Assessment of Metals Loading in an Acid Mine Drainage Watershed

Watershed-scale modeling can be useful in identifying the main environmental factors and the physical mechanisms responsible for acid mine drainage (AMD) formation, attenuation, and impacts. Since flow rates and water quality of the AMD and receiving streams are related to the rainfall-runoff relationship and associated contaminant dissolution, we thought that hydrologic analysis of the mined area and surrounding drainage basin should be the starting point in documenting the source and fate of AMD contaminants. Further modeling of AMD pollutants could then be performed in terms of metal concentrations and loading at the watershed scale. In this study, monitoring was conducted in the Geopung mine watershed; the watershed analysis risk management framework (WARMF) model was used to evaluate the effect of AMD contributions to downstream metal concentrations. The hydrologic model of the basin was calibrated and verified with rainfall and streamflow data, and the water quality model was calibrated for the dissolved concentrations of metals (Cd, Cu, Zn, and Pb), using discharge data gathered in 2009. There was a strong correlation (r = 0.93) between the observed and simulated runoff values plus high Nash–Sutcliffe model efficiency (NSE = 0.89) and low average percent difference between predicted and measured values (%Diff = 0.46). Subsequent model validation using data gathered in 2010 also showed good agreement (%Diff = 9.76; NSE = 0.77; r = 0.91) between the observed and simulated values. For the metals, the model was calibrated using data from 2010; the correlation between the observed and simulated values was quite good (r = 0.80–0.41).     

Detection of process anomalies using an improved statistical learning framework

Maintenance technologies have been progressed from a time-based to a condition-based manner. The fundamental idea of condition-based maintenance (CBM) is built on the real-time diagnosis of impending failures and/or the prognosis of residual lifetime of equipment by monitoring health conditions using various sensors. The success of CBM, therefore, hinges on the capability to develop accurate diagnosis/prognosis models. Even though there may be an unlimited number of methods to implement models, the models can normally be classified into two categories in terms of their origins: using physical principles or historical observations. We have focused on the latter method (sometimes referred as the empirical model based on statistical learning) because of some practical benefits such as context-free applicability, configuration flexibility, and customization adaptability. While several pilot-scale systems using empirical models have been applied to work sites in Korea, it should be noted that these do not seem to be generally competitive against conventional physical models. As a result of investigating the bottlenecks of previous attempts, we have recognized the need for a novel strategy for grouping correlated variables such that an empirical model can accept not only statistical correlation but also some extent of physical knowledge of a system. Detailed examples of problems are as follows: (1) missing of important signals in a group caused by the lack of observations, (2) problems of signals with the time delay, and (3) problems of optimal kernel bandwidth. This paper presents an improved statistical learning framework including the proposed strategy and case studies illustrating the performance of the method.     

Synthesis of branched carbon nanotubes by carbonization of solid polyvinylidene fluoride fibers

Branched carbon nanotubes (b-CNTs) were synthesized by carbonization of polyvinylidene fluoride (PVDF) fibers containing a Pt catalyst. The solid fibrous polymer converted into carbon nanotubes with simultaneous growth of branches on the surface of the tubes during carbonization. The Pt particles were expected to decompose PVDF polymer inside the tubes into volatile carbonaceous species leaving a hollow center, and also to act as catalytic sites for the growth of carbon branches. The resulting b-CNTs had a high degree of graphitization and a large electrochemical surface area, and also showed a possibility as a supporting material for electrocatalysts.     

Prediction of the heat transfer characteristics on natural convection in a long rectangular enclosure with a twisted cylinder using direct simulation and artificial neural networks

Journal of Mechanical Science and Technology - A three-dimensional direct simulation was carried out to examine the natural convective fluid flow and heat transfer characteristics in a long...     

Characterisation of low‐digestible starch fractions isolated from amylosucrase‐modified waxy corn starch

Amylosucrase (AS) modification of starch increases the slowly digestible (SDS) and resistant starch (RS) fractions. However, the characteristics and formation mechanism of each fraction of AS‐modified starch have not been determined yet. Therefore, this study isolated SDS and/or RS from AS‐modified waxy corn starches and investigated their structural characteristics. The amount of SDS+RS and RS had a positive correlation with the proportion of the medium length (13–24 of degree of polymerisation) branched chains of amylopectin. The relative crystallinity increased in the order of AS‐modified starch < SDS+RS < RS, while maintaining the B‐type crystalline structure. The thermal transition temperature ranges of the isolated fractions were also higher than those of undigested starches. The medium branched chains of amylopectin were presumably the clincher for the SDS and/or RS formation in AS‐modified starches. The principal causes of SDS and RS formation were the chain length elongation and the subsequent retrogradation‐like process, respectively.     

Background Information of Deep Learning for Structural Engineering

Since the first journal article on structural engineering applications of neural networks (NN) was published, there have been a large number of articles about structural analysis and design problems using machine learning techniques. However, due to a fundamental limitation of traditional methods, attempts to apply artificial NN concept to structural analysis problems have been reduced significantly over the last decade. Recent advances in deep learning techniques can provide a more suitable solution to those problems. In this study, versatile background information, such as alleviating overfitting methods with hyper-parameters, is presented. A well-known ten bar truss example is presented to show condition for neural networks, and role of hyper-parameters in the structures.     

Crystal structure,magnetic properties,and magnetization variation in Bi0.84La0.16Fe1-xTixO3 polycrystalline ceramic

The crystal structure, microstructure, and magnetic properties of Bi_0.84La_0.16Fe_1-xTi_xO₃ (x?=?0.1, 0.14, 0.16, 0.2) ceramic samples prepared by solid-state reaction are investigated by combining X-ray diffraction, Raman scattering, scanning electron microscope, and magnetic measurements. The analysis of crystal structure reveals the presence of multiphase structures and the gradual formation of Aurivillus Bi₅FeTi₃O₁₅ phase when increasing Ti concentration. The microstructure study shows several grain shapes, corresponding to the different crystal structures and phases present in the sample. The competition in lattice strain of the coexisting phases leads to the isothermal structural transition and the self-change in magnetization. A drastic decrease in coercivity and an increase in magnetization are observed with increasing ferromagnetic-like hysteresis loops. The change in magnetic properties of the samples is strongly dependent on the coexisting phases in the samples.