带自由面压铸模具冷却系统的最优设计摘 要这个研究是关于有限元法和运用到带自由表面压铸模具的推断网络法。研究的目的是发现最佳的冷却系统参数和减
少压铸模具的变形。为了避免众多的影响因素，压铸模具的自由表面采用等价的多项式函数非线性的结构。根据非线性函数，包括空间开槽和孔道直
径的冷却系统参数被适当调整了。一个模仿压铸冷却参数的推断网络系统已经被构造出来。这个推断网络包括许多函数节。一旦这冷却系统的参数被
给定，这个网络系统便可以精确地预测压铸件的变形量。一个带有性能指标的模拟退火最佳化算法然后被应用到神经网络，目的是探索最优的冷却系
统参数和获得一个满意效果。关键词：压铸模具；自由形态的；神经网络；模拟退火1.引言典型的传统压力铸造法包括高气压的充填，冷却，凝固
和顶出阶段。冷却阶段具有非常的重要性，因为它能较大地影响生产能力继而影响压铸件的数量。众所周知压模铸件大约有八成的循环时间被花费于
对热熔进行充分地冷却，目的是使铸件可以被没有翘曲的顶出。一个成功的冷却系统的设计可以显著地减小各种因数影响，它可以减少冷却时间、减
少翘曲和相应地增加部件的质量。冷却过程的主要目的是维持充填和冷却的均匀温度。相应地,当考虑冷却系统和建立冷却过程条件时至少有二个重
要的准则供设计师参考：1、达到均匀温度；2、缩小循环时间。要实现这两个目标，设计师可能需要一个最佳的计算机辅助设计系统来完成一个快
速和均匀的冷却体系。在充填和冷却过程期间，最优系统设计需要进行热传递分析。这个热分析工具将预测压铸件的温度梯度和变形。一般而言，传
统的压模设计仍然依靠经验，由于缺乏铸造流动和热传递的有利分析，设计师不能评价和控制由于压铸材料、膨胀和收缩引起的变形。不同冷却系统
参数可以引起大的温度梯度和不同的变形。虽然有限元法软件能够分析一个在不同的冷却系统的压铸模具的注射金属压力和热应力、热膨胀和温度分
布情况的填充流动和冷却条件，分析模型的建立是很难的，特别是三维自由形态的几何学。除了了解多腔模、金属流动和固化过程非常必要，设计师
还应该完全地掌握基本的有限元软件。只要完全的了解压出板制造是能有效避免人员移动麻烦的过程，引用软件就可以达到和节省大量金钱和时间。
首先，对冷室压模铸件注射和压铸型腔填充排气孔和溢流口进行设计。当金属铸件使用一个活塞进入到一个腔内，他会考虑金属发生的变化和型腔内
的空气被熔化的金属置换。随后Garber将会显示太大的或太小的活塞速度可能影响铸件的质量Groenevelt和Kaiser研讨了注
射入型腔内熔融金属的速度的影响、流经距离和腔内产品铸件上表面的温度。根据压铸件不同的浇注初始温度实验可得太低的预先加热温度可能引起
铸塑料液堵塞流道，从而发生故障。但较高的温度可能增加冷却时间和减少生产能力。Truelove使用一个冷却系统控制压铸过程的整个温度
，目的是获得一最佳的传热特征，减少压铸件的热节问题的发生，从而改善铸件的质量。Jong和一些人发明了一个用于熔融金属在高气压压铸期
间流动和凝固数学方程，以便分析型腔内压铸元件的温度情况和冷却应力。kenichiro和一些人使用有限元法分析和设计压铸模；结果不仅
改善精度，但被考虑的因素还有许多，如压模铸件的压力、浇铸的液体的流动速度、粘性和材料随温度和相变化的机械特性。使用CAD\CAE故
障软件对压出板的系统设计过程的研究，目的是减少压模设计过程中的人为误差。它使用CAD软件创造一个形式自由的模型、使用有限元软件分析
压铸过程的情况。它模拟压铸件和在不同参数（冷却线长度R、开槽中心距L、孔道直径D）下的铸件变形的温度分布，如图1所示。它使用一个推
断系统建立了关系图1，冷却通道和自由形态压出板之间的关系，变形和冷却系统参数模型之间的关系。根据推断模拟方法，它能描绘输入和输出变
量之间的复杂的和不确定的关系。一旦推断系统构造出输入和输出压力铸造参数的关系，一个合理的带有性能指标的优选法能探索出最佳的铸造参数
。在这里，一个模拟退火的测深优化方法被采用。这个模拟退火算法是通过模拟退火过程来减少性能指标。它已经被成功地应用到压铸模具设计中等
。这个基础理论可以被广泛地应用。2.压铸流动理论在压铸过程中大约80%的时间花费在冷却过程中。压模铸件的变形是由于浇铸过程中不均匀
的温度分布引起的，它影响铸件的质量。冷却系统的设计者不得不考虑整体循环和计算压铸过程中不同阶段的变形。铸造过程分析包括三主要阶段：
第一，浇铸过程必须保证浇铸充满型腔。主要的压模铸件流动方程被分成五阶段。在填充阶段，模槽在高压下充满浇铸的塑性流体。3.建立冷却系
统和压铸模具变形之间的关系高气压注射铝合金压铸的铸造压力大约是30–150 Mpa；通常注射压力随时间而变。为了研究铸造过程压力的
影响，Dochler和Borton使用阴极射线示波器和照相机分析铸造过程中的气压变化，那就是说高压充填、冷却和顶出。压铸模具的设计
包括流道、均匀的型腔布置、分析压铸模具的寿命（残余应力）、冷却系统等等。这个研究结果的目的是找到适合铸造任何工件的压铸模具的最佳冷
却系统。假定铸造条件是：压铸压力120 Mpa、浇注速度2.8 m / s、压铸循环时间20s 每次、预先加热温度150oC、注射
温度700oC。对流入型腔的液体的基本假定：1、三维流动；2、牛顿流体；3、层流；4、不可压缩流体；5、流体在垂直的和水平方向无流
速差别。根据在零件面的闭合截面不同的冷却参数，有15个设计数据用于模拟压铸过程。根据三维流动模型的模型流动分析，基本布局是一个自由
形态的表面。压铸模具的表面温度被预先加热到一定温度(150度)，熔融金属的温度被控制在700度。注射口的温度特别需要维持在700度
。模具的温度是150度、冷却水温度是40度、其它的温度在填充时需要即时使用有限元分析控制，如图2和3所示。因为获得其临界温度条件需
进行分析，变形分析使用三维流动和非线性条件对实体模型进行分析。各节点的配置温度作为初始条件被输入。使注射口的边界约束。条件充当三维
热应变条件。机械性能随温度变化而变。它是从非线性稳定阶段分析中获得。上面讨论的型腔流动分析结果用应变分析阶段执行显示在表格1上。铸
造过程中的参数是很复杂的和难以控制的。在各参数和指标函数之间的关系很难明确决定。用于实际压铸条件的实验方法和统计法是不同的。在实际
运用中有很多地限制。研究使用一个神经网络去学习和培养一个系统，它用于压铸件的变形和铸造过程中的变形，使用这个神经网络完成各参数的进
一步地分析。 图2、温度梯度 图3 变形分布同样地，输入参数关系的建立（冷却系统参数： R,冷却线距离: D,孔道直径: L,开槽
中心距）和在铸造过程期间输出参数（变形）被显示在附录上。建设一个完整的推断系统，第一个必要条件是建立数据库。由输入和输出产生的信息
必须足够的多。因而推断网络训练的导流因素（冷却系统参数）应该完美和制造没有缺点的产品。表格1阐明了从三维模型流动分析获得的压铸件的
冷却系统参数和极限变形。根据压铸模型的发展，三层推断系统能自动地综合处理，它由冷却系统参数和铸件结果（变形）组成。不同的控制参数被
用于这个系统，它能够预先模拟压铸模型在不同的控制参数下面的变形。全部的多项式方程被登记在附录（PSE = 5.43 X 10_7）
。表格2比较这些出错预测的模型和模拟情况。这模拟情况是从因为建立这个模型进行压铸模拟试验而设置的20个装置中得到的这个数据集被用来
检定这模型建立的合理性。我们看得见来源于表2的故障大约2% ，则可以得到建立这模型的目的。附 录（英文版）The Optimal
Design of a Cooling System for a Die-Casting Die With a Free Form
SurfaceAbstractThis study is on the finite element and abductive
network method application to die-casting dies with free-form su
rfaces.The study aims to find the optimal cooling system paramete
rsand decrease in deformation of a die-casting die. In order to a
void the numerous influencing factors, the free-form surface of a
die-casting die is created as a non-linear Eq. of a polynomial f
unction. The parameters of the cooling system, including the chan
nel space and channel diameter, are adjusted according to the non
-linear Eq..An abductive network has been built for modelling the
diecasting cooling parameters. The abductive network is composed
of a number of functional nodes. Once the cooling system paramet
ers are given, this network can predict the deformation of the di
e-casting accurately. A simulated annealing optimization algorith
m with a performance index is then applied to the neural network
for searching for the optimal cooling system parameters and to ob
tain a satisfactory result.Keywords： Die-casting die；Free-form；Ne
ural network；Simulated annealing1. IntroductionThe typical, tradi
tional die-casting process includes high-pressure filling, coolin
g, solidification and ejection stages. The cooling stage is of gr
eat importance because it significantly affects both the producti
vity and the quality of the die-cast part. It is well known that
about 80% of the cycle time of die-casting is spent in cooling th
e hot melt sufficiently so that the cast part can be ejected with
out warp. The design of a successful die can be considerably affe
cted by perfect filling, which reduces the cooling time, reduces
warp and in turn increases the quality of the part. The main aim
of the cooling process is to maintain a uniform temperature of th
e filling and cooling cycle. Accordingly, there are at least two
important concepts for the designer when considering the cooling
system and in establishing cooling processing conditions: (1) ach
ieving uniform temperature and. (2) mini-mising the cycle time.To
achieve these two aims, the designer may need an optimal compute
r-aided design system to achieve a rapid and uniform cooling syst
em. The design of an optimal system needs analysis of 3D heat tra
nsfer during the filling and cooling processes.The thermal analys
is tool should predict the temperature gradient and deformation o
f the die-body.Generally speaking, traditional die design still d
epends on experience, due to the lack of analytic ability in moul
d flow and heat transfer, so the designer is unable to evaluate a
nd handle the deformation resulting from material and thermal exp
ansion and shrinkage of the die. The parameters of different cool
ing systems can cause large temperature gradients, and different
deformations.Although FEM software is capable of analysing the fi
llingflow and cooling conditions of pressure-injected metal and t
he heat stress, heat strain and temperature distribution conditio
ns of a die-casting die under various cooling systems, the establ
ishment of an analytic model is very difficult, especially for 3D
free-form geometry. Besides understanding the requirements of mu
lti-cavity dies, and the metal flow and solidification process, t
he designer should be fully acquaintanted with the basic finite e
lement software. Integration can be achieved and can save a lot o
f money and time only if a complete understanding of the process
of die manufacturing is available and eliminate the annoyance cau
sed by moving of personnel.Initially, consider the design of the
vent gate and overflow gate in the process of injection and flow
to fill the die cavity during cold room die-casting as investigat
ed by simulation by Garber [2]. When metal casting using a plunge
r into a cavity,he considered the change occurring in the metal,
and the replacement of the air in the cavity by molten metal. Sub
sequently, Garber [3,4] showed that too large or too small a plun
ger speed will affect the cast quality. Groenevelt and Kaiser [5]
studied and discussed the influence of the speed of injection of
the molten metal into the cavity, and flow distance and cavity t
emperature on the quality of product after casting. According to
the experiment, the distance of molten metal flow increases linea
rly as the die-casting speed increases, as does the die temperatu
re. The range of temperature is approximately 121–288oC. Other st
udies [6,7] proposed the importance of initial temperature (pre-h
eat temperature) of the die, and pointed out that too low a pre-h
eat temperature would tend to cause failure in filling up the cav
ity inside the die by the die-casting liquid, and result in forma
tion failure. A higher temperature may increase the cooling time
and reduce productivity. Truelove [8] used a cooling system to co
ntrol the overall temperature of the die, in order to obtain an o
ptimal heat transmission characteristic, and reduce the occurrenc
e of hard point phenomena in the cast piece, thereby improving th
e quality of the piece.Jong et al. [9] developed a mathematical E
q. for the flow and solidification of molten metal during high-pr
essure diecasting, in order to analyse the temperature conditions
and solidification strain of die-cast components in the cavity.K
enichiro et al. [10] used a finite element method to analyse and
design the die; the result was not only improved accuracy, but th
e factors to be considered are increased too, and the pressure of
die-casting, the speed of molten liquid flow, viscosity, and the
mechanical nature of the material changed with temperature and p
hase.This study uses CAD\CAE error software for a systemic design
process of a die, in order to minimise human error in die design
[11–13]. It uses the CAD software to create a freeform model, an
d the finite element software to analyse the conditions of die-ca
st processing. It simulates the temperature distribution of the d
ie-body and deformation after casting under various parameters (c
ooling-line distance R, channel center distance L, channel diamet
er D), as shown in Fig. 1. It uses an abductive network to establ
ish the relationship of the Fig. 1. Relationship between cooling
channel and free-form die. deformation and the cooling system par
ameters model. Based on the abductive modelling technique, it is
able to represent the complicated and uncertain relationships bet
ween the input and the output variables.Once the abductive networ
k has constructed the relationships of the input and output die-c
asting variables, an appropriate optimisation algorithm with a pe
rformance index is able to search for the optimal casting paramet
ers. In this paper, a sound optimisation method of simulated anne
aling [14] is adopted. The simulated annealing algorithm is a sim
ulation of the annealing process for minimising the performance i
ndex.It has been successfully applied to die-casting die design [
15], etc. The basic theory can be widely applied.2. Die-Casting F
low TheoryIn the die-casting process about 80% of the time is spe
nt in cooling cycle. The deformation of the die-casting die is ca
used by the non-uniform temperature distribution of the casting p
rocessing, which affects the quality of casting part. The designe
r of the cooling system has to think about the total cycle and co
mpute the deformation at every stage of the diecasting process. T
he die-casting process analysis includes three major stages: (1)
filling stage; (2) cooling and solidification stage; (3) ejection
, i.e. stress residue stage. Firstly, the casting processing has
to ensure that the melt fills the cavity. The major die-casting f
low equations are divided into five stages. In the filling stage,
the mould cavity fills with molten plastic fluid under high pres
sure. 3. Create the Relationship Between Cooling System and Die-C
asting Die DeformationThe die-casting pressure for high-pressure
injection in Al alloy die-casting is approximately 30–150 Mpa; ge
nerally the injection pressure varies with time. To examine the i
nfluence of the processing pressure, Dochler and Borton used a ca
thode ray oscilloscope and camera to analyse the pressure variati
on in the die-casting process, i.e. high-pressure filling, coolin
g and ejection.Design of a die-casting die involves the design of
a runner, cavity balance, analysis of life span of the die (resi
due stress),cooling system, etc. The purpose of this study is to
find the optimal cooling system of a die-casting die for casting
any workpiece. The assumed casting conditions are: die-casting pr
essure 120 Mpa, casting speed 2.8 m/s, die-casting cycling time 2
0 s/cycle, pre-heat die temperature 150oC, injection temperature
700oC. The basic assumption for the flow in the cavity is: (1) 3D
flow; (2) Newton fluid; (3) laminar flow; (4) incompressible flu
id; (5) zero speed of fluid in the vertical and horizontal wall d
irections.According to the different cooling parameters in the cl
osed section of the part surface, there are 15 sets of data to si
mulate die-casting processing. The basic configuration is a free-
form surface according to mould flow analysis of the 3D flow mode
l. The surface temperature is set at the pre-heat temperature (15
0oC) of the die, while the temperature of the molten metal is 700
oC. The different cooling parameters of the cooling system produc
e a total of 15 sets, as shown in Table 1.The method for heat tra
nsfer and deformation analysis is identical, the set injection te
mperatures are all 700oC, so it is only necessary to maintain a t
emperature of 700oC at the injection gate. The mould temperature
is 150oC, cooling water temperature is 40oC, and other temperatur
es are obtained using finite element analysis for the temperature
at the instant of filling up, as shown in Figs 2 and 3. Fig. 2.
Temperature gradient. Fig. 3. Deformation distribution.The deform
ation analysis uses a solid model analysed by 3D flow and non-lin
ear conditions, for finding its temperature at the boundary condi
tion required for performing the analysis. The configured tempera
ture of each node is input as the initial condition. Making the i
njection gate a boundary constraint condition serves as a 3D ther
mal strain condition. The mechanical properties change with the a
ccompanying the change of temperature. It is obtained from an ana
lysis of the non-linear stable stage. The results of the cavity f
low analysis discussed above performed by the strain analysis sta
ge are shown in Table 1.The parameters of the die-casting process
are complicated and hard to control. There is no definite determ
ination in the relation of each parameter and target function. It
is different when using experimental and statistical methods for
the condition of the actual die-casting. There is much restricti
on in the application. This study employs a neural network to lea
rn and train the network for the deformation of the die-casting d
ie, and the deformation of the die in the die-casting process, an
d uses this neural network to perform further analysis on each pa
rameter.Similarly, the establishment of the relation of the input
parameter (cooling system parameters: R, cooling line distance;
D, channel diameter; L, channel-centre distance) and output param
eter (deformation) during the die-casting process is shown in the
Appendix. To build a complete abductive network, the first requi
rement is to train the database. The information given by the inp
ut and output data must be sufficient. Thus the training factor (
cooling system parameters) for the abductive network training should be good and make defect-free product. Table 1 illustrates the cooling system parameters and the maximum deformation of the die-casting die obtained from 3D mould-flow analysis.Based on the development of the die-casting model, threelayer abductive networks, which are composed of cooling system parameters and the casting results (deformation), are synthesised automatically. The process is capable of predicting accurately the die-casting die deformation under various control parameters. All polynomial equations used in this network are listed in the Appendix (PSE = 5.43 X 10_7).Table 2 compares the error predicted by the abductive model and the simulation case. The simulation case is excluded from the 20 sets of simulation cases for establishing the model. This set of data is used to test the appropriateness of the model established above. We can see from Table 2 that the error is approximately 2%, which shows that the model established above is suitable for this purpose.7 |
|
