NUMERICAL SIMULATION OF METHANE-AIR TURBULENT JET FLAME USING A NEW SECOND-ORDER MOMENT MODEL

A new second-order moment model for turbulent combustion is applied in the simulation of methane-air turbulent jet flame. The predicted results are compared with the experimental results and with those predicted using the well-known EBU-Arrhenius model and the original second-order moment model. The comparison shows the advantage of the new model that it requires almost the same computational storage and time as that of the original second-order moment model, but its modeling results are in better agreement with experiments than those using other models. Hence, the new second-order moment model is promising in modeling turbulent combustion with NOx formation with finite reaction rate for engineering application.

Comparison of a Full Second-Order Moment Model and an Algebraic Stress Two-Phase Turbulence Model for Simulating Bubble-Liquid Flows in a Bubble Column

A full second-order moment (FSM) model and an algebraic stress (ASM) two-phase turbulence modelare proposed and applied to predict turbulent bubble-liquid flows in a 2D rectangular bubble column. Predictiongives the bubble and liquid velocities, bubble volume fraction, bubble and liquid Reynolds stresses and bubble-liquidvelocity correlation. For predicted two-phase velocities and bubble volume fraction there is only slight differencebetween these two models, and the simulation results using both two models are in good agreement with the particleimage velocimetry (PIV) measurements. Although the predicted two-phase Reynolds stresses using the FSM are insomewhat better agreement with the PIV measurements than those predicted using the ASM, the Reynolds stressespredicted using both two models are in general agreement with the experiments. Therefore, it is suggested to usethe ASM two-phase turbulence model in engineering application for saving the computation time.

A two-scale second-order moment two-phase turbulence model for simulating dense gas-particle flows

A two-scale second-order moment two-phase turbulence model accounting for inter-particle collision is developed, based on the concepts of particle large-scale fluctuation due to turbulence and particle small-scale fluctuation due to collision and through a unified treatment of these two kinds of fluctuations. The proposed model is used to simulate gas-particle flows in a channel and in a downer. Simulation results are in agreement with the experimental results reported in references and are near the results obtained using the sin- gle-scale second-order moment two-phase turbulence model superposed with a particle collision model （USM-θ model） in most regions.

EFFECT OF EMPIRICAL COEFFICIENTS ON SIMULATION IN TWO-SCALE SECOND-ORDER MOMENT PARTICLE-PHASE TURBULENCE MODEL

A two-scale second-order moment two-phase turbulence model accounting for inter-particle collision is developed, based on the concept of particle large-scale fluctuation due to turbulence and particle small-scale fluctuation due to collision. The proposed model is used to simulate gas-particle downer reactor flows. The computational results of both particle volume fraction and mean velocity are in agreement with the experimental results. After analyzing effects of empirical coefficient on prediction results, we can come to a conclusion that, inside the limit range of empirical coefficient, the predictions do not reveal a large sensitivity to the empirical coefficient in the downer reactor, but a relatively great change of the constants has important effect on the prediction.

SECOND-ORDER MOMENT MODEL FOR DENSE TWO-PHASE TURBULENT FLOW OF BINGHAM FLUID WITH PARTICLES

The USM-θ model of Bingham fluid for dense two-phase turbulent flow was developed, which combines the second-order moment model for two-phase turbulence with the particle kinetic theory for the inter-particle collision. In this model, phases interaction and the extra term of Bingham fluid yield stress are taken into account. An algorithm for USM-θ model in dense two-phase flow was proposed, in which the influence of particle volume fraction is accounted for. This model was used to simulate turbulent flow of Bingham fluid single-phase and dense liquid-particle two-phase in pipe. It is shown USM-θ model has better prediction result than the five-equation model, in which the particle-particle collision is modeled by the particle kinetic theory, while the turbulence of both phase is simulated by the two-equation turbulence model. The USM-θ model was then used to simulate the dense two-phase turbulent up flow of Bingham fluid with particles. With the increasing of the yield stress, the velocities of Bingham and particle decrease near the pipe centre. Comparing the two-phase flow of Bingham-particle with that of liquid-particle, it is found the source term of yield stress has significant effect on flow.

Fast Non-Local Means Algorithm Based on Krawtchouk Moments

Non-local means(NLM)method is a state-of-the-art denoising algorithm, which replaces each pixel with a weighted average of all the pixels in the image. However, the huge computational complexity makes it impractical for real applications. Thus, a fast non-local means algorithm based on Krawtchouk moments is proposed to improve the denoising performance and reduce the computing time. Krawtchouk moments of each image patch are calculated and used in the subsequent similarity measure in order to perform a weighted averaging. Instead of computing the Euclidean distance of two image patches, the similarity measure is obtained by low-order Krawtchouk moments, which can reduce a lot of computational complexity. Since Krawtchouk moments can extract local features and have a good antinoise ability, they can classify the useful information out of noise and provide an accurate similarity measure. Detailed experiments demonstrate that the proposed method outperforms the original NLM method and other moment-based methods according to a comprehensive consideration on subjective visual quality, method noise, peak signal to noise ratio(PSNR), structural similarity(SSIM) index and computing time. Most importantly, the proposed method is around 35 times faster than the original NLM method.

Moments and Mean First—Passage Time of Parabolic—Bistable Potential System Driven by Colored Noise

A parabolic-bistable potential system driven by colored noise is studied. The exact analytical expressions of the stationary probability distribution (SPD) and the moments of the system are derived. Furthermore, the mean first-passage time is calculated by the use of two approximate methods, respectively. It is found that (i) the double peaks of SPD are rubbed-down into a flat single peak with the increasing of noise intensity; (ii) a minimum occurs on the curve of the second-order moment of the system vs. noise intensity at the point ; (iii) the results obtained by our approximate approach are in good agreement with the numerical calculations for either small or large correlation time , while the conventional steepest descent approximation leads to poor results.

MEAN-SQUARE RADIUS OF GYRATION AND DIPOLE MOMENT OF POLY(METHYLPHENYLSILOXANE) CHAINS

The mean-square radius of gyration <S^2>,the mean-square dipole moment <D^2>,the mean-square end-to-end distance <R^2> and their temperature coefficients of unsymmetrical disubstituted poly(methylphenylsiloxane) (PMPS) chains, as a function of stereochemical structure,confomational energies and length of polymers,were studied by using an improved configurational-confomational statistical method based on the rotational-isomeric-state theory.It is found that the increase in isotacticity of P...

基于耿贝尔-线性矩和K-Means算法的短时强降水等级划分研究

短时强降水地域特点突出,致灾性强,是气象防灾减灾的重点和难点.本文基于1951—2021年国家气象观测站逐小时降水资料,采用耿贝尔—线性矩法拟合2215个国家气象观测站的1 h、3 h、6 h短时强降水的重现期分布,统计显示99%以上的国家气象观测站分布函数通过0.05显著性检验;同时,根据我国防洪标准、室外排水设计规范及中央气象台山洪预警等标准,将短时强降水按照重现期2 a、20 a、50 a、100 a划分为一般、较强、强、特强四个等级.然后,采用K-Means算法,应用耿贝尔函数的位置参数和尺度参数、短时降水的百分位值和极值等变量聚类,将我国划分为三个短时强降水区,量化了三个区的短时强降水等级并应用2019—2021年70490个区域自动观测站和国家气象观测站监测统计验证了合理性.研究成果可以满足气象业务应用需求,也为气象监测、预报预警和影响评估提供技术支撑.

基于Mean Shift算法和NMI特征的目标跟踪算法研究

针对传统Mean shift跟踪算法对空中运动目标跟踪效果不理想的问题,提出了基于Mean shift算法和归一化转动惯量(Normalized moment of inertia,NMI)特征的目标跟踪算法.算法中引入了目标NMI特征,建立了基于虚警概率最小原则和相似度二级判决门限的跟踪策略,对目标模型进行更新.同时利用卡尔曼滤波,在目标被遮挡后进行估计预测.实验表明该算法在空中运动目标存在较大形变、被遮挡等情况下,能够进行实时、稳定跟踪.

带宽自适应的Mean Shift目标跟踪算法

针对传统均值漂移(Mean Shift)目标跟踪算法中核函数带宽缺乏良好自适应调整的缺点,提出了自适应调整核函数带宽的Mean Shift目标跟踪算法.该算法首先采用核函数计算目标颜色特征值的概率密度,在视频当前帧目标的最优位置区域由目标颜色特征概率投影生成目标概率密度分布图;然后根据概率密度零阶矩值调整下一帧跟踪窗口宽度,从而实现核函数带宽的自适应调整;最后通过矩运算计算椭圆参数,用椭圆锁定目标来实现复杂背景下目标的空间、尺度和方向定位.人脸跟踪实验结果表明:与传统MeanShift目标跟踪算法相比,文中算法可以实时地对目标进行缩放锁定,且能够估计目标姿态;与Cam Shift算法相比,文中算法抵抗相似颜色干扰的性能较好.

模板自适应的Mean Shift红外目标跟踪

为了解决Mean Shift跟踪算法中目标模板只能从单一图像建立且很难更新问题,提出了一种结合改进的Mean Shift与增量式支持向量机的红外目标跟踪算法。首先,根据目标区域的灰度直方图对目标进行描述,然后采用标准Mean Shift搜索目标,结合子图图像矩特征进行二次搜索,再计算下一帧搜索的窗口大小,以解决目标尺寸明显变化时造成目标丢失的问题。同时,针对目标遮挡易导致跟踪失败的问题,引入机器学习理论,采用增量式支持向量机自适应更新模板,则目标跟踪问题转换为目标和背景的分类问题。实验结果表明:提出的改进算法在目标尺寸、姿态发生变化或出现部分遮挡时,能有效跟踪目标。

一种尺度和方向适应性的Mean Shift跟踪算法

针对原始的Mean Shift跟踪算法虽能准确地估计目标位置,但对目标尺度和方向不能实现自适应估计,结合目标模型与候选目标区域的候选模型得到了反向投影图,此反向投影图可表示图像中像素点属于目标的概率,将反向投影图的矩特征应用到原始Mean Shift跟踪算法框架,实现了目标尺度和方向适应性Mean Shift跟踪.实验结果表明:该算法能有效跟踪尺度和方向变化的目标.

基于改进K-Means算法的车辆识别方法

车辆的检测和识别一直是道路监控、安全辅助驾驶、车辆自主导航等领域的重要研究内容。文中基于机器视觉方法,在Lab颜色空间通过对经典K-Means算法的聚类中心、聚类数和距离测度三方面的改进实现了L分量的聚类,从而达到图像分割的目的。提取图像的矩形度、Hu矩和Affine矩的特征,针对不同的车型建立各自的模板,利用改进Hu不变矩和仿射不变矩的组合不变矩对分割后图像进行车型的识别。实验结果表明,文中提出的方法对于复杂环境下的车辆检测和识别具有良好的可靠性和鲁棒性。

基于余弦定理和K-means的植物叶片识别方法

为了提高植物叶片识别准确率,提出一种基于余弦定理和K-means的识别方法.该方法首先通过提取叶片的Hu不变矩和形状特征得到叶片的综合特征向量,然后使用K均值聚类(Kmeans)对各类叶片训练样本的特征向量集合进行聚类以获得聚类中心特征向量,紧接着使用余弦定理计算目标叶片和训练样本的相似度并排序.仿真实验表明:在Flavia植物叶片数据库中进行测试,该文方法以96.03%的概率在前5位发现目标,优于KNN、BP神经网络方法,因此,该方法具有一定的实用价值.

采用K-means的脑肿瘤磁共振图像分割与特征提取

大脑肿瘤分割对于医师判断肿瘤恶化程度非常重要。然而,由于肿瘤的不规则形状、与周围组织的低对比度以及出现位置的不固定,给脑肿瘤的精确分割带来很大的困难。传统的K-means分割方法仅仅利用图像的灰度特征,很难准确分割肿瘤边界。利用灰度共生矩阵提取出的纹理特征,并结合图像几何不变矩特征对分割出的脑肿瘤图像进行特征提取。灰度共生矩阵定义为像素对的联合概率分布,是一个对称矩阵,它不仅反映图像灰度在相邻的方向、相邻间隔、变化幅度的综合信息,也反映了相同的灰度级像素之间的位置分布特征,是计算纹理特征的基础;几何矩(不变矩)具有旋转、平移、尺度等特性,能将图像分解为有限特征值,并且通过对比所提取出的同一病人的肿瘤图像的不变矩参数,可以获得该肿瘤几何形状变化程度。实验结果表明,该方法可以同时从纹理和几何特征对图像特征进行描述,与分别采用灰度共生矩阵和不变矩方法进行特征提取相比较,降低了算法计算量,同时提升了算法的抗噪性。

随机年龄结构固定资产系统倒向Euler法的p阶矩耗散性

在单边Lipschitz条件下,研究了一类随机年龄结构固定资产系统倒向Euler法数值解的p阶矩耗散性.当0<p<1时,步长满足一定条件可以得到系统的p阶矩耗散性;而p=2时,在对步长没有任何限制条件的情况下,得到了系统的均方耗散性.最后,通过数值例子验证了理论结果的可行性和有效性.

基于惯量矩的自适应带宽Mean Shift目标跟踪算法

基于核的传统均值漂移目标跟踪算法,对于目标的尺度和角度缺乏良好的自适应能力。为此,提出一种基于惯量矩的自适应调整核函数带宽均值漂移跟踪算法。该算法首先用颜色模型在最优迭代位置投影创建目标概率密度分布,然后计算此密度分布的形心主轴惯量矩和旋转角度,最后用惯量矩的方法对目标进行椭圆拟合。得到目标的长度和宽度,递归滤波后自适应调整核函数带宽。实验结果表明,该算法在光照、尺度变化情况下可以准确跟踪目标,并估计目标旋转角度。

自适应跟踪窗的Mean Shift算法

针对经典Mean Shift算法在目标跟踪过程中不能自适应目标的形状变化,提出了一种跟踪窗自适应的Mean Shift目标跟踪算法。该算法利用权值图像的矩特征推导出了一种目标真实面积的估计方法,同时利用估计出的目标真实面积校正可表达目标窗口内图像特征的协方差矩阵,从而计算出目标区域的主轴长度和方向。实验结果表明,该方法保留了Mean Shift算法简单、有效的特点,同时该算法的跟踪窗具备当目标发生旋转和尺度变化时的自适应能力,并且具有较强的鲁棒性。

基于不变矩和Mean Shift聚类的视频关键帧提取

随着通信和多媒体技术的迅速发展,检索和浏览海量多媒体数据成为日益迫切的问题。关键帧提取技术在基于内容的视频检索中扮演了重要角色。提取的关键帧有两个主要作用:一是用它来静态表示视频的主题;其二是希望从关键帧中提取特征数据,作为多媒体数据库的数据源。以视频场景中运动目标的不变矩为特征提出了一种基于非监督聚类的关键帧提取算法。首先在视频序列中分离出运动目标,然后计算运动目标区域的各阶不变矩,并以不变矩向量作为特征,运用Mean Shift算法聚类有相似特征的视频帧,进而在每类中选取有代表性的视频帧作为关键帧。不同场景下的实验结果证实了本算法的可行性。