gè)標(biāo)準(zhǔn)澆口位于腔的一側(cè)，距零件邊緣10 mm，相距60 mm。 采用聚酰胺6（Durethan B30S，Lanxess）進(jìn)行研究。在注射成型之前，材料在80℃注射工藝條件保持恒定，模具溫度為90℃， 當(dāng)熔體溫度設(shè)定在280℃ 使用不同的切換點(diǎn)設(shè)置使用短射技術(shù)制作試樣。在樣品上測量熔體前沿的相遇角，作為流動(dòng)距離的函數(shù)。 測量結(jié)果繪制在圖5上。從中可以清楚地看出，會(huì)合角隨流長的增加而增大。在7毫米的流動(dòng)距離，它達(dá)到了一個(gè)可測量的焊縫角約28°，當(dāng)距離為22毫米時(shí)，角度為100°。在較長的流動(dòng)中，由于熔體前沿的輪廓，無法測量會(huì)合角。 通過在澆口位置中心使用同心圓對(duì)熔體前沿進(jìn)行可視化，還構(gòu)建了會(huì)合角。結(jié)果表明，拉伸會(huì)合角的增加沒有測量值高。在理論流動(dòng)距離為10毫米時(shí)，它很好地代表了測量值，但在較長的距離時(shí)，它低估了實(shí)驗(yàn)尺度。 3、分析 有限元模擬注射成型是目前設(shè)計(jì)注射模最先進(jìn)的技術(shù)。市場上有不同級(jí)別的節(jié)目。這些基礎(chǔ)知識(shí)對(duì)產(chǎn)品設(shè)計(jì)有一定的幫助，可以在不了解塑料制造的情況下使用。更復(fù)雜的程序能夠模擬整個(gè)注射成型過程，這樣人們就可以看到模具是否能夠完美工作。這種軟件覆蓋了大量的材料和機(jī)械數(shù)據(jù)庫，設(shè)計(jì)者必須具備塑料制造的專業(yè)知識(shí)。 對(duì)于注塑模擬，在大多數(shù)情況下，采用二維三角形單元或三維四面體單元來描述型腔，其中兩節(jié)點(diǎn)管單元用于流道、連接件和通道。用控制體積法計(jì)算了熔體前沿的變化。在每一步中都可以得到壓力場、溫度場和速度場。這些結(jié)果構(gòu)成了應(yīng)力和變形分析以及焊縫結(jié)果的基礎(chǔ)。 Moldflow Plastics Insight 6.2用于模擬分析實(shí)驗(yàn)中使用的零件模型。在分析過程中，使用了三種不同的中間平面網(wǎng)格類型：原始網(wǎng)格、理想網(wǎng)格和平滑網(wǎng)格。每種網(wǎng)格類型在4個(gè)網(wǎng)格邊長度中完成：1、2、2.5和5 mm。 原始網(wǎng)格是指模型由等邊三角形組成，沿估計(jì)焊縫的節(jié)點(diǎn)不產(chǎn)生直線。這種網(wǎng)格類型的優(yōu)點(diǎn)是具有良好的長寬比。網(wǎng)格單元的長寬比非常重要，因?yàn)樗绊懡Y(jié)果的精度。比率定義了三角形的最長邊與三角形面積之間的相關(guān)性，而中面網(wǎng)格的推薦最大縱橫比為約6?？梢钥闯?，在每一條邊的長度上，這種網(wǎng)格三角形的平均長寬比都大于1.5。 理想網(wǎng)格由具有共線節(jié)點(diǎn)的等腰三角形構(gòu)成。生成這種網(wǎng)格類型的優(yōu)點(diǎn)是，它可以很好地自動(dòng)化，但是，由于較差的縱橫比，即2，它不如原始網(wǎng)格那樣精確。在平滑網(wǎng)格的情況下，原始網(wǎng)格的節(jié)點(diǎn)收斂形成一條線，在預(yù)測的焊縫區(qū)域中創(chuàng)建網(wǎng)格三角形邊的更均勻路徑。它是從原始網(wǎng)格類型生成的，并在焊縫區(qū)域進(jìn)行了修改。將焊縫上的節(jié)點(diǎn)靠近理想焊縫位置。 模擬分析的工藝設(shè)置與實(shí)驗(yàn)注射成型相同，模具恒溫，熔體溫度為90℃和280℃。 焊縫分析結(jié)果與實(shí)驗(yàn)結(jié)果進(jìn)行了比較。在大多數(shù)情況下，理想的網(wǎng)格類型最適合測量結(jié)果。在邊長為1mm的情況下，采用理想和平滑網(wǎng)格類型的分析接近于流量長度為7-10mm之間的測量結(jié)果。原始網(wǎng)格類型計(jì)算的值沿整個(gè)檢測流長在測量結(jié)果周圍波動(dòng)，不接近測量值，而其他網(wǎng)格類型與流動(dòng)開始時(shí)的測量值不同。同時(shí)觀察到，在距離為10 mm后，所有網(wǎng)格都預(yù)測出焊縫角急劇增加。 當(dāng)網(wǎng)格長度為2 mm且流動(dòng)距離較短時(shí)，振動(dòng)再次明顯。與測量結(jié)果相比，原始網(wǎng)格給出的結(jié)果最不準(zhǔn)確。角度值變化較大：計(jì)算出焊縫角為0°距離10.6毫米但147°12毫米。除原始網(wǎng)格外，在較長的流動(dòng)路徑下與測量結(jié)果的差異小于在邊緣長度為2 mm時(shí)的差異。 邊緣長度增加到2.5mm，測量結(jié)果和模擬結(jié)果之間的相似性降低。在流動(dòng)距離為15～20 mm的區(qū)域，用理想網(wǎng)格模擬計(jì)算的焊縫線角與實(shí)測值接近，但其它網(wǎng)格變化不符合實(shí)測值的變化趨勢。 使用5 mm的邊緣長度，曲線之間的一致性很弱。盡管分析結(jié)果顯示出一些相似性，但出乎意料的是，結(jié)果僅在少數(shù)流動(dòng)長度下接近測量值。 比較每個(gè)邊緣長度處的不同網(wǎng)格類型，可以注意到在每種情況下，理想網(wǎng)格與測量數(shù)據(jù)的相關(guān)性最好，在0.95和0.98之間變化。結(jié)果還表明，理想網(wǎng)格的最佳相關(guān)度在高邊長處，即5mm處，但隨著邊長的減小，相關(guān)度降低的幅度相對(duì)較小。對(duì)于原始網(wǎng)格類型，相關(guān)性最低，但隨著邊緣長度的增加，相關(guān)性顯著提高，但這種網(wǎng)格類型并沒有達(dá)到理想的相關(guān)性值。使用平滑網(wǎng)格，相關(guān)度隨著邊緣長度的增加而提高，但也沒有達(dá)到理想網(wǎng)格類型的值。 參考文獻(xiàn) [1] T. Tábi, J.G. Kovács, Examination of injection molded thermoplastic maize starch. Express Polym. Lett. 12 (2007) 423. [2] L. Mészáros, T. Tábi, J.G. Kovács, T. Bárány, The effect of EVA content on the processing parameters and the mechanical properties of LDPE/ground tire rubber blends. Polym. Eng. Sci. 48 (2008) 868. [3] E. Lafranche, P. Krawczak, J.P. Ciolczyk, J. Maugey, Injection moulding of long glass fibre reinforced polyamide 6-6: guidelines to improve flexural properties. Express Polym. Lett. 7 (2007) 456. [4] G. Dogossy, T. Czigány, Modeling and investigation of the reinforceing effect of maize hull in PE matrix composites. Polym. AdvanTechnol. 17 (2006) 825. [5] S. Hashemi, Effect of temperature on tensile properties of injection moulded short glass fibre and glass bead filled ABS hybrids. Express Polym. Lett. 7 (2008) 474. [6] K. Banik, Effect of mold temperature on short and long-term mechanical properties of PBT. Express Polym. Lett. 2 (2008) 111. [7] J. Shoemaker, Moldflow Design Guide. Carl Hanser Verlag, Munich,2006. [8] R.A. Malloy, Plastic Part Design for Injection Molding. Hanser Publishers, 1994. [9] S.-C. Chen, W.-R. Jong, J.-A. Chang, Dynamic mold surface temperature control using induction heating and its effects on the surface appearance of weld line. J. Appl. Polym. Sci. 101 (2006) 1174. [10] S. Hashemi, Y. Lepessova, Temperature and weldline effects on tensile properties of injection moulded short glass fibre PC/ABS polymer composite. J. Mater. Sci. 42 (2007) 2652. [11] S. Hashemi, Thermal effects on weld and unweld tensile properties of injection moulded short glass fibre reinforced ABS composites. Express Polym. Lett. 1 (2007) 688. [12] R. Seldén, Effect of processing on weld line strength in five thermoplastics. Polym. Eng. Sci. 37 (1997) 205. [13] S. Hashemi, Influence of temperature on weldline strength of injection moulded short glass fibre styrene maleic anhydride polymer composites. Plast. Rubber Compos 31 (2002) 318. [14] C. Lu, S. Guo, L. Wen, J. Wang, Weld line morphology and strength of polystyrene/polyamide-6/poly(styrene-co-maleic anhydride) blends.Eur. Polym. J. 40 (2004) 2565. [15] N. Merah, M. Irfan-ul-Haq, Z. Khan, Temperature and weld-line effects on mechanical properties of CPVC. J. Mater. Process. Tech. 142 (2003) 247. [16] C. Bier?gel, W. Grellmann, T. Fahnert, R. Lach, Material parameters for the evaluation of PA welds using laser extensometry. Polym.Test. 25 (2006) 1024. [17] H. Zhou, D. Li, Computer evaluation of weld lines in injectionmolded parts. J. Reinf. Plast. Comp. 24 (2005) 315. [18] Y. Zhou, P.K. Mallick, Effects of melt temperature and hold pressure on the tensile and fatigue properties of an injection molded talcfilled polypropylene. Polym. Eng. Sci. 45 (2005) 755. [19] C.L. Au, A geometric approach for injection mould filling simulationInt. J. Mach. Tools Manuf 45 (2005) 115. [20] S. Fathi, A.H. Behravesh, Visualization analysis of flow behavior during weld-line formation in injection molding process. Polym. Plast. Technol. 47 (2008) 666. [21] M. Zhai, Y. Lam, C. Au, Runner sizing and weld line positioning for plastics injection moulding with multiple gates. Eng. Comput. 21(2006) 218. [22] K. Mezghani In: The 6th Saudi Engineering Conference, Dharan,2002, pp. 335–347. [23] H. Zhou, D. Li, Modelling and prediction of weld line location and properties based on injection moulding simulation. Int. J. Mater. Prod. Technol. 21 (2004) 526. [24] M.-Y. Chen, H.-W. Tzeng, Y.-C. Cheng, S.-C. Chen, The application of fuzzy theory for the control of weld line positions in injection molded part. ISA T 47 (2008) 119. [25] D.H. Chun, Cavity filling analyses of injection molding simulation:bubble and weld line formation. J. Mater. Process. Tech. 89-90 (1999) 177.  2．外文資料原文（與課題相關(guān)，至少1萬印刷符號(hào)以上）： Experimental validation of simulated weld line formation in injection moulded parts J.G. Kovács*, B. Sikló Abstract：The interest in weld line analysis of injection-moulded parts has increased in the past few years, mainly because of the ever-increasing requirements for the performance of injec- tion-moulded items. Weld lines are formed when two melt fronts come in contact with each other. Whereas the total elimination of weld lines is not always possible without modifying the part geometry, their negative in?uence on part performance and appear- ance can be minimized. This can be done by trial and error experiments or by model prediction. The cost and time ef?ciency of the latter makes it a preferred route for weld lines analysis. Computer simulation packages of injection moulding are capable of accu- rately predicting the weld line location, but none of the current ones can predict the weld line contact angle or mechanical properties quantitatively. This paper focuses on the analysis of weld line formation and suggests ways to modify the ?nite element mesh to get better results. Keywords：Weld line ；Knit line；Injection moulding ；Simulation；Finite element mesh 1.Introduction Injection moulding is one of the most productive processes used to form plastic parts [1–6]. The effectiveness of the method depends on the quality of the product, which can be hindered by inadequate process settings or mould construction causing various de?ciencies. Many kind of defect such as weld lines, warpage, jetting or sink marks can reduce the quality of the injection moulded parts, worsening productivity. The occurrence of a weld line means a signi?cant problem both aesthetically and mechanically in the design of injection moulded parts. Weld lines are formed when two melt fronts come in contact with each other. In a part with multiple gates, variable wall thicknesses, holes or cores form separate melt fronts during mould ?lling and the separated melt fronts create weld lines, causing numerous troubles in the part [7,8]. It not only worsens the local mechanical properties, but creates optical imperfections, especially when using high gloss materials. The surface marks of weld lines can be eliminated by the application of induction heating in surface temperature control, which was investigated on ABS tensile bars by Chen et al. [9]. Many parameters have an effect on the properties of a weld line and these factors have been investigated from many aspects. As regards mechanical properties, analysis of weld line strength and modulus was performed and showed that the weld line did not have a signi?cant effect on tensile modulus [10,11]. Several researchers [12–15] used the weld line factor (WL-factor), de?ned as: strength of specimens with weld line/strength of specimens without weld line, to evaluate their experiments. Highest WL- factors were obtained for un?lled materials and using high melt temperature, high holding pressure and low mould temperature. Weld lines were studied using laser exten- someter and acoustic emission, and the conclusion was that a weld line is not a simple discontinuity in the material, but a locally extended disturbance of the stress and strain distribution [16]. The interest in weld line analysis of injection-moulded parts has increased greatly in the past few years, mainly because of the ever-increasing requirements for the performance of injection-moulded items. Whereas the total elimination of weld lines is not always possible without modifying the part geometry, their negative in?uence on part performance and appearance can be minimized. This can be done by trial and error experiment or by model prediction. The cost and time ef?ciency of the latter makes it a preferred route for weld line analysis. Computer simulation packages of injection moulding are capable of accurately predicting the weld line location, but none of the current ones can predict the weld line prop- erties quantitatively. This is mainly because a mathematical model for weld line properties is, to date, unavailable [17]. In their article, Zhou and Li [17] presented an evaluation model for weld line strength based on the arti?cial neural network method (ANN). For the input of the network, the factors affecting weld line properties were chosen; those are the orientation coef?cient of the material, the meeting angle and the melt mobility history coef?cient. Comparison with experimental results shows that the presented model is capable of predicting weld line properties quantitatively for engineering design. Zhou et al. [18] examined the effects of melt temperature and hold pressure on the mechanical properties of specimens with weld lines and found that the yield and fatigue strengths of the specimens increased with increasing hold pressure as well as increasing melt temperature. They explained the observed differences in properties in terms of a skin-core morphology, which was in?uenced by both the melt temperature and the holding pressure. Au [19] used a geometrical approach to generate the ?lling patterns of plastic parts and determine the approx- imate location of possible weld lines. Fathi and Behravesh[20] studied the kinematical behaviour of the ?ow during weld formation with a visualization technique, while Zhou and Li [17] developed an arti?cial network to predict weld line properties. The affecting factors were analyzed in detail in order to identify the input parameters for the network. The formation and positioning in noncritical areas of unavoidable weld lines are also investigated with simula- tion analyses. The controlling of the ?ow for weld line positioning for multi-gated parts was carried out with a runner resizing method [21]. Mezghani [22] compared the simulated weld line location results with the real position on injection moulded parts. Zhou and Li [23] presented a weld detector algorithm, which is based on the characteristics of the initial meeting node. Chen [24] applied fuzzy theory for controlling the weld line position by varying the wall thickness and the gate location in part simulation models. Chun [25] showed by simulation the effect of wall thickness and gate location on the formation and position of weld lines. 2.Experimental The experiments were performed on an Arburg Allrounder 320C 600-250 injection moulding machine using a two cavity-injection mould (Fig. 1.). This special mould has changeable inserts to be able to inject with different gate types (standard, ?lm, special-?lm, multi gates, etc.), with different mould surface ?nishes (polished, ?ne eroded, rough eroded) and to inject different thickness specimens (0.5–4 mm). The thickness of the samples is set by a moving part to position the depth of the cavities. The ejection system of the injection mould differs from the conventional one; it does not include ejector pins but operates on the whole part surface area, so eliminating deformation of the sample. The gate type can be varied with the change of an insert interposed between the cavi- ties without dismounting the mould from the injection moulding machine. For the experiments, a ?ne eroded surface ?nish was used and an insert with double standard gates was set in the mould. Each part, having nominal dimensions of 80 mm 80 mm 2 mm, was injection moulded from two points (Fig. 2.). The two standard gates are located on one side of the cavity 10 mm from the part edge and 60 mm apart. Polyamide 6 (Durethan B30S, Lanxess) was used for the investigations. Before injection moulding, the material was dried at 80 ○C for 4 h. The injection processing conditions were kept constant; the mould temperature was 90 ℃while the melt temperature was set at 280 ○C. The speci- mens were produced with short shot technology using different switch-over point settings (Fig. 3.). The meeting angle of the melt front was measured on the samples as a function of the ?ow distance (Fig. 4.). The results of the measurement are plotted on Fig. 5. It can be clearly seen that the meeting angle increased with the ?ow length. At a ?ow distance of 7 mm, it reached a measurable weld line angle of about 28○, while at a distance of 22 mm the angle achieved was 100○. At longer ?ows the measurement of the meeting angle was not possible because of the pro?le of the melt front. The meeting angles were also constructed from the visualization of the melt fronts using concentric circles centred at the gate locations. The results showed that the increase of the drawn meeting angle was not as high as the measured values (Fig. 5.). At a theoretical ?ow distance of 10 mm it represented the measured values well but at a longer distance it underestimated the experimental scale. 3、Analysis Injection moulding simulation with the ?nite element method is the most advanced technique for designing injection moulds. There are different levels of program available on the market. The basic ones are helpful in product design, which can be used without having deep knowledge of plastic manufacturing. The more complex programs are able to simulate the whole injection moulding process so one can see whether the mould will be able to work perfectly or not. Such software cover enor- mous databases of materials and machines and the designers must have professional knowledge of plastic manufacturing. For an injection moulding simulation, in most cases two-dimensional triangular elements or three-dimensional tetrahedron elements are used to describe the cavity, with two-node tube elements for the runners, connectors and channels. The melt front advancements are calculated by the control volume method. The pressure, temperature and velocity ?eld can be obtained in each time step. These results constitute the basis of the stress and deformation analysis as well as results for weld lines. Mold?ow Plastics Insight 6.2 was used for the simulation analyses with a model of the part used in the experiments (Fig. 6.). During the analyses, three different mid plane mesh types were used and compared: original mesh, ideal mesh and smoothed mesh. Each mesh type was completed in 4 mesh edge lengths: 1, 2, 2.5 and 5 mm. Original mesh means that the model consists of equilateral triangles and the nodes along the estimated weld line did not produce a straight line. The advantage of this mesh type is the good aspect ratio. The aspect ratio of the mesh elements is important because it affects the accuracy of the results. The ratio de?nes the correlation between the longest side of the triangle and the triangle area, and the recommended maximum aspect ratio for a mid plane mesh is about 6. It can be seen that at every edge length this type of mesh triangle was greater than an average aspect ratio of 1.5. Ideal mesh is made up of isosceles triangles with collinear nodes. The advantage of the generation of this mesh type is that it can be well automated, however, because of the worse aspect ratio, namely 2, it was not as accurate as the original mesh (Fig. 7). In the case of the smoothed mesh, the nodes of the original mesh are converged to form a line creating a more uniform path of the mesh triangle sides in the area of the predicted weld line. It was generated from the original mesh type with modi?cation at the weld line region. The nodes positioned on the weld line were made nearer to the ideal weld line position. The process settings for the simulation analyses were identical to the experimental injection moulding, constant mould temperature and melt temperature namely 90 ℃ and 280 ℃。 The weld line analysis results were compared to the experimental. In most cases, the ideal mesh type best ?tted the results of the measurements. At an edge length of 1 mm, the analyses with ideal and smoothed mesh type came close to the measurement results between flow lengths of 7 and 10 mm (Fig. 8.). The values calculated with original mesh type flucated around the measured results along the whole examined flow length and did not approach them, while the other mesh types differed from the measured at the beginning of the ?ow. It was also observed that after a distance of 10 mm a steep weld line angle increase was predicted by all of the meshes. At a mesh length of 2 mm and a short ?ow distance an oscillation was again noticeable (Fi