【作者】 成志强；

【导师】 柳葆生；

【作者基本信息】 西南交通大学 ， 固体力学， 2006， 博士

【摘要】 金属注射成形(MIM)是一种制造复杂三维小零件的近净成形技术。与其它的加工艺相比,如压模成型、机加工和铸造等,MIM技术能够低成本地生产高机械性能的复杂零件,因此有着明显的优势。MIM工艺主要包括四个过程:1)填料的混合 2)注射 3)脱塑 4)烧结。 MIM作为一种新的制造技术,数值模拟在其高效率的应用上有着重要作用。在MIM各个工艺阶段的模拟中,注射阶段的模拟显得更为重要。因为在这个阶段出现的任何缺陷,在接下来的脱塑和烧结阶段都会被放大,以至于影响产品的最终质量。温度、压力等注射工艺参数的正确选取及合理的模具设计是保证产品质量的关键。金属注射成形的数值模拟软件提供了预测填充过程任意时刻填料的流动前沿面、速度场、压力场和温度场的强大功能。除了能够预测注塑成形过程的各种缺陷外,金属注射成形过程中的离析现象在双相流模型中也能够被预测。离析是脱塑过程中产生塌陷和烧结过程中导致翘曲、变形的根本原因。 为了预测金属注射成形过程中的粉末离析现象,引入混合理论是必要的,这样求解两个耦合的Navier-Stokes方程不可避免。如果应用传统的算法,计算量会异常的繁重以至于影响其实际应用。为了达到高效模拟的目的,专门用于模拟金属注射填充的矢量化显式算法首先得以实施。在此基础上,扩展为针对双相流的矢量化算法以模拟粉末离析现象。在新的双相流算法中,既没有全局求解也没有构筑任何全局矩阵,矩阵操作仅在单元一级进行。从整体来看,全局操作完全矢量化,因此计算效率非常高。新算法的另一显著特点,是采用等阶插值的普通单元代替MINI单元,不可压缩条件通过反馈修正的方法得到满足,通过系统化的光顺处理来避免由于采用等阶插值而导致的自锁。同时引入了另外一个光顺处理用来避免由于采用离散的有限元模型引起的初始填充阶段的计算不稳定。新算法在求解大规模工业问题上显示出其明显的计算速度优更多还原

【Abstract】 Metal Injection Moulding (MTM) is a near net shape technology to manufacture the small intricate parts in 3D shapes. Compared with other fabrication methods, such as die compaction, machining and casting etc, MIM technology has some evident advantages to produce complex components with high mechanical properties and low costs. The MIM technology includes four main stages: 1) mixture of feedstock. 2) injection moulding to get the green parts. 3) debinding stage to remove the binder. 4) final sintering stage to get the condensed part in pure metallic material.As a new manufacturing technology, numerical simulation plays an important role in its efficient applications. Among the simulations of different stages, the simulation of injection moulding is perhaps the most important one. Any defects appearing in this stage will be amplified in the next debinding and sintering stage to corrupt the quality of final parts. The properly chosen process parameters in injection stage, such as the temperature, pressure and well-designed mould configurations, are essential for ensuring the quality of produced parts. The numerical simulation of MIM injection provides a powerful tool to predict the filling state, velocity fields, pressure field and temperature field in the mould cavities at each instant of the injection process. Except for predicting the defects that may happen commonly in the process of polymer injection, the powder segregation effect in MIM injection can be predicted by the simulation based on a bi-phasic model. The segregation effect is the source of collapse in debinding, as well as the warpage and distortion in sintering.For the prediction of powder segregation in injection stage of MIM, the application of mixture theory is necessary, which induces two coupled Navier-Stokes equations. The computation becomes an extremely heavy charge for the traditional algorithms, which prevents severely its real application. To achieve the efficientMIM simulation, a fully vectorial explicit algorithm for simulation of the rilling process is proposed firstly, then the important extension to bi-phasic simulation for the prediction of powder segregation is proposed to realize a fully explicit and vectorial solver. In the new solver for simulation of the bi-phasic injection model, there is no more global solution or the construction of any global matrix except the small operations at element level. In global sense, the computation is carried out only by the vectorial operations. So the computation is very efficient. Moreover, the new solver employs the elements of equal-order interpolations instead of the MINI elements. Incompressibility of the mixture is retained by a corrective feedback operation, with a procedure of systematic smoothing to avoid the possibility of mesh locking. Another smoothing method is proposed to avoid the instability at starting stage of the filling simulation. This instability is caused by the sudden change of filled zone because of the discretization of element model. The developed new solver shows the proper advantages for simulation of the industrial problems in large scale. Furthermore, the software is easy to be parallelized on a high performance system with multi-clusters because of its vectorial feature. The comparison of numerical results with experiments and previous simulations proves the validity and efficiency of new algorithm.更多还原