The production of volatile compounds in model casein systems with varying fat levels as affected by low-frequency ultrasound

The effects of low-frequency ultrasound on the production of volatile compounds in model casein protein systems containing various fat concentrations of 2%, 4% and 6% (w/w) were investigated. Ultrasound application was performed at 20 kHz for up to 10 min which corresponded to energy densities ranging from 9.54 to 190.8 J mL⁻¹. Similar volatile compounds were detected both in pure fat and mixtures of casein and fat (CF) systems. These volatiles belonged to the groups of aldehydes, ketones, esters, alcohols and hydrocarbons, which were the products of oxidation of lipids or protein degradation due to acoustic cavitation. The amount of fat in the casein systems had minor effects on the production of volatiles, whereas the production of volatile compounds was significantly affected by the ultrasound treatment. Short sonication times <5 min generated similar volatile profiles to the untreated samples. In contrast, prolonged sonication for 5 and 10 min considerably increased the production of volatile compounds and the amounts of fatty acids. Thus, the application of low–frequency ultrasound for short periods should be considered to minimise the production of volatile compounds which can ultimately affect the taste.     

Intelligent control of the feeding of aluminum electrolytic cells using neural networks

To be efficient, the control of alumina feeding of the electrolytic cell must be based on cell resistance, alumina concentration, and cell state. Most control schemes now in use are based on cell resistance only, and, thus, constitute an open-loop control that lacks robustness because their decision criteria are not explicitly tied to concentration nor to cell state. This results in the cell operating at nonoptimal concentrations, and cell efficiency is diminished. An optimal operation requires a knowledge of concentration and an adjustment of the decision criteria as a function of concentration. A learning vector quantization (LVQ) type of neural network was built and trained to recognize the cell state. Knowing the state of the cell and its resistance, concentration can be estimated using predetermined regression functions. The decision criteria for the control logic are then consequently adapted. A closed-loop control scheme is thus obtained. Results show that, with its control so structured, the cell can operate at or near optimal concentrations independently of its state. This flexible and intelligent character of the neural control can provide a considerable advantage as compared to the standard control.     

Spatial prediction of landslide susceptibility using data mining-based kernel logistic regression,naive Bayes and RBFNetwork models for the Long County area (China)

The main goal of this study is to assess and compare three advanced machine learning techniques, namely, kernel logistic regression (KLR), naïve Bayes (NB), and radial basis function network (RBFNetwork) models for landslide susceptibility modeling in Long County, China. First, a total of 171 landslide locations were identified within the study area using historical reports, aerial photographs, and extensive field surveys. All the landslides were randomly separated into two parts with a ratio of 70/30 for training and validation purposes. Second, 12 landslide conditioning factors were prepared for landslide susceptibility modeling, including slope aspect, slope angle, plan curvature, profile curvature, elevation, distance to faults, distance to rivers, distance to roads, lithology, NDVI (normalized difference vegetation index), land use, and rainfall. Third, the correlations between the conditioning factors and the occurrence of landslides were analyzed using normalized frequency ratios. A multicollinearity analysis of the landslide conditioning factors was carried out using tolerances and variance inflation factor (VIF) methods. Feature selection was performed using the chi-squared statistic with a 10-fold cross-validation technique to assess the predictive capabilities of the landslide conditioning factors. Then, the landslide conditioning factors with null predictive ability were excluded in order to optimize the landslide models. Finally, the trained KLR, NB, and RBFNetwork models were used to construct landslide susceptibility maps. The receiver operating characteristics (ROC) curve, the area under the curve (AUC), and several statistical measures, such as accuracy (ACC), F-measure, mean absolute error (MAE), and root mean squared error (RMSE), were used for the assessment, validation, and comparison of the resulting models in order to choose the best model in this study. The validation results show that all three models exhibit reasonably good performance, and the KLR model exhibits the most stable and best performance. The KLR model, which has a success rate of 0.847 and a prediction rate of 0.749, is a promising technique for landslide susceptibility mapping. Given the outcomes of the study, all three models could be used efficiently for landslide susceptibility analysis.

     

Evaluating a color-based active basis model for object recognition

Wu and coworkers introduced an active basis model (ABM) for object recognition in 2010, in which the learning algorithm tends to sketch edges in textures. A grey-value local power spectrum was used to find a common template and deformable templates from a set of training images and to detect an object in new images by template matching. In this paper, we propose a color-based active basis model (color-based ABM for short), which incorporates color information. We adopt the framework of Wu et al. in the learning, detection, and classification of the color-based ABM. However, in order to improve the performance in object recognition, we modify the framework of Wu et al. by using different color-based features in both the learning and template matching algorithms. In this color-based ABM approach, two types of learning (i.e., supervised learning and unsupervised learning) are also explored. Moreover, the usefulness of the color-based ABM for practical object recognition in computer vision applications is demonstrated and its significant improvement in recognizing objects is reported.     

EvaluatingWater Vapor Permeance Measurement Techniques for Highly Permeable Membranes

The cup method and dynamic moisture permeation cell (DMPC) method are two common techniques used to determine the water vapor permeation properties of a membrane. Often, ignoring the resistance of boundary air layers to the transport of water vapor results in the water vapor permeance of the membrane being underestimated in practical tests. The measurement errors are higher with highly permeable membranes. In this study, the two methods were simulated using COMSOL Multiphysics platform and the extent of the error was evaluated. Initial results showed that the error is equally high in both methods. With the correction for the still air gap, the cup method produces a relatively reduced error. In the DMPC method, reducing the error caused by the boundary air layer by increasing the sweep speed can produce higher instrument error. Highly accurate and precise instrument is needed for DMPC method; however, its error is still higher than that in the cup method. Simulations also show that lowering the test pressure is favorable to both methods.     

Optimal control of an aluminum casting furnace: Part I. The control model

This is a series of two articles on the control of an aluminum casting furnace to bring a mass of liquid aluminum from a known initial temperature to a desired final temperature in a given time with minimal fuel cost. An analytic model of the furnace already exists but is too complex for control purposes. Here in Part I, a simplified nonlinear control model is derived from the analytic model. In Part II, an optimization of the fuel flow is performed on the control model using Pontryagin’s maximum principle. The first article shows that despite the complexity of the analytic model, a tenth-order nonlinear control model with good representativity can be obtained. The second article shows that the maximum principle applied to this problem leads to a solution with optimal fuel cost. If modeling industrial processes is a complex problem in itself, obtaining a control model therefrom is just as delicate. This series of articles proposes an approach that works for the casting furnace and is indeed applicable to other industrial processes as well. In the existing analytic model, the casting furnace is treated as two one-dimensional conducting media (metal and refractories), while its chamber is seen as a well-stirred reactor. In this article, a control model is derived therefrom by a statistical method. The analytic model is run several times to obtain a set of predicted data on which a least-squares approximation is performed to determine the best parameter values to be used for the control model equations. The conduction equations in the two media are linear. The expressions for heat generation in the chamber and for radiative-convective heat transfer from the chamber to the two media are nonlinear and are kept to ensure maximum representativity. The result is a highly representative tenth-order control model, the degree of representativity being assessed by comparing the temperature outputs and the energy balances obtained from the analytic model with those obtained from the control model.     

Indentation of poly(formaldehyde), reinforced nylons and poly(ethyl terephthalate)

Indentation tests were performed on flat coupons of poly(formaldehyde), three types of reinforced nylon and reinforced poly(ethyl terephthalate) using a device developed for a typical universal tensile testing machine. Spherical, conical and cylindrical indenters were selected for this study and the effects of indenter shape combined with the rate of indentation on the behavior of these materials were examined. The load‐depth relationship during the indentation process revealed the emergence of step‐like transitions at high indentation rates. The effects of the assessed parameters on the indentation resistance, as defined in this study, were analyzed. Interest was also focused on the investigation of the deformation work as well as the plastic work taking place for a total depth that incorporated both the skin and part of the bulk of the material. The modeling of the data was based on a previously developed approach and the introduction of a modified parameter that took not only the geometrical characteristics of the indenter into account but also the indentation depth. The proposed method has proven to be useful since a steady state of the indentation pressure was observed for all materials after a certain depth of indentation. This was valid for the results obtained using both conical and cylindrical indenters. Polym. Eng. Sci. 44:1439–1447, 2004. © 2004 Society of Plastics Engineers.     

InP HBT integrated circuit technology with selectively implanted subcollector and regrown device layers

We describe a quasi-planar HBT process using a patterned implanted subcollector with a regrown MBE device layer. Using this process, we have demonstrated discrete SHBT with f/sub t/>250 GHz and DHBT with f/sub t/>230 GHz. The process eliminates the need to trade base resistance for extrinsic base/collector capacitance. Base/collector capacitance was reduced by a factor of 2 over the standard mesa device with a full overlap between the heavily doped base and subcollector regions. The low proportion of extrinsic base/collector capacitance enables further vertical scaling of the collector even in deep submicrometer emitters, thus allowing for higher current density operation. Demonstration ring oscillators fabricated with this process had excellent uniformity and yield with gate delay as low as 7 ps and power dissipation of 6 mW/CML gate. At lower bias current, the power delay product was as low as 20 fJ. To our knowledge, this is the first demonstration of high-performance HBTs and integrated circuits using a patterned implant on InP.