一种新型叶片涡轮分级机流场的数值模拟

建立了一种带后弯导板的新型叶片形状涡轮分级机内腔叶片之间气-固两相湍流的双流体理论模型,并利用通用的流体力学计算软件进行了数值模拟。结果表明:该涡轮分级机在正常工作状态时,内腔流场遵循强湍流流动规律;当采用带后弯导板的叶片形状时,混合气流通过两叶片间流场会产生一种正方向的涡流,它可以抵消使用直叶片时由于涡轮高速旋转而不能克服的附加反方向涡旋,从而使流场更加稳定,容易形成整流,提高涡轮分级机的分级效率和分级精度。     

Preparation and evaluation of high drug content particles

To determine how to prepare high drug content particles using a Wurster fluidized bed to determine realizing the miniaturization of solid dosage forms, aspirin was selected as the model drug and granulated without any additive. In this study, the emphasis was on evaluating the key operation factors of airflow rate and atomizing flow volume. The properties of the resulting particles, such as the average diameter, particle strength, appearance, and compressibility using different airflow rates and atomizing flow volumes, were investigated. Furthermore, detailed optimization of the operation conditions was conducted by artificial neural network (ANN) analysis. The relationship between the controlling factors (powder supplied, concentration of spray liquid, the amount of consumed spray liquid, and spray rate) and the response variables (product yield, median diameter, angle of repose, and degradation of aspirin) was investigated after evaluating the airflow rate and atomizing flow volume effects. The resulting granules under optimum operation conditions showed excellent physicochemical properties such as particle size uniformity, flowability, and compressibility.     

River segmentation using satellite image contextual information and Bayesian classifier

Satellite-based remote sensing imaging can provide continuous snapshots of the Earth’s surface over long periods. River extraction from remote sensing images is useful for the comprehensive study of dynamic changes of rivers over large areas. This paper presents a new method of extracting rivers by using training samples based on the mathematical morphology, Bayesian classifier and a dynamic alteration filter. The use of a training map from erosion morphology helps to extract the non-predictive river’s curves in the image. The algorithm has two phases: creating the profile to separate river area via evaluated morphological erosion and dilation, namely, a training map; and improving the river’s image segmentation using the Bayesian rule algorithm in which two consecutive filters swipe false positive (non-water area) along the image. The proposed algorithm was tested on the Kuala Terengganu district, Malaysia, an area that includes a river, a bridge, dam and a fair amount of vegetation. The results were compared with two standard methods based on visual perception and on peak signal-to-noise ratio, respectively. The novelty of this approach is the definition of the contextual information filtering technique, which provides an accurate extraction of river segmentation from satellite images.     

CFD simulation and optimization of the flow field in horizontal turbo air classifiers

This paper mainly presents a numerical study of the gas flow in horizontal turbo air classifiers. The effect of the air-inlet direction on the performance of classifier was also investigated through powder classification experiments. The simulated results show that the vertical vortex is the dominant flow in conical part of the classifier and there exists the horizontal vortex in the classifying chamber. The tangential velocity profile resembles a Rankine vortex inside the rotor cage. The vertical vortex intensity increases with increasing the inlet air velocity, while the rotor cage speed has limited effect on the control of gas pathlines in the classifying chamber. Horizontal turbo air classifiers are divided into four quadrants according to the air-inlet direction. For classifiers in quadrants I and III, a double-layer flow with opposite directions generates around the rotor cage which causes a secondary vortex. The secondary vortex is eliminated and the airflow becomes uniform in the classifier that belongs to quadrant Π or IV. The experimental results with fluidized catalytic cracking catalysts and fly ash demonstrate that cut sizes of this classifier decrease averagely by 5 μm and 2.2 μm respectively, and the classification accuracy increases by 7.5–10.3%.     

电力变压器故障诊断专家系统研究现状及开发前景

在搜集了国内外大量有关文献的基础上,对用模糊理论、人工神经网络方法开发变压器的故障诊断专家系统的原理、现状作了总结,提出了改进方向。     

基于小波包特征的纹理影像分割

利用小波包变换来检测纹理图象上的灰度变化，以得到多分辨率、多方向性的微观统计特征图象，在此基础上用Ｅｎｖｅｌｏｐ算法提取基于边缘信息的纹理特征，为了克服纹理特征之间的相关性，采用了子空间分类器对纹理特征图象进行分类，取得了较好的分割效果。     

环境空气SO2和NO2浓度的GA_ANN预测模型研究

空气中污染物浓度的预测是一个复杂的非线性问题。国内外的研究表明神经网络能够比回归模型更好地预报空气污染物。设计并实现了将用于选择最优预报因子的遗传算法和神经网络算法相结合的GA_ANN空气质量预测模型,利用某市2003~2006年的数据建立神经网络空气质量预测模型,对该市2007年全年SO₂和NO₂的预测实验表明,GA_ANN模型比单纯的神经网络模型具有更高的预报精度。     

结合灰色扩充的GA-BP神经网络模型在渭河水质遥感反演中的应用

以渭河陕西段水域为研究对象,用遗传算法改进的BP神经网络,结合灰色理论,建立了一种结合灰色扩充的GA-BP神经网络模型,对渭河水质中的主要污染指标CODmn(高锰酸盐指数)、COD(化学需氧量)、NH₃-N(氨氮)、DO(溶解氧)进行了遥感反演建模。实验证明：改进后的人工神经网络模型在预测精度上高于普通的BP神经网络模型和传统的多元线性回归模型,可用于渭河水质遥感反演建模。
     

基于两级分类器的人脸检测系统设计

人脸识别技术拥有广泛的应用前景,但是目前不少实现方式存在一些不尽人意之处。在对OPENCV与SVM分类器进行分析的基础上,阐述了基于两级分类器的人脸检测方法的原理和实现过程,首先分析了两级分类器的构建,引入人脸图像的矩形特征向量,将图像的矩形特征作为分类的依据,随后论述了系统设计与实现,包括灰度变换过程、直方图均衡过程、图像平滑过程以及金字塔序列化的实现。这种检测模式能够加快处理速度,提升效率。     

Particle swarm optimization for ensembling generation for evidential k-nearest-neighbour classifier

The problem addressed in this paper concerns the ensembling generation for evidential k-nearest-neighbour classifier. An efficient method based on particle swarm optimization (PSO) is here proposed. We improve the performance of the evidential k-nearest-neighbour (EkNN) classifier using a random subspace based ensembling method. Given a set of random subspace EkNN classifier, a PSO is used for obtaining the best parameters of the set of evidential k-nearest-neighbour classifiers, finally these classifiers are combined by the “vote rule”. The performance improvement with respect to the state-of-the-art approaches is validated through experiments with several benchmark datasets.