The selection of optimal cooling parameters for the multi-cavity die

The subject of this study is a finite element and abductive network method application for the multi-cavity die. In order to select the optimal cooling system parameters to minimise the warp of a die-casting die, the Taguchis method and the abductive network are used. These methods are applied to create an efficient model with functional nodes for the considered problem. Once the cooling system parameters are developed, this network can be used to predict the warp for the die-casting die accurately. A simulated annealing (SA) optimisation algorithm with a performance index is then applied to the neural network for searching for the optimal cooling system parameters, and obtains a rather satisfactory result as compared with the corresponding finite element verification.     

<Emphasis Type="Bold">The Optimal Design of a Cooling System for a Die-Casting Die With a Free Form Surface</Emphasis>

This study is on the finite element and abductive network method application to die-casting dies with free-form surfaces. The study aims to find the optimal cooling system parameters and decrease in deformation of a die-casting die. In order to avoid the numerous influencing factors, the free-form surface of a die-casting die is created as a non-linear Eq. of a polynomial function. The parameters of the cooling system, including the channel space and channel diameter, are adjusted according to the non-linear Eq.. An abductive network has been built for modelling the die-casting cooling parameters. The abductive network is composed of a number of functional nodes. Once the cooling system parameters are given, this network can predict the deformation of the die-casting accurately. A simulated annealing optimisation algorithm with a performance index is then applied to the neural network for searching for the optimal cooling system parameters and to obtain a satisfactory result.     

Digitized die forming system for sheet metal and springback minimizing technique

The design of the runner and gating systems is of great importance to achieving a successful injection moulding process. The subjects of this study are the finite element and abductive neural network methods applied to the analysis of a multi-cavity injection mould. In order to select the optimal runner system parameters to minimize the warp of an injection mould, FEM, Taguchi’s method and an abductive network are used. These methods are applied to train the abductive neural network. Once the runner and gate system parameters are developed, this network can be used to accurately predict the warp of the multi-injection mould. A simulated annealing (SA) optimization algorithm with a performance index is then applied to the neural network in order to search the gate and runner system parameters. This method obtains a satisfactory result as compared with the corresponding finite element verification.     

Optimum Gate Design of FreeForm Injection Mould using the Abductive Network

This study uses the injection position and size of the gate as the major control parameters for a simulated injection-mould. Once the injection parameters (gate size and gate position) are given, the product performance (deformation) can be accurately predicted by the abductive network developed. To avoid the numerous influencing factors, first the part-line of the parameter equation is created by an abductive network to limit the range of the gate. The optimal injection parameters can be searched for by a simulation annealing (SA) optimisation algorithm, with a performance index, to obtain a perfect part. The major purpose is searching for the optimal gate location on the part-surface, and minimising the air-trap and deformation after part formation. This study also uses a practical example which has been and proved by experiment to achieve a satisfactory result.     

The runner optimisation design of a die-casting die and the part produced

The accurate prediction of die-casting performance is essential for the design of die-casting dies. The main purpose of die design is to produce an accurate part, but a key role in a good design is also the design of the runner. In order to understand the practical conditions in die casting, this paper studies the runner part and die body parts which have been considered separately during the experiment. A die-casting test has been designed according to experimental methods and the runners have a total of 23 sets of tests.A finite-element method was used to obtain the displacement from the measurement of the temperature in the cavity, and an abductive network has been built for modelling the diecasting process. The abductive network is composed of a number of functional nodes. Once the die-casting parameters (high-speed injected position, runner injected angle, runner sectional ratio) are given, the die-casting performance (the displacement) can be accurately predicted by this network. The optimal die-casting parameters can be searched for by a simulation annealing optimisation algorithm using a performance index to obtain a perfect part.     

The selection of the optimal design parameters applied to the gas-assisted injection mould of an L-shape plates design

This study examined the finite element and abductive network method applied to thegas-assisted injection mould (GAIM). From these models, the effect of the GAIM condition includes the melt formation by percent volume filled, the base-plane thickness, and the ratio of L-shape plane thickness. The L-shape plane can reduce by about 50 per cent the expense of materials, reduce the injection cycle time, and increase the L-shape structure. In order to select the optimal design parameters to minimize the warp of a GAIM, the experimental design method and abductive network were used. These methods were applied to create an efficient model with functional nodes. Once the GAIM parameters were developed, this network was used to predict GAIM warp accurately. A simulated annealing (SA) optimization algorithm with a performance index was then applied to the neural network for searching the optimal GAIM parameters, obtaining a more satisfactory result compared to the corresponding finite element verification.     

Design of a Universal Fixture for Laser Scanning

In this paper, a universal fixture for 3D laser scanning is presented. In order to reconstruct the surface or CAD model from the scanned data of an existing part, it is necessary to integrate the point data from several views of an object into a common reference frame. When the registration is implemented with a feature-based method using tooling balls, the reference coordinate system is established by several common tooling balls from different point clouds. The laser scanner is often used to acquire the surface information of the part, but is limited in its measuring direction, which is fixed only along the z-axis. A universal fixture using several joints and an adjustable tooling ball holder is designed considering the convenience of the part set-up and the accuracy of the registration. The location of the tooling balls can be arranged to avoid the occlusion of the part and to minimise the registration error. The universal fixture was applied to a wrap-around part having freeform surfaces to demonstrate the effectiveness of the proposed design.     

基于贝叶斯推理的浇注系统选型方法研究

针对压铸模浇注系统设计过程中存在大量不精确数据的情况,提出基于贝叶斯推理算法的压铸模浇口类型选择方法.结合浇注系统的类型划分,建立相应的类型标准模型以及典型属性模糊向量,并采用模糊数学的方法处理设计信息中不精确约束值,最后将贝叶斯理论应用于推理过程中.该方法将概率逻辑与模糊数据处理的思想结合在一起,将贝叶斯推理方法应用于基于智能化设计的浇注排溢系统中,提高了对设计信息的获取能力以及产品设计的智能化程度.     

通用量子门神经网络在齿轮故障诊断中的应用

为进一步提高齿轮故障诊断能力,结合目前神经网络机理的研究进展,建立了一种基于通用量子门的量子神经元模型,提出了通用量子门神经网络(universal  quantum  gate neural network,UQGN)算法。首先,该算法将转换后的量子态训练样本作为输入。然后,利用量子旋转门和通用量子门完成旋转、选择、翻转和聚合等一系列操作,并完成网络参数的更新。最后,将训练后的结果输出。在数学上,证明了UQGN算法的泛化能力。利用该算法对齿轮的正常、齿面磨损、齿根裂纹和断齿4种情况进行了模式识别。实验结果表明,与普通神经网络和普通量子神经网络相比,UQGN算法在泛化性能、鲁棒性、准确率和执行时间等方面具有较好的效果。     

基于Moldflow的电器面板翘曲分析

以电器面板外壳为实例,针对目前模具注塑过程中发生的收缩翘曲问题,通过Moldflow软件进行模拟分析,给出了注塑件影响翘曲的主要因素。主要是对电器面板进行工艺分析,建立翘曲理论公式考虑影响翘曲的主要因素,确定合理的浇口方案,然后通过Moldflow对两个不同的浇口方案进行分析比较,包括纤维的取向分析以及各方向变形量进行一定的比较。软件研究表明,浇口的位置和数量对翘曲变形有很大程度的影响,优化浇口方案可以减小翘曲量。     

Feature-Based Parametric Design of a Gating System for a Die-Casting Die

This paper presents a semi-automated approach for designing the gating system of a die-casting die. This approach combines the P-Q 2 technique and feature-based parametric design for achieving the automatic generation of the geometries for the gating system. It also enables users to integrate their expertise in the design at an early stage. User-defined gating features are pre-defined and stored in a gating library, and are then retrieved from the library and applied to the gating part with the desired parameters and locations during the design process. Algorithms based on the P-Q 2 technique are proposed for calculating the related parameters of the gating system. A prototype design has been developed using this approach, and the system is able to shorten the time for constructing the 3D geometries of the gating elements such as the gate, runner, overflow, shot sleeve, sprue, etc.     

Automatic Detection of UXO from Airborne Magnetic Data Using a Neural Network

Recent developments in airborne magnetic detection systems have made it possible to detect small ferro-metallic objects. However, airborne magnetic data can be really large and, therefore, there is an increasing need for a fully automatic interpretation technique that could be used to make decisions regarding the nature of the sources in the field in real time. The massively parallel processing advantage of artificial neural networks makes them suitable for hardware implementations; therefore, using these networks in conjunction with a magnetic system has the potential to greatly speed up the detection of ferro-metallic objects. In this paper, we have developed a new method for detection and characterization of unexploded ordnance (UXO) using a Hopfield neural network as applied to airborne magnetic data. The Hopfield network is used to optimize the magnetic moment of a dipole source representing the magnetic object at regular locations. For each location, the Hopfield network reaches its stable energy state. The location of the object corresponds to the location yielding the minimum Hopfield energy. Output results include position in two dimensions (horizontal location and depth), magnetic moment, and effective inclination. Theoretical and actual field examples show that the Hopfield network is accurate and objective tool for the detection of UXO. Moreover, because the Hopfield neural network is a natural analog-to-digital converter, it is ideally suited for incorporation into airborne magnetic instrumentation systems.     

Laser Based Flexible Fabrication of Functionally Graded Mould Inserts

A CO₂ laser-based freeform fabrication process with emphasis on difficult-to-shape and functionally effective materials was investigated with regard to fabrication of dies and moulds. Square and circular moulds were built by use of a material additive process of layers of TiC and Ni-alloy composite. The effects of laser processing on the quality, microstructures, and hardness of the moulds were determined. Additionally, the mould performance was evaluated in die-casting, injection moulding, and in a thermal fatigue environment. The TiC core-Ni-alloy shell mould outperformed moulds made of hardened H13 die steel, 304 stainless steel, and TiC-coated stainless steel in withstanding stresses and in retaining dimensional stability at elevated temperatures.     

Prediction of surface reconstruction uncertainties for freeform surface inspection

A method is introduced to predict uncertainties of the B-spline freeform surfaces that are reconstructed from the measurement points for inspection of freeform surfaces. The uncertainties of a reconstructed B-spline surface are modeled by variances of coordinates of points on this freeform surface. Prediction of the uncertainty at any location on the reconstructed B-spline surface is carried out in two steps: (1) estimation of variances of the B-spline surface’s control points introduced by the surface reconstruction process, and (2) propagation of the variances from the control points to the points on the B-spline freeform surface. In this research, the variances of the control points of the reconstructed B-spline surface in all three directions are considered to improve the accuracy of uncertainty prediction in the regions with significant changes of geometric shapes. Both the errors and uncertainties at different locations on the reconstructed surface are considered in freeform surface inspection to compare the manufactured surface against the design surface and its tolerance. The developed method has been applied to two case studies to demonstrate its effectiveness.     

MATLAB在注塑件翘曲变形工艺参数优化中的应用

本文利用BP神经网络对注塑工艺参数及其相对应的翘曲变形量样本进行训练,得到了描述工艺参数到翘曲量映射关系的ANN模型,并且验证了此模型的准确性,得出了工艺参数与塑件注塑翘曲变形量的内在联系。     

Cover die service life improvement by biomimetic laser-remelting process and CAE simulation

Die service life improvement is an important problem in high-pressure die-casting industry. As many factors affecting the service life, die service life improvement is a systematic project. In this paper, a cover die-casting die was investigated in order to increase its life time. Different die materials, computer-aided engineering (CAE) analysis, biomimetic laser-remelting process, and die structure optimization are adopted in order to enhance its performance. The die service lives are traced for dies of different die materials (H13 and SKD61). The die service lives are compared for dies with/without structure optimization with the help of CAE simulation. The application shows that CAE simulation is a useful tool for die structure optimization. Experiment result on SKD61 die steel after laser-remelting process shows that biomimetic laser-remelting process provides desirable microstructural and microhardness changes in biomimetic units, which induces the intensified particles for improving the service life.     

制造领域中反向工程曲面推理的神经网络法

将神经网络技术用于制造领域中反向工程曲面的推理,建立了神经网络计算模型。使用神经网络方法推理实体表面的突出优点是不需要表面知识,只需要测得表面的有限点数即可,网络将学习参数变量之间的组合关系,推断出曲面的函数形式,计算出未知坐标点的值,避免了许多的人为假设和繁杂的数学计算,体现出一种智能推理过程。本文列举了有关推理实例,包括初等解析曲面和自由曲面,从理论和实际计算证明了这种方法的可行性和有效性。结果表明,神经网络对于反向工程曲面的推理是一种非常好的辅助分析手段。这对于复杂的、不完整的、部分磨损或损坏的实体表面具有实际用途。     

基于神经网络的汽车发动机机械设备故障定位

为提升发动机故障定位方法的定位性能,提出基于神经网络的汽车发动机机械设备故障定位。该方法首先依据汽车发动机具体结构,分析汽车发动机的各项故障类型,结合小波原理分析发动机振动信号,获取发动机故障特征向量值;再依据获取的特征向量,建立基于人工神经网络的汽车发动机故障定位模型;最后通过遗传算法更新模型参数,计算模型输出值,实现发动机的故障定位。结果表明,运用该方法对汽车发动机实施故障定位时,故障定位精度高、定位效果好以及定位性能优秀。     

高速经编机电子送经控制系统的设计

为了保证经编机在高速运转下送经量的恒定,运用当今经编机电子送经系统中最常见的送经原理,即利用反馈信号构建闭环控制系统,从而实现了同一送经量下以恒定线速度进行反馈控制,设计了一种基于高级精简机器指令(ARM)和现场可编程门阵列(FPGA)控制的高速经编机电子送经控制系统,并运用了PID控制理论.研究结果表明,该系统能实现...     

CL surface deformation approach for a 5-axis tool path generation

Pressure die-casting condition selection mainly relies on the experience and expertise of individuals working in production industries. Systematic knowledge accumulation regarding the manufacturing process is essential in order to obtain optimal process conditions. It is not safe a priori to presume that rules of thumb, which are widely used on the shop floor, always lead to fast prototype production calibration and to increased productivity. Thus, neural network meta-models are suggested in this work in order to generalise from examples connecting input process variables, such as gate velocity, mould temperature, etc., to output variables, such as filling time, solidification time, defects, etc. These examples, or knowledge, are gathered from experiments conducted on casting simulation software, which are designed systematically using orthogonal arrays (DoE). They could also be based on experiments from industrial practice. Neural models derived in this way can help in avoiding excessive numbers of what-if scenarios examined on the casting simulation software, which can be very time-consuming. Furthermore, they can be employed in the fitness function of a genetic algorithm that can optimise the process, i.e. yield the combination of input parameters which achieves the best output parameter values.