一种“KooN”结构的马尔可夫简并建模和简化计算的方法

IEC61508等标准提出了几种计算安全仪表系统需求时平均失效概率的方法,但是,对于具有冗余配置的复杂系统,随着组件数量的增加,系统的中间状态数量快速增长,用户难以构建马尔可夫模型,即便借助计算机来建模运算也较为耗时。提出了一种同型“K oo N”简并状态的马尔可夫建模的通用方法,首先是根据降级状态进行判断,将符合条件的状态进行简并,然后对标记为危险失效状态的概率进行计算。通过严格的理论推导,该简并状态方法可以在不损失精度的前提下简化马尔可夫建模。     

Combined effect of channel to rib width ratio and gas diffusion layer deformation on high temperature – Polymer electrolyte membrane fuel cell performance

The present study investigates the combined influence of Channel to Rib Width (CRW) ratio and clamping pressure on the structure and performance of High Temperature-Polymer Electrolyte Membrane Fuel Cell (HT-PEMFC) using a three-dimensional numerical model developed previously. It also considers the impact of interfacial contact resistance between the Gas Diffusion Layer (GDL) and Bipolar Plate (BPP). The structural analysis of the single straight channel HT-PEMFC geometry shows that the von-Mises stress greatly increases in the GDL under the ribs as the CRW ratio increases resulting in considerably high deformation. The cell performance analysis depicts the significance of ohmic resistance and concentration polarization for different CRW ratios, particularly at higher operating current densities. However, in low to medium current density regions, the CRW ratio has little influence on cell performance. A substantial impact on the species, overpotential, and current distributions is observed. The findings also reveal that the CRW ratio significantly affects the temperature distribution in the cell.     

A Smart Deep Convolutional Neural Network for Real-Time Surface Inspection

A proper detection and classification of defects in steel sheets in real time have become a requirement for manufacturing these products, largely used in many industrial sectors. However, computers used in the production line of small to medium size companies, in general, lack performance to attend real-time inspection with high processing demands. In this paper, a smart deep convolutional neural network for using in real-time surface inspection of steel rolling sheets is proposed. The architecture is based on the state-of-the-art SqueezeNet approach, which was originally developed for usage with autonomous vehicles. The main features of the proposed model are: small size and low computational burden. The model is 10 to 20 times smaller when compared to other networks designed for the same task, and more than 700 times smaller than general networks. Also, the number of floating-point operations for a prediction is about 50 times lower than the ones used for similar tasks. Despite its small size, the proposed model achieved near-perfect accuracy on a public dataset of 1800 images of six types of steel rolling defects.     

Acidification of the electrolyte during the galvanic corrosion of AA7075: A numerical and experimental study

The electrochemical interactions between aluminum alloy 7075 and low-carbon steels under gelled electrolytes were studied. Such electrolytes provided the opportunity to investigate both thick and thin electrolyte systems. The electrolyte was chemically modified to visually track the acidic fronts during the anodic reaction and the subsequent hydrolysis process. Two mathematical models were validated for both thick and ultrathin electrolytes. The acidification of thick electrolytes was extended some millimeters beyond the aluminum alloy surface, whereas the acidic front was localized next to the metallic joint using ultrathin electrolytes. The combination of both numerical and experimental results allows proving (and explaining why) that the acidification process is more aggressive under dilute than under concentrated electrolytes.     

Form selection of concomitant polymorphs: A case study informed by crystallization kinetics modeling

Molecular mechanisms and process kinetics of crystallizing concomitant polymorphs remain poorly understood. Solvent-mediated phase transformation and concomitant crystallization are difficult to be distinguished in practice, as multiple forms can be detected at the same time. Herein, we developed a population balance model to simulate a concomitant crystallization process of two polymorphs of tolfenamic acid. Our kinetic modeling aims to understand concomitant crystallization and help guide form selection of such a molecular system. Crystallization kinetics of ethanolic solutions were uncovered from induction time measurements, as well as seeded and unseeded crystallization experiments. Experimental and simulation results demonstrate that the stable form I crystallizes concomitantly with the metastable form II. The faster growing form II results in an intermediate decline in the composition of form I in crystallized samples, a characteristic feature of the concomitantly crystallized system. A four-quadrant scheme of attainable polymorph outcome was simulated under various crystallization conditions.     

A novel adaptive sampling based methodology for feasible region identification of compute intensive models using artificial neural network

Identification of feasible region of operations in multivariate processes is a problem of interest in several fields. This is particularly challenging when the process model is black-box in nature and/or is computationally expensive, as analytical solutions are not available and the number of possible model evaluations is limited. An efficient methodology is required to identify samples where the model is evaluated for developing a computationally efficient surrogate model. In this work, an artificial neural network based surrogate model is proposed which is integrated with a statistical-based approach (Jack-knifing) to estimate the variance of the surrogate model prediction. This allows implementation of an adaptive sampling approach where new samples are identified close to the feasible region boundary or in regions of high prediction uncertainty. The proposed approach performs better than a previously published kriging based method for different dimensionality case studies.     

Mathematical Modeling of Drying Processes of Selected Fruits and Vegetables

This study investigates the behavior of fruit and vegetable samples during drying. The experimental data are fitted to several different thin-layer drying models. Regression analysis is used to determine model parameters, while statistical indicators serve to evaluate the goodness of fit. The power function model gives the best fit for all examined samples. Based on this model, different drying and heat storage technologies can be combined to ensure that the required residual moisture content of an agricultural product is reached. It is demonstrated on the case of a specific Togolese processing plant that under favorable conditions, fossil fuel consumption can be decreased by 33 %.     

Sequential alignment attention model for scene text recognition

Scene text recognition has been a hot research topic in computer vision due to its various applications. The state-of-the-art solutions usually depend on the attention-based encoder-decoder framework that learns the mapping between input images and output sequences in a purely data-driven way. Unfortunately, there often exists severe misalignment between feature areas and text labels in real-world scenarios. To address this problem, this paper proposes a sequential alignment attention model to enhance the alignment between input images and output character sequences. In this model, an attention gated recurrent unit (AGRU) is first devised to distinguish the text and background regions, and further extract the localized features focusing on sequential text regions. Furthermore, CTC guided decoding strategy is integrated into the popular attention-based decoder, which not only helps to boost the convergence of the training but also enhances the well-aligned sequence recognition. Extensive experiments on various benchmarks, including the IIIT5k, SVT, and ICDAR datasets, show that our method substantially outperforms the state-of-the-art methods.