基于知识模板的复杂产品设计重用方法研究

为实现复杂产品设计知识的莺用,提出了基于面向对象技术的知识模板概念.用统一建模语言建立了复杂产品设计阶段的功能-行为-结构模型.通过功能分解和结构映射,将功能信息和向何信息集成封装在一个类模板中.基于功能域、行为域和结构域三个映射域,设计了功能分解算法,从而形成可重用知识.知识模板通过组成模板的静态和动态知识提取并加以综合获得,利用相似度数值的计算进行模板的选择与匹配.最后以叉车产品功能分解和结构映射所形成的知识模板,说明了所提出的设计重用方法的实际应用.     

Selectivity of Silicon Carbide/Stainless Steel Solid-State Reactions and Discontinuous Decomposition of Silicon Carbide

Solid-state reactions of SiC with a multicomponent system, stainless steel, have been studied at 1125°C. Four-layered reaction products consisting of modulated carbon precipitation zone, random carbon precipitation zone, four-phase mixture zone, and grain-boundary precipitation zone were formed in the reaction zone. The carbon precipitates were embedded in a matrix of complex metal silicides. In addition, extensive interfacial melting was noted. Carbon atom was found to diffuse faster than Si and selectively reacted with Cr to form Cr carbide(s) along the grain boundaries of stainless steel. No Fe carbides or Ni carbides were ever detected. Among the consituents existing in stainless steel, Ni atoms have the highest affinity for Si. An uphill diffusion of Ni toward the SiC reaction front was observed. While the diffusion of Cr and Fe toward SiC followed a downhill concentration gradient, very small amounts of Cr reached the SiC interface. The selective reactions of Si and C with Ni, Fe, and Cr are discussed on the basis of Gibbs free energy of formation of various compounds. The diffusion kinetics of C and Si atoms in selected metal/SiC reactions are discussed on the basis of their chemical affinities for respective metals. The modulation of carbon precipitation is correlated with previous results from Ni/SiC, Ni₃Al/SiC, Fe/SiC, Co/SiC, and Pt/SiC, reactions. A general model describing discontinuous decomposition of SiC is proposed to explain the origin of carbon modulation.     

Coherent Phase Diagram for Spinodal Decomposition in the Tetragonal Titanium Dioxide—Tin Oxide System

A microscopic calculation was carried out for spinodal decomposition in the tetragonal TiO₂—SnO₂ system by taking into account the contribution of the elastic free energy. Necessary elastic constants for the solid solutions and the elastic free energy of modulated structure were calculated in terms of interatomic potentials. A coherent phase diagram was obtained unambiguously from the calculated elastic free energy by using the interaction parameter determined by the experimental binodal line. The results were compared with the calculation based on the regular-solution model.     

金属有机物分解法制备钛酸铅薄膜

用金属有机物分解法制备了钛酸铅薄膜。研究了成膜的技术条件。一次甩胶可以获得厚约0.1μm均匀、无开裂、无针孔的薄膜,较厚的膜可通过多次甩胶得到。X射线衍射证实,经700℃热处理后,膜呈现四方钙钛矿的单相结构。Rutherford背散射表明,即使在多次甩胶制得的膜中,组份的厚度分布也是均匀的。     

方向性纹理织物疵点检测方法研究

通过分析方向性织物纹理的特点,提出了一种织物疵点检测新的方法。首先根据正常纹理Hough变换确定织物纹理的纹路方向;然后采用方向性小波对织物纹理图像进行方向性的分解,并在此基础上从分解后的各细节子图中提取子窗口的特征;最后通过BP神经网络进行织物疵点识别。实验结果表明了该方法的有效性。     

电压闪变影响下的功率计算模型与电能计量方法研究

当前针对谐波电能计量技术开展了大量研究,而含有电压波动与闪变畸变信号影响下的电能计量技术研究较少。文章构建了电压波动与闪变影响下的功率数学模型,推导功率的简化计算式,提出了基于小波包分解与重构的电压波动与闪变下的电能计量方法;利用Db45小波函数,采用5层小波包变换,进行电压波动与闪变影响下稳态、非稳态信号的有功与无功电能计量的仿真分析,实验结果证明了算法的可行性与准确性。     

小波分析在低压负荷预测数据处理中的应用

低压电力负荷预测在保证低压用户的用电可靠性及经济性上具有积极的意义。通过对历史数据分析可清楚了解低压用户用电负荷规律,为企业用电规划提供保障。小波变换对含白噪声信号具有良好的滤波能力,使用小波处理历史负荷数据。采用db3小波基函数对扩展后的原始信号进行三层小波分解,对比硬、软阈值降噪效果后选取软阈值降噪更好实现滤波效果,达到了使数据波形平滑的目的,为负荷预测模型选择提供可靠依据。     

计及电量分解的高比例水电电网短期调度方法

中长期交易电量与短期发电计划衔接不当是引发结构性弃水或缺电的关键问题之一,如何分解交易电量96点曲线对于高比例水电电网发电计划编制至关重要.本文依托云南电网实际工程,提出计及电量曲线分解的短期发电调度方法,引入电站分类策略以贯序确定标准负荷曲线,兼顾公平性提出耦合标准负荷曲线的日电量分解方法,并以交易电量完成度相对偏差...     

基于混合方式的贝叶斯网络等价类学习算法

贝叶斯网络(BN)是不确定知识表示和推理的主要方法之一,是人工智能中重要的理论模型.针对现有混合方法学习BN结构不稳定、容易陷入局部最优等问题,本文将图论中的最大主子图分解理论与条件独立(CI)测试相结合,同时引入少量的局部评分搜索,提出一种新的基于混合方式的BN等价类学习算法.新算法通过确定所有变量的Markov边界构造网络的无向独立图,并对无向图进行最大主子图分解,从而将高维的结构学习问题转化为低维问题,然后利用低阶CI测试和局部评分搜索识别子图中的V结构.理论证明以及实验分析显示了新算法的正确性和有效性.     

基于EEMD近似熵的水电机组振动信号特征提取

对于水电机组非平稳非线性振动信号特征提取方法的研究近年来一直是水电机组故障诊断领域研究热点,特征提取的有效性直接关系到故障诊断的准确性。本文提出基于集合经验模态分解(EEMD)和近似熵的水电机组振动信号特征提取方法,将信号经EEMD分解后筛选得到的本征模态分量(IMF)近似熵特征值输入概率神经网络(PNN)进行模式识别。采用经验模态分解(EMD)和近似熵特征提取方法进行对比实验。识别结果表明:采用EEMD和近似熵的特征提取方法,能有效区分机组不同的运行状态,可为实际工程应用提供理论依据。