SELF-SIMILAR CRACK EXPANSION METHOD FOR THE STRESS INTENSITY FACTOR ANALYSIS

The Self-Similar Crack Expansion (SSCE) method is proposed to evaluate stress intensity factors at crack tips, whereby stress intensity factors of a crack can be determined by the crack opening displacement over the crack, not just by the local displacement around the crack tip. The crack expansion rate is estimated by taking advantage of the crack self-similarity. Therefore, the accuracy of the calculation is improved. The singular integrals on crack tip elements are also analyzed and are precisely evaluated in terms of a special integral analysis. Combination of these two techniques greatly increases the accuracy in estimating the stress distribution around the crack tip. A variety of two-dimensional cracks, such as subsurface cracks, edge cracks, and their interactions are calculated in terms of the self-similar expansion rate. Solutions are satisfied with errors less than 0.5% as compared with the analytical solutions. Based on the calculations of the crack interactions, a theory for crack interactions is proposed such that for a group of aligned cracks the summation of the square of SIFs at the right tips of cracks is always equal to that at the left tips of cracks. This theory was proved by the mehtod of Self-Similar Crack Expansion in this paper.

A SELF-SIMILAR CRACK EXPANSION METHOD FOR THREE-DIMENSIONAL CRACKS IN AN INFINITE OR SEMI-INFINITE MEDIUM

The Self-Similar Crack Expansion (SSCE) method is used to calculate stress intensity factors for three-dimensional cracks in an infinite medium or semi-infinite medium by the boundary integral element technique, whereby, the stress intensity factors at crack tips are determined by calculating the crack-opening displacements over the crack surface. For elements on the crack surface, regular integrals and singular integrals are precisely evaluated based on closed form expressions, which improves the accuracy. Examples shaw that this method yields very accurate results for stress intensity factors of penny-shaped cracks and elliptical cracks in the full space, with errors of less than 1% as compared with analytical solutions. The stress intensity factors of subsurface cracks ate in good agreement with other analytical solutions.

Study of Delay and Loss Behavior of Internet Switch-Markovian Modelling Using Circulant Markov Modulated Poisson Process (CMMPP)

Most of the classical self-similar traffic models are asymptotic in nature. Therefore, it is crucial for an appropriate buffer design of a switch and queuing based performance evaluation. In this paper, we investigate delay and loss behavior of the switch under self-similar fixed length packet traffic by modeling it as CMMPP/D/1 and CMMPP/D/1/K, respectively, where Circulant Markov Modulated Poisson Process (CMMPP) is fitted by equating the variance of CMMPP and that of self-similar traffic. CMMPP model is already the validated one to emulate the self-similar characteristics. We compare the analytical results with the simulation ones.

基于多特征融合的红外图像行人检测

针对基于单特征红外图像行人识别准确率低的问题,提出一种基于梯度方向直方图(HOG)、积分通道特征(ICF)和强度自适应特征(ISS)的多特征融合红外图像行人检测的新方法。首先,分别提取训练样本的HOG、积分通道和ISS特征,用主成分分析(PCA)算法对提取的ISS特征进行降维,然后通过并行加权特征融合方法把HOG、积分通道和降维后的ISS特征相融合,并用融合后的特征训练支持向量机(SVM)分类器,最后用训练好的SVM分类器进行行人识别检测。LSI Far Infrared Pedestrian Dataset红外行人图像数据库上的实验证明,基于多特征的红外图像行人检测方法明显优于经典的HOG和局部二值模式(LBP)单特征方法,提高了检测精度,降低了误检率。