用RBF神经网络预测材料力学性能的研究

由于径向基函数(RBF)神经网络有易学,动态仿真性强,较强的输入输出映射功能和全局最优逼近的结构特点,因此将之用于预测麦杆增强复合板材力学性能。高斯函数表示形式简单,径向对称,光滑性好和解析性好,所以模型采用高斯函数作为隐含层基函数,k均值聚类法确定径向基函数的参数,运用最小二乘法确定权值。结合影响复合板材力学性能因素的特点和变化规律,以成型温度、成型压力、纤维含量、保温时间、拉伸强度、冲击韧性等为对象建立预测复合板材力学性能的模型,用它来优化模压成型的工艺参数,找出最佳工艺参数的范围。结果表明,径向基函数神经网络具有较好的学习和泛化能力,在预测力学性能中效果较好。     

基于小生境遗传神经网络的材料力学性能预测

对工业材料测试,建立优化系统,关于建立材料力学性能与组成、工艺等相关的预测模型,可以减少试验次数、提高效率、实现工艺的优化.提出利用遗传算法优化的神经网络建立材料性能影响因子到力学性能的非线性映射.在遗传算法中采用基于淘汰相似结构机制的小生境技术,使预定距离之内仅存一个优良个体,维护了群体的多样性,从而提高全局搜索能力.以麦杆增强复合材料为例进行仿真研究,建立其力学性能预测的小生境遗传神经网络模型,利用模型优化注塑成型的工艺参数进行仿真.结果表明所建模型具有较好的学习和泛化能力,对于优化成型工艺参数具有可行性,在材料性能研究领域有着较好的应用和推广价值.     

基于支持向量机的复合材料力学性能预测

通过建立材料力学性能与工艺参数相关的预测模型,可以减少试验次数、实现工艺优化,提高产品质量;该文提出利用支持向量机建立材料性能影响因素到力学性能的非线性映射;以麦杆增强复合材料为例,建立其力学性能预测的支持向量机模型,对材料的注塑工艺参数进行分析,得出其注塑成型的最佳工艺参数;结果表明所建模型具有较好的学习和泛化能力,对于优化成型工艺参数具有可行性,在材料性能研究领域有着较好的应用和推广价值.     

采用神经网络预估法建立板材力学性能预测模型

为提高产品质量,降低产品成本,开发了板材的屈服强度、抗拉强度、延伸率等力学性能的预测模型;介绍了建立热轧带钢力学性能质量模型的数据挖掘过程;用普通神经网络建立起由工艺参数预测力学性能的质量模型,模型预测结果的5%命中率是0.508;然后,提出一种新的建模方法--预估法,该方法是以三层BP神经网作为基本模型,通过增加模型层数,缩小底层子模型的预测范围,从而提高模型的泛化能力,这种方法的关键问题是能够对测试数据实现正确分类;利用该方法建立起质量模型,预测结果的5%命中率达到0.706,完全可以满足现实生产需要.     

人工神经网络在带钢力学性能预测中的应用

本文运用神经网络原理建立了带钢力学性能预测的神经网络模型，并用该模型对某钢厂带钢力学性能进行了预测，还对预测结果和真实数据进行了比较，结果表明：相对误差低于1%，而且收敛速度快，泛化能力好。预测模型可减少或取消实际生产中的破坏性检测试验，从而提高经济效益。另外，文中还总结了提高神经网络性能的方法。     

基于遗传算法优化投影寻踪技术的神经网络集成模型及其应用

首先利用遗传算法优化的投影寻踪技术对神经网络学习矩阵降维,再利用Bagging技术和不同的神经网络学习算法生成集成个体,并再次用遗传算法进化的投影寻踪技术对神经网络个体集成.建立基于遗传算法优化的投影寻踪技术神经网络集成模型,通过上证指数开盘价、收盘价进行实例分析,计算结果表明该方法具有较好的学习能力和泛化能力,在股市预测中预测精度高、稳定性好.     

基于支持向量机的中药工艺参数优化研究

提出了基于SVM的滴丸生产工艺参数优化方法,较好地预测了滴丸含水量,给出了各工艺参数取值范围,在实际生产中取得了良好效果。理论分析和仿真研究表明,该方法学习速度快、跟踪性能好、泛化能力强、对样本的依赖程度低,比基于BP神经网络的建模具有更好的推广能力。     

最小一乘回归神经网络集成方法股市建模研究

提出了一种新的神经网络集成股市建模方法,采用偏最小二乘方法构造神经网络输入矩阵,利用Bagging技术和不同的神经网络学习算法生成集成个体,再用遗传算法选择参与集成的个体,以"误差绝对值和最小"为最优准,建立最小一乘回归神经网络集成模型,通过上证指数开盘价、收盘价进行实例分析,计算结果表明该方法具有较好的学习能力和泛化能力,在股市预测中预测精度高、稳定性好.     

基于AGA-DBSCAN优化的RBF神经网络构造煤厚度预测方法

在构造煤厚度的预测中,经常出现因各种限制性因素而导致预测精度不高的问题,因此提出了利用自适应遗传算法优化密度聚类(DBSCAN)优化RBF神经网络参数的方法对构造煤厚度进行预测.首先,对采区三维地震属性数据进行预处理,采用主成分分析算法(PCA)对该数据降维并消除变量之间的线性相关性.然后,构建预测构造煤厚度的RBF神经网络模型,并利用DBSCAN获取最佳核心点数据,通过计算得到k-means聚类的初始聚类中心,以此优化k-means算法,进而得到RBF神经网络隐含层基函数最优的中心向量,提高该模型预测的精准性和鲁棒性.同时,针对遗传算法存在容易陷入局部最优的问题,通过随着进化次数的增多自适应地改变交叉率和变异率来改善遗传算法的全局和局部搜索能力,使之逃离局部最优点,获得更优的进化结果.此外,为了增强模型的泛化能力,对模型权重参数加入了L2正则化项,有效避免了噪声对模型泛化能力的影响.最后,将该模型应用到芦岭煤矿II六采区8#煤层中,模型预测构造煤的厚度与实际地质资料具有较高的一致性.因此,所提构造煤厚度预测模型的实际预测精度较高、误差较小,可以推广到实际采区构造煤厚度的预测.     

基于神经网络预测模型输入参数配置方法的实现

基于数据挖掘中的关联概念,提出了一种针对神经网络预测模型训练参数的选择方法,有效地提高了神经网络模型在毛纺工艺中对纱线断头率的预测精度;该方法通过生产中的训练参数记录进行关联规则的提取,可快速的排除产生负面影响的训练参数,迅速选择可以提高预测精度的训练参数,从而达到提高神经网络模型预测性能的目的;实验证明,利用关联算法进行参数配置,可以有效提高神经网络输入模型的预测精度.     

New approaches in microcasting: permanent mold casting and composite casting

For several years, microcasting was based on investment casting. New approaches are now the permanent mold casting and composite casting of micro parts. Casting was performed with aluminum bronze of the type CuAl₁₀Ni₅Fe₄. Permanent mold casting was commenced with steel mold inserts in a lost mold. The development of a band heater enabled the heating of permanent molds inside the casting machine. This shall ensure sufficient form filling of micro cavities. For permanent mold casting micro structured steel molds and graphite molds were used. Composite casting was investigated for surrounding a micro part by melt, and for pouring a metal melt into a micro structured part. In both cases different materials were combined. For metal–ceramic composite casting ZrO₂, Al₂O₃ or Si₃N₄ were used in connection with Al bronze. For metal–metal composite casting Al bronze was cast around steel. First results for both new approaches are presented.     

Study of Hot Embossing Using Nickel and Ni–PTFE LIGA Mold Inserts

Hot embossing is one of the main process techniques for polymer microfabrication, which helps X-ray lithography, electroplating, and molding (LIGA) to achieve low-cost mass production. Most problems in polymer micromolding are caused by demolding, especially for hot embossing of high-aspect-ratio microstructures. The demolding forces are related to the sidewall roughness of the mold insert, the interfacial adhesion, and the thermal shrinkage stress between the mold insert and the polymer. The incorporation of polytetrafluoroethylene (PTFE) particles into a nickel matrix can have the properties such as antiadhesiveness, low friction, good wear, etc. To minimize the demolding forces and to obtain high-quality polymer replicas, a Ni-PTFE composite microelectroforming has been developed, and the hot embossing process using Ni and Ni-PTFE LIGA mold inserts has been well studied in this paper. The morphologies, sidewall roughness, and friction coefficient have been explored in the fabricated Ni-PTFE LIGA mold insert. Finally, the comparison of embossed microstructures with various aspect ratios and the comparison of the embossing lifetimes of mold inserts have been carried out between Ni and Ni-PTFE mold inserts, which show a better performance of the Ni-PTFE mold and its potential applications.     

基于模糊推理神经网络的诊断模型应用研究

模糊推理系统能处理和利用不精确的知识,是一种有效的智能建模方法.但是模糊推理系统缺乏有效的设计方法.本文通用神经网络的自适应性和诊断的建模方法,建立了一种新的故障诊断模型-模糊神经网络诊断模型.     

Micromolding fabrication of microresistors with a composite of carbon nanotubes and SU-8 polymer and the application in Wilkinson power divider

This paper reports a new method to fabricate microresistors for applications in micro-electromechanical systems. The fabrication is based on ultraviolet (UV) lithography and micromolding replication. A master mold was first made using UV lithography of a negative-tune photoresist, SU-8. The SU-8 master mold was then used to produce a polydimethylsiloxane (PDMS) intermediate mold. The PDMS intermediate mold was used to replicate the microresistor using composite mixture of multi-walled carbon nanotubes (MWCNTs) and SU-8 photoresist. The replicated microresistors were thermally cured at 95 °C for 6 h. The performances of the replicated micro-resistors were then tested. The experiment had also been conducted using the micro-resistors as isolation resistance in a Wilkinson power divider to test their functionality. Because passive communication components such as Wilkinson power dividers can potentially be made with lithography/micromolding technologies and using polymer as structural material, microresistors as reported in this paper may potentially be suitable for using in such applications.     

并行环境下注塑件智能设计支撑技术研究

注塑件设计是注塑模设计制造的第一步,注塑件设计的优劣以及注塑件产品模型信息的完备与否,对注塑模设计制造的难易以及成形工艺参数的选择都具有十分重要的作用。本文从构造新一代注塑模设计制造系统的出发,探讨注塑件智能设计的支撑技术,提出采用并行工程设计思想,以基于特征的产品建模方法构造注塑件产品模型;结合神经网络和专家系统的特长,构造神经网络与专家系统混合系统;利用特征和知识紧密的关系,实现注塑件以及注塑     

基于Internet的网络课堂系统的设计与开发

本文从分析网络课程设计的现状出发,对网络课程的内涵及其教学本质进行了探讨,提出了网络课程教学的概念.网络课程是网上教学过程的重要环节,是教学评价的主要途径.利用计算机、网络技术构建的基于校园网络的在线课程设计、在线练习、在线答疑等模块,不仅可以提高网络课件的交互性,同时也提高了网络课堂教学的质量.     

C＆C08交换机企业专网与公网直连的实现

本文以宣钢局在用华为32模C＆C08数字程控电话交换机为例,介绍了企业专网与公网程控局直连数据配置的实现步骤,以及数据配置参数。     

基于入侵检测系统构建的研究

随着网络技术的快速发展,网络入侵事件的发生也渐渐的增多。从网络安全立体、纵深、多层次防御的角度出发,入侵检测系统和技术得到的高度重视。另议方面随着技术的发展,网络日趋复杂,防火墙技术所表现出来的不足引出了人们对入侵检测系统(IDS)技术的研究和开发。该文就入侵检测系统定义、入侵检测系统分类、入侵检测系统功能模块作了一定的分析,并对入侵检测系统作了简单的前景预测。     

The composite banyan network

A new multipath multistage interconnection network called the composite banyan network is proposed. The network incorporates both the banyan and the reverse banyan networks and is constructed by superimposing the two. The basic building blocks in the composite banyan network are 3×3 switching elements with log₂N stages. A major advantage of the composite banyan network over existing networks with 3×3 SEs is an efficient and fast control algorithm that sets up a path between any source and destination pair. Instead of complex numerical calculations, the network can easily generate a primary routing tag and alternate tags through simple binary operations. Also, the network has a lot of favorable features, including regularity, symmetry, and easy rerouting capability under faults and conflicts. It is shown that at least two totally disjoint paths exist between any source and destination pair, which increase the degree of fault-tolerance. A deterministic permutation routing algorithm is also developed for the 8×8 composite banyan network, Using a simple tabular method, it is shown that the algorithm always finds a set of conflict-free tags     

Determination of back-pressure profile for forward extrusion using sequential approximate optimization

This paper proposes a method for determining optimal back-pressure profile in forging of aluminum alloy using a sequential approximate optimization (SAO). In forging, it is important to improve the mold filling for the product quality. In addition, it is preferable to produce a product with a minimum forming energy. To achieve these objectives simultaneously, a forging method with back-pressure profile is proposed. Here, the back-pressure profile implies that the back-pressure varies through the stroke. In this paper, a multi-objective optimization (MOO) problem is formulated. To improve the mold filling, an unfilled area is taken as the first objective function. Furthermore, a forming energy during the forging is taken as the second objective function. Numerical simulation in the forging is so expensive that the SAO using the radial basis function (RBF) network is adopted, and the pareto-frontier is identified with a small number of simulation runs. Based on the numerical result, the experiments are also conducted. It can be found from these results that, the back-pressure profile approach is valid for improving the mold filling as well as the forming energy.