Hydrodynamics of gas fluidized beds with mixture of group D and B particles

The dynamics of a gas‐solid fluidized bed containing Geldart Group D particles mixed with a small proportion of Geldart Group B particles are investigated using pressure fluctuations data. Time series analysis, using a variety of nonlinear dynamics tools, shows that the slugging present with Group D particles can be suppressed by the addition of a small proportion of Group B particles. The power spectra and the auto‐correlation function are used for a preliminary evaluation of dominant slug frequencies. It is shown that the bed fluidized with the mixture of Group D and B particles behaves in a less periodic manner and is dominated by more random bubble motion. On the other hand, the correlation integral is used to analyze the chaotic behaviour of the flow, through evaluation of the fractal structure of the reconstructed attractors. It is shown that the bed fluidized with Group D particles is characterized by a combination of steady slug motion and irregular particles motion. The behaviour of the fluidized bed of Group D/B particles mixture is characterized by a single correlation dimension for a wide range of fluidization velocities.     

Corrosion protection of commercial steel using stainless steel coatings deposited by Cathodic Arc Plasma Deposition technique

The commercial steel AISI 1010 was coated with AISI 316L steel using Cathodic Arc Plasma Deposition (CAPD) technique. The coatings were deposited in vacuum and in the presence of nitrogen, acetylene and mixture of the two as reactive gases. The coatings were deposited as a function of time while other parameters remained constant. The coatings 0.75 to 1.3 μm thick were adherent and amorphous. The aqueous corrosion properties of the coated samples in 3.5 wt % NaCl solution were studied by Tafel, cyclic and potentiodynamic polarization measurements. The derived corrosion parameters were then compared among the various uncoated and coated samples. The study revealed that the coated samples were more corrosion resistant than the uncoated one. Similarly, the samples coated in the nitrogen + acetylene mixture atmosphere were more corrosion resistant than the samples coated in only nitrogen and acetylene atmospheres. The corrosion parameters were also compared as a function of coating time to ascertain best coating thickness.     

Effect of seasonal variations and lactation times on aflatoxin M1 contamination in milk of different species from Punjab, Pakistan

During October 2009 to September 2010, aflatoxin M₁ (AFM₁) levels were analyzed by HPLC-FLD in 356 milk samples of different lactating species (buffalo, cow, goat, sheep and camel) from Punjab (Pakistan). Recoveries of AFM₁ ranged from 92 to 97% and the limit of detection (LOD) was 0.004 μg/L. For all lactating species the mean concentration of AFM₁ was significantly higher in winter season than in summer (p < 0.05). The results showed that 55, 56, 32, 58 and 27% of winter milk samples of buffalo, cow, goat, sheep and camel exceeded the EU maximum limit (0.05 μg/kg), compared with 38, 33, 21, 36 and 14% of summer milk samples, respectively. For all lactating species the mean concentration of AFM₁ was significantly higher in morning milks than in evening milks (p < 0.05). The percentage of morning milk samples exceeding the EU maximum limit was 72, 67, 69, 71 and 44% for buffalo, cow, goat, sheep and camel, while for evening milks percent non compliant rates were 39, 30, 18, 33 and 25%, respectively. The level of AFM₁ tended to be higher in animal species fed mainly on concentrate mixtures (buffalo and cow) than in other species grazing on fresh greens.     

Siamese multi-layer perceptrons for dimensionality reduction and face identification

This paper presents a framework using siamese Multi-layer Perceptrons (MLP) for supervised dimensionality reduction and face identification. Compared with the classical MLP that trains on fully labeled data, the siamese MLP learns on side information only, i.e., how similar of data examples are to each other. In this study, we compare it with the classical MLP on the problem of face identification. Experimental results on the Extended Yale B database demonstrate that the siamese MLP training with side information achieves comparable classification performance with the classical MLP training on fully labeled data. Besides, while the classical MLP fixes the dimension of the output space, the siamese MLP allows flexible output dimension, hence we also apply the siamese MLP for visualization of the dimensionality reduction to the 2-d and 3-d spaces.     

Removal and recovery of nickel(II) from aqueous solution by loofa sponge-immobilized biomass of Chlorella sorokiniana: characterization studies

The biosorption process for the removal of nickel(II) by loofa sponge-immobilized biomass of Chlorella sorokiniana (LIBCS), a newly developed immobilized biosorbent, was characterized. Effects of environmental factors on metal uptake capacity of LIBCS were studied and compared with free biomass of C. sorokiniana (FBCS). Nickel(II) removal by LIBCS was found to be influenced by pH of the solution, initial metal concentration, and biomass concentration. The biosorption of nickel(II) ions by both LIBCS and FBCS increased as the initial concentration of nickel(II) ions increased in the medium. No loss to biosorption capacity of LIBCS for nickel(II) was found due to the presence of loofa sponge, indeed as compared to FBCS an increase of 25.3% was noted in the biosorption capacity of LIBCS. Maximum biosorption capacities for FBCS and LIBCS were found as 48.08 and 60.38 mg nickel(II)/g, respectively, whereas the amount of nickel(II) ions adsorbed on the plain loofa sponge was 6.1mg/g. During these biosorption studies, LIBCS exhibited excellent physical and chemical stability without any significant release/loss of microalgal biomass from loofa sponge matrix. The kinetics of nickel(II) removal was extremely fast reaching at equilibrium in about 15 min for LIBCS and 20 min for FBCS. The biosorption equilibrium was well described by the Langmuir and Freundlich adsorption isotherms. The biosorption capacities were found to be solution pH dependent and the maximum adsorption was found at a solution pH 4-5. The LIBCS could be regenerated using 75 mM HCl, with up to 98% recovery. The LIBCS were shown to be robust and stable with little decrease in the nickel(II) uptake capacity when used in consecutive seven biosorption-desorption cycles. Continuous removal of nickel(II) from electroplating effluent by LIBCS packed in fixed bed column bioreactor confirm the possibility of developing a biological treatment process for the removal of toxic metals from authentic wastewater.     

Automatic Quality Inspection of Bakery Products Based on Shape and Color Information

基于形状和颜色信息的烘焙产品的自动质量检测

Babar Khan,韩芳*,王直杰,Ather Iqbal

（东华大学 信息科学与技术学院,上海 201620）

创新点说明：

利用支持向量机和基于生物启发式HMAX模型的形状辨识器与RGB对抗颜色通道辨识器集成而形成一种有效的智能分类系统,并将其用于烘焙产品的质量检测。

研究目的：

本文要对烘焙产品进行更有效的质量检测,需要提出一种智能方法。

研究方法：

本文采用了支持向量机、基于生物启发式HMAX模型的形状辨识器与RGB对抗颜色通道辨识器集成,形成一种高级自动分类系统,对烘焙产品根据其形状和颜色（烘焙时间）进行对应分类。

结果：

利用本文方法对烘焙产品进行质量检测的结果与其它方法相比较,发现本文提出的基于生物启发的机器视觉系统运行更加精确和有效,同时发现采用One Versus One支持向量机和Directed Acyclic Graph支持向量机可以得到最大的分类精确率。本文方法对饼干形状和颜色的辨识分别达到了95%和100%的分类精确率。同时,算法稳定。

结论：

本文方法因其更加可靠和精确可用于代替现有的烘焙产品检测方法。

关键词：烘焙产品质量检测,计算机视觉,HMAX, 对抗颜色通道,RGB颜色辨识器,支持向量机

     

Brain tumor classification from multi-modality MRI using wavelets and machine learning

In this paper, we propose a brain tumor segmentation and classification method for multi-modality magnetic resonance imaging scans. The data from multi-modal brain tumor segmentation challenge (MICCAI BraTS 2013) are utilized which are co-registered and skull-stripped, and the histogram matching is performed with a reference volume of high contrast. From the preprocessed images, the following features are then extracted: intensity, intensity differences, local neighborhood and wavelet texture. The integrated features are subsequently provided to the random forest classifier to predict five classes: background, necrosis, edema, enhancing tumor and non-enhancing tumor, and then these class labels are used to hierarchically compute three different regions (complete tumor, active tumor and enhancing tumor). We performed a leave-one-out cross-validation and achieved 88% Dice overlap for the complete tumor region, 75% for the core tumor region and 95% for enhancing tumor region, which is higher than the Dice overlap reported from MICCAI BraTS challenge.