【正文】 echnical requirement of the user, the user training time from the past few weeks shorter for a few hours. Figure 1 is based on the central layer model and surface model based on 3D solid / flow analysis simulation parison chart.Finite element, finite difference, the control volume methodsInjection molding products are thin products, products in the thickness direction of size is much smaller than the other two dimensions, temperature and other physical quantities in the thickness direction of the change is very large, if the use of a simple finite element and finite difference method will cause analysis time is too long, can not meet the actual needs of mold design and manufacturing. We in the flow plane by using finite element method, the thickness direction by using finite difference method, were established and plane flow and thickness directions corresponding to the size of the grid and coupling, while the accuracy is guaranteed under the premise of the calculation speed to meet the need of engineering application, and using the control volume method is solved. The moving boundary problem in. For internal and external correspondence surface differences between products, can be divided into two parts the volume, and respectively formed the control equation, the junction of interpolation to ensure that the two part harmony contrast.Numerical analysis and artificial intelligence technologyOptimization of injection molding process parameters has been overwhelming majority of mold design staff concerns, the traditional CAE software while in puter simulation of a designated under the conditions of the injection molding conditions, but is unable to automatically optimize the technical parameters. Using CAE software personnel must be set to different process conditions were multiple CAE analysis, bined with practical experience in the program were pared between, can get satisfactory process scheme. At the same time, the parts after the CAE analysis, the system will generate a large amount of information about the project ( product, process, analyzes the results ), which often results in a variety of data form, requiring the user to have the analysis and understanding of the results of CAE analysis ability, so the traditional CAE software is a kind of passive putational tools, can provide users with intuitionistic, effective engineering conclusion, to software users demand is too high, the influence of CAE system in the larger scope of application and popularization. In view of the above, HSCAE3D software in the original CAE system based on accurate calculation function, the knowledge engineering technology is introduced the system development, the use of artificial intelligence is the ability of thinking and reasoning, instead of the user to plete a large number of information analysis and processing work, directly provide guiding significance for the process of conclusions and remendations, effectively solve the CAE of the plexity of the system and the requirements of the users of the contradiction between, shortening of the CAE system and the distance between the user, the simulation software by traditional passive putational tools to active optimization system. HSCAE3D system artificial intelligence technology will be applied to the initial design, the results of the analysis of CAE interpretation and evaluation, improvement and optimization analysis of3 aspects.附錄：科技英文中文翻譯注塑模具設(shè)計的技術(shù)1．用三維實體模型取代中心層模型　　傳統(tǒng)的注塑成形仿真軟件基于制品的中心層模型。用戶首先要將薄壁塑料制品抽象成近似的平面和曲面，這些面被稱為中心層。在這些中心層上生成二維平面三角網(wǎng)格，利用這些二維平面三角網(wǎng)格進行有限元計算，并將最終的分析結(jié)果在中面上顯示。而注塑產(chǎn)品模型多采用三維實體模型，由于兩者模型的不一致，二次建模不可避免。但由于注塑產(chǎn)品的形狀復(fù)雜多樣、千變?nèi)f化，從三維實體中抽象出中心層面是一件十分困難的工作，提取過程非常繁瑣費時，因此設(shè)計人員對仿真軟件有畏難情緒，這已成為注塑成形仿真軟件推廣應(yīng)用的瓶頸?！　SCAE　3D主要是接受三維實體/表面模型的STL文件格式。現(xiàn)在主流的CAD/CAM系統(tǒng)，如UG、Pro/ENGINEER、 CATIA和SolidWorks等，均可輸出質(zhì)量較高的STL格式文件。這就是說，用戶可借助任何商品化的CAD/CAE系統(tǒng)生成所需制品的三維幾何模型的STL格式文件，HSCAE　3D可以自動將該STL文件轉(zhuǎn)化為有限元網(wǎng)格模型，通過表面配對和引入新的邊界條件保證對應(yīng)表面的協(xié)調(diào)流動，實現(xiàn)基于三維實體模型的分析，并顯示三維分析結(jié)果，免去了中心層模擬技術(shù)中先抽象出中心層，再生成網(wǎng)格這一復(fù)雜步驟，突破了仿真系統(tǒng)推廣應(yīng)用的瓶頸，大大減輕了用戶建模的負擔，降低了對用戶的技術(shù)要求，對用戶的培訓時間也由過去的數(shù)周縮短為幾小時。圖1為基于中心層模型和基于三維實體/表面模型流動分析模擬情況對比圖。　　2．有限元、有限差分、控制體積方法的綜合運用　　注塑制品都是薄壁制品，制品厚度方向的尺寸遠小于其他兩個方向的尺寸，溫度等物理量在厚度方向的變化又非常大，若采用單純的有限元或有限差分方法勢必造成分析時間過長，無法滿足模具設(shè)計與制造的實際需要。我們在流動平面采用有限元法，厚度方向采用有限差分法，分別建立與流動平面和厚度方向尺寸相適應(yīng)的網(wǎng)格并進行耦合求解，在保證計算精度的前提下使得計算速度滿足工程的需要，并采用控制體積法解決了成形中的移動邊界問題。對于內(nèi)外對應(yīng)表面存在差異的制品，可劃分為兩部分體積，并各自形成控制方程，通過在交接處進行插值對比保證這兩部分的協(xié)調(diào)。3．數(shù)值計算與人工智能技術(shù)的結(jié)合　　優(yōu)選注塑成形工藝參數(shù)一直是廣大模具設(shè)計人員關(guān)注的問題，傳統(tǒng)的CAE軟件雖然可以在計算機上仿真出指定工藝條件下的注塑成形情況，但無法自動對工藝參數(shù)進行優(yōu)化。CAE軟件使用人員必須設(shè)置不同的工藝條件進行多次CAE分析，并結(jié)合實際經(jīng)驗在各方案之間進行比較，才能得出較滿意的工藝方案。同時，在對零件進行CAE分析后，系統(tǒng)會產(chǎn)生有關(guān)該方案的大量信息(制品、工藝條件、分析結(jié)果等)，其中分析結(jié)果往往以各種數(shù)據(jù)場的形式出現(xiàn)，要求用戶必須具備分析和理解CAE分析結(jié)果的能力，所以傳統(tǒng)的CAE軟件是一種被動式的計算工具，無法提供給用戶直觀、有效的工程化結(jié)論，對軟件使用者的要求過高，影響了CAE系統(tǒng)在更大范圍內(nèi)的應(yīng)用和普及。針對以上不足，HSCAE　3D軟件在原有CAE系統(tǒng)準確的計算功能基礎(chǔ)上，把知識工程技術(shù)引入系統(tǒng)的開發(fā)中，利用人工智能所具有的思維和推理能力，代替用戶完成大量信息的分析和處理工作，直接提供具有指導(dǎo)意義的工藝結(jié)論和建議，有效解決了CAE系統(tǒng)的復(fù)雜性與用戶使用要求的簡單性之間的矛盾，縮短了CAE系統(tǒng)與用戶之間的距離，將仿真軟件由傳統(tǒng)的“被動式”計算工具提升為“主動式”優(yōu)化系統(tǒng)。HSCAE　3D系統(tǒng)主要將人工智能技術(shù)應(yīng)用于初始工藝方案設(shè)計、CAE分析結(jié)果的解釋和評價、分析方案的改進與優(yōu)化3個方面。35