基于设计特征模型的加工特征映射技术

为了实现数控编程过程中能够自动确定加工区域,对基于特征设计系统构建产品模型的方法、产品信息模型、信息模型间的关联关系、飞机结构件加工特征定义分类、设计特征与加工特征间的内在关系进行了深入的研究,提出了基于设计特征模型映射加工特征的方法.该方法通过遍历特征设计树,判断特征间的关系,查询特征组成面信息,从而构建出特征属性邻接图,进而采用基于图匹配的特征识别方法得到加工特征,并提取加工特征信息.该方法压缩了图匹配空间,大大提高了特征识别的速度,应用此方法实现了飞机壁板类零件的特征识别.      

基于潜在成分的时变系统损伤的概率神经网络分类

提出了一种新的时变系统健康监控的损伤分类方法。将函数级数时变自回归平滑时序模型应用于时变系统的振动信号,以估计TAR／TMA参数和革新方差。这些参数是时间的函数,将它们在以特定的基函数构成的某种函数子空间上展开得到相应的投影系数组。所估计得到的TAR／TMA参数和革新方差可进一步用来计算潜在成分（LCs）,将LCs用于健康评估比原来的参数更具信息。并将LCs联合并归化为数值得到特征集,输入概率神经网络（PNN）进行损伤分类。为了评价该方法,对一个时变系统进行了仿真,以各种不同的质量和刚度减少来模拟不同的损伤类别。算例表明：该方法能够在时变系统的背景下对损伤进行归类。     

适于黏性计算的自适应笛卡儿网格生成及其应用

复杂外形的黏性网格生成已经成为计算流体力学(CFD)在工程应用中的主要瓶颈,高效而鲁棒的网格生成技术研究显得尤为迫切。使用基于物面外形自适应加密的笛卡儿网格方法生成计算网格,投影方法拟合壁面边界,将投影得到的柱形单元沿法向分层获得适于黏性计算的网格。应用特征恢复技术恢复凹面特征并改善凸面处的网格质量。流场解算中应用最小二乘法重构获得2阶精度,Roe格式计算无黏通量,Holmes-Connell方法计算黏性通量。加入Spalart-Allmaras一方程湍流模型以获得湍流解。上下对称高斯 赛德尔迭代(LU-SGS)格式隐式时间推进,对网格重排序保证格式的稳定性和平衡性。算例显示网格生成方法是快速的和鲁棒的,流场计算精度满足工程需求,因此该方法非常适合应用于实际工程计算。     

Beamlet变换提取线特征的改进算法

文章首先利用二进方块中Beamlet基的对称、平行等几何性质对离散Beamlet基进行分类;然后通过定义一个新的二进方块统计量和建立一个新的能量函数,对Beamlet变换方法中利用BD-RDP提取线特征的算法进行改进.通过实验对比,该算法进一步减少了运算量,提高了线特征提取的速度,且对复杂结构图像中的线特征提取效果更为清晰.     

基于SVM的浮动车行驶模式判断模型

浮动车在低速情况下存在两种行驶模式,如不能对上述模式进行准确区分,将严重影响浮动车实时路况计算的精度和效率.研究和设计了一个基于支持向量机(SVM,Support Vector Machine)的浮动车行驶模式判断模型,并针对性地提出了一种简单的基于隶属度矩阵的特征评价和选择方法.实验表明通过上述方法选择的特征子集所训练的分类器在测试样本集上具有92.6%的分类准确性;经过行驶模式分析后,浮动车系统的准确性有显著提升.      

设计特征模型中的装配特征识别

研究了由设计特征模型识别装配特征的算法 ,该算法采用凹凸特征模型为中间模型实现设计特征模型向装配模型的转化。识别算法用特征设计与特征识别相结合的方法有效地识别出设计特征指定面的所有凹凸特征 ,利用判别装配特征的充分必要条件得到所求的装配特征。该方法克服了传统算法的缺点 ,提高了装配特征识别的有效性和可靠性     

基于表观细粒度辨别网络的近海船舶目标检测方法

近海船舶目标检测是一项非常具有挑战性的任务,受到学者专家广泛关注。基于卷积神经网络（CNN）和注意力机制的检测器在近海船舶目标检测方面的应用取得了显著成就。然而,船舶目标检测存在着表观相似和背景干扰导致检测过程中出现误检的问题。为此,本文提出了一种用于Faster RCNN（更快的基于区域的卷积神经网络）的表观细粒度辨别的检测头模块。该模块包括类别细粒度分支和高效全维动态卷积定位分支。其中类别细粒度分支通过全局特征建模和灵活的感知范围来挖掘和利用类别细粒度辨别特征,高效全维动态卷积定位分支通过高效灵活的感知船舶边界信息来区分目标与背景,从而减少误检漏检问题。通过在近海船舶公开数据集Seaships7000 上进行实验验证,本文算法减少了误检漏检,提升了检测器性能。     

A review of topology optimization for additive manufacturing: Status and challenges

Topology optimization was developed as an advanced structural design methodology to generate innovative lightweight and high-performance configurations that are difficult to obtain with conventional ideas. Additive manufacturing is an advanced manufacturing technique building as-designed structures via layer-by-layer joining material, providing an alternative pattern for complex components. The integration of topology optimization and additive manufacturing can make the most of their advantages and potentials, and has wide application prospects in modern manufacturing. This article reviews the main content and applications of the research on the integration of topology optimization and additive manufacturing in recent years, including multi-scale or hierarchical structural optimization design and topology optimization considering additive manufacturing constraints. Meanwhile, some challenges of structural design approaches for additive manufacturing are discussed, such as the performance characterization and scale effects of additively manufactured lattice structures, the anisotropy and fatigue performance of additively manufactured material, and additively manufactured functionally graded material issues, etc. It is shown that in the research of topology optimization for additive manufacturing, the integration of material, structure, process and performance is important to pursue high-performance, multi-functional and lightweight production. This article provides a reference for further related research and aerospace applications.     

基于自组织神经网络的软件功能测试数据自动生成

针对面向软件功能的测试数据自动生成问题,提出了一种动态自组织特征映射方法,用于生成揭示软件功能故障的测试数据(简称故障数据)。该方法主要有两部分组成,①采用具有全局多峰搜索特性的小生境遗传算法,在输入空间内搜索功能测试数据,生成少量的初始故障数据;②由初始故障数据,采用具有联想和分类能力的可变结构自组织特征映射,不断迭代生成大量相近而不同的故障数据,以便给开发者提供引发这些软件故障的信息,从而确定软件故障行为的模式或假设。用某型空空导弹发射控制软件进行了实验,运行结果表明了方法的有效性,故障数据生成效率高于遗传算法和随机法。     

Online condition diagnosis for a two-stage gearbox machinery of an aerospace utilization system using an ensemble multi-fault features indexing approach

China manned space station is designed to operate for over ten years. Long-term and sustainable research on space science and technology will be conducted during its operation. The application payloads must meet the ‘‘long life and high reliability" mission requirement. Gearbox machinery is one of the essential devices in an aerospace utilization system, failure of which may lead to downtime loss even during some disastrous catastrophes. A fault diagnosis of gearbox has attracted attentions for its significance in preventing catastrophic accidents and guaranteeing sufficient maintenance. A novel fault diagnosis method based on the Ensemble Multi-Fault Features Indexing(EMFFI) approach is proposed for the condition monitoring of gearboxes. Different from traditional methods of signal analysis in the one-dimensional space, this study employs a supervised learning method to determine the faults of a gearbox in a two-dimensional space using the classification model established by training the features extracted automatically from diagnostic vibration signals captured. The proposed method mainly includes the following steps. First, the vibration signals are transformed into a bi-spectrum contour map utilizing bi-spectrum technology,which provides a basis for the following image-based feature extraction. Then, Speeded-Up Robustness Feature(SURF) is applied to automatically extract the image feature points of the bi-spectrum contour map using a multi-fault features indexing theory, and the feature dimension is reduced by Linear Discriminant Analysis(LDA). Finally, Random Forest(RF) is introduced to identify the fault types of the gearbox. The test results verify that the proposed method based on the multi-fault features indexing approach achieves the target of high diagnostic accuracy and can serve as a highly effective technique to discover faults in a gearbox machinery such as a two-stage one.