• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.50 no.3
• /
• pp.181-192
• /
• 2022

Active vibration control system(AVCS) for helicopter enables to control the vibration generated from the main rotor and has the superb vibration reduction performance with low weight compared passive vibration reduction device. In this paper, FxLMS algorithm-based vibration control software of the light civil helicopter tansmits the control command calculated using the signals of the tachometer and accelerometers to the circular force generator(CFG) is developed and verified. According to the RTCA DO-178C/DO-331, the vibration control software is developed through the model based design technique, and real-time operation performance is evaluated in PILS(processor in-the loop simulation) and HILS(hardware in-the loop simulation) environments. In particular, the reliability of the software is improved through the LDRA-based verification coverage in the PIL environments. In order to AVCS to light civil helicopter(LCH), the dynamic response characteristic model is obtained through the ground/flight tests. AVCS configuration which exhibits the optimal performance is determined using system optimization analysis and flight test and obtain STC certification.

• Composites Research
• /
• v.35 no.5
• /
• pp.317-321
• /
• 2022

In this study, the PIC design method with machine learning that automatically assigning different stacking sequences according to loading types was applied bumper beam. The input value and labels of the training data for applying machine learning were defined as coordinates and loading types of reference elements that are part of the total elements, respectively. In order to compare the 2D and 3D implementation method, which are methods of representing coordinate value, training data were generated, and machine learning models were trained with each method. The 2D implementation method is divided FE model into each face and generating learning data and training machine learning models accordingly. The 3D implementation method is training one machine learning model by generating training data from the entire finite element model. The hyperparameter were tuned to optimal values through the Bayesian algorithm, and the k-NN classification method showed the highest prediction rate and AUC-ROC among the tuned models. The 3D implementation method revealed higher performance than the 2D implementation method. The loading type data predicted through the machine learning model were mapped to the finite element model and comparatively verified through FE analysis. It was found that 3D implementation PIC bumper beam was superior to 2D implementation and uni-stacking sequence composite bumper.

• The Journal of the Convergence on Culture Technology
• /
• v.8 no.5
• /
• pp.497-502
• /
• 2022

Information delivery among young generation has a distinct tendency to prefer visual to text as means of information distribution and sharing recently, and it is natural to distribute information through Youtube or one-man broadcasting on Internet. That is, young generation usually get their information through this kind of distribution procedure. Many young generation are also drastic and more aggressive for decorating themselves very uniquely. It tends to create personal characteristics freely through drastic expression and attempt of face makeup, hair styling and fashion coordination without distinction of sex. Especially, face makeup becomes an object of major concern among males nowadays, and female of course, then it is the major means to express their personality. In this study, to meet the demands of the times, we design and implement the intellectual smart mirror application that efficiently retrieves and recommends the related videos among Youtube or one-man broadcastings produced by famous professional makeup artists to implement the face makeup congruous with our face shape, hair color & style, skin tone, fashion color & style in order to create the face makeup that represent our characteristics. We also introduce the AI technique to provide optimal solution based on the learning of user's search patterns and facial features, and finally provide the detailed makeup face images to give the chance to get the makeup skill stage by stage.

• Journal of Intelligence and Information Systems
• /
• v.17 no.4
• /
• pp.241-254
• /
• 2011

Financial time-series forecasting is one of the most important issues because it is essential for the risk management of financial institutions. Therefore, researchers have tried to forecast financial time-series using various data mining techniques such as regression, artificial neural networks, decision trees, k-nearest neighbor etc. Recently, support vector machines (SVMs) are popularly applied to this research area because they have advantages that they don't require huge training data and have low possibility of overfitting. However, a user must determine several design factors by heuristics in order to use SVM. For example, the selection of appropriate kernel function and its parameters and proper feature subset selection are major design factors of SVM. Other than these factors, the proper selection of instance subset may also improve the forecasting performance of SVM by eliminating irrelevant and distorting training instances. Nonetheless, there have been few studies that have applied instance selection to SVM, especially in the domain of stock market prediction. Instance selection tries to choose proper instance subsets from original training data. It may be considered as a method of knowledge refinement and it maintains the instance-base. This study proposes the novel instance selection algorithm for SVMs. The proposed technique in this study uses genetic algorithm (GA) to optimize instance selection process with parameter optimization simultaneously. We call the model as ISVM (SVM with Instance selection) in this study. Experiments on stock market data are implemented using ISVM. In this study, the GA searches for optimal or near-optimal values of kernel parameters and relevant instances for SVMs. This study needs two sets of parameters in chromosomes in GA setting : The codes for kernel parameters and for instance selection. For the controlling parameters of the GA search, the population size is set at 50 organisms and the value of the crossover rate is set at 0.7 while the mutation rate is 0.1. As the stopping condition, 50 generations are permitted. The application data used in this study consists of technical indicators and the direction of change in the daily Korea stock price index (KOSPI). The total number of samples is 2218 trading days. We separate the whole data into three subsets as training, test, hold-out data set. The number of data in each subset is 1056, 581, 581 respectively. This study compares ISVM to several comparative models including logistic regression (logit), backpropagation neural networks (ANN), nearest neighbor (1-NN), conventional SVM (SVM) and SVM with the optimized parameters (PSVM). In especial, PSVM uses optimized kernel parameters by the genetic algorithm. The experimental results show that ISVM outperforms 1-NN by 15.32%, ANN by 6.89%, Logit and SVM by 5.34%, and PSVM by 4.82% for the holdout data. For ISVM, only 556 data from 1056 original training data are used to produce the result. In addition, the two-sample test for proportions is used to examine whether ISVM significantly outperforms other comparative models. The results indicate that ISVM outperforms ANN and 1-NN at the 1% statistical significance level. In addition, ISVM performs better than Logit, SVM and PSVM at the 5% statistical significance level.

• Journal of Korean Society of Water and Wastewater
• /
• v.21 no.4
• /
• pp.493-501
• /
• 2007

The membrane pilot plant has being operated in the Hyeondo pumping station to find the optimal operation technique of Gong-Ju membrane water treatment plant (WTP) which is constructing in $250m^3/d$ scale. The pilot plant was consisted of two trains which can treat $30,000m^3/d$ per train. First train was operated for one year under the condition of flux $1m^3/m^2{\cdot}d$ while the effects of flux variation and addition of powdered activated carbon(PAC) were evaluated in second train. The turbidity of membrane product water of first train which is operated on Flux $1m^3/m^2{\cdot}d$ was always below 0.05 NTU regardless of raw water turbidity. And also, the trance-membrane pressure(TMP) was maintained at $0.3{\sim}0.5kgf/cm^2$ for about 9 months and increased rapidly to $1.8kgf/cm^2$ which is maximum operating TMP. However, TMP was rapidly increased to $1.8kgf/cm^2$ within 2 months as flux was increased from 1 to $2m^3/m^2{\cdot}d$, especially, within 10 days under high turbidity(30~50NTU). This reault means that if Gongju membrane WTP is operated in flux $1m^3/m^2{\cdot}d$, chemical cleaning period can be maintained over 6 months. Only 10% of dissolved organic carbon (DOC) was removed in membrane process while the removal efficiencies of manganese and iron were 60% and 77% respectively. However, because only solid manganese and iron were removed in membrane process, an additional process for treating soluble manganese is required if souble manganese is high in raw water. 70% of 70ng/L 2-MIB which is causing taste & odor was removed in powdered activated carbon (PAC) tank with 50mg/L PAC which is design concentration of Gongju WTP. In addition, TMP was reduced with addition of 50mg/L PAC regardless of flux. Because TMP was not influenced even if 100mg/L PAC was added, the high taste and odor problem can be controled by additional injection of PAC.

• Korean Chemical Engineering Research
• /
• v.55 no.3
• /
• pp.332-340
• /
• 2017

This study aims to propose a new optimization-based approach for design and analysis of the stand-alone hybrid energy supply system using renewable energy sources (RES). In the energy supply system, we include multiple energy production technologies such as Photovoltaics (PV), Wind turbine, and fossil-fuel-based AC generator along with different types of energy storage and conversion technologies such as battery and inverter. We then select six different regions of Korea to represent various characteristics of different RES potentials and demand profiles. We finally designed and analyzed the optimal RES stand-alone energy supply system in the selected regions using multiobjective optimization (MOOP) technique, which includes two objective functions: the minimum cost and the minimum $CO_2$ emission. In addition, we discussed the feasibility and expecting benefits of the systems by comparing to conventional systems of Korea. As a result, the region of the highest RES potential showed the possibility to remarkably reduce $CO_2$ emissions compared to the conventional system. Besides, the levelized cost of electricity (LCOE) of the RES-based energy system is identified to be slightly higher than conventional energy system: 0.35 and 0.46 $/kWh, respectively. However, the total life-cycle emission of $CO_2$ ($LCE_{CO2}$) can be reduced up to 470 g$CO_2$/kWh from 490 g$CO_2$/kWh of the conventional systems.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.18 no.10
• /
• pp.42-47
• /
• 2017

In this study,we conducted an injection molding analysis of metal powder materials for the development of flanges, which are necessary adapters for optical communication. The metal powder injection molding process is a technique for producing an injection molded article having a complicated shape by mixing ceramic or stainless powder and binders. It is used to produce products which require complex processing technology or for which the productivity is low. The purpose of this study is to minimize the manufacturing processing of products which are manufactured through existing mechanical processing procedures. For the injection molding analysis, we mixed stainless STS316 metal powder with binders at a ratio of 6 to 4 to make molding materials consisting of granular pellets. Then, three-dimensional modeling and meshing were carried out to obtain the optimal injection molding analysis conditions(molding temperature, melting temperature, injection time, injection temperature, injection pressure, packing time and cooling time). As a result of the analysis, it was discovered that the inlet became available 13.29 seconds after the first injection. Also, as the flowing and packing in the melt through the sprue, runner and gate were stable, it is expected that good molds can be manufactured.

• Journal of Korean Society for Geospatial Information Science
• /
• v.23 no.1
• /
• pp.73-79
• /
• 2015

Recently, BIM proliferates at high speeds so that BIM planning is trying to be used in various kinds of engineering. However, when using different software during design phase and construction phase, the problem of mutual compatibility is coming out. Even the BIM technology has been used or the practical applicability has been made into result, it would be fair to say that BIM has limitations in the visual level. In this research, it is meaningless to obtain the BIM result as the primary purpose. As the usefulness of it is judged incomplete, we committed to master the trend and problems of terrain spatial information systems and BIM. Furthermore, the plan of building the BIM in the civil field, especially the civil-BIM based on the technology ofterrain spatial information has been presented. It can be judged through this research that the high-capacity DB of BIM occurred during the whole process may cause poor performance of the following stage the structure system which connects the terrain spatial information and civil-BIM. In order to manage the optimal full cycle, the spatial analysis technology of the stages after choosing the DB has been described.

• Korean Chemical Engineering Research
• /
• v.48 no.6
• /
• pp.717-724
• /
• 2010

The flash point is one of the most important physical properties used to determine the potential for fire and explosion hazards of flammable liquids. Despite the needs of the experimental flash point data for the design and construction of chemical plants, there is often a significant gap between the demands for the data and their availability. This study have built and compared two models of partial least squares(PLS) and support vector machine(SVM) to predict the experimental flash points of 893 organic compounds out of DIPPR 801. As the independent variables of the models, 65 functional groups were chosen based on the group contribution method that was oriented from the assumption that each fragment of a molecule contributes a certain amount to the value of its physical property, and the logarithm of molecular weight was added. The prediction errors calculated from cross-validation were employed to determine the optimal parameters of two models. And, an optimization technique should be used to get three parameters of SVM model. This work adopted particle swarm optimization that is one of heuristic optimization methods. As the selection of training data can affect the prediction performance, 100 data sets of randomly selected data were generated and tested. The PLS and SVM results of the average absolute errors for the whole data range from 13.86 K to 14.55 K and 7.44 K to 10.26 K, respectively, indicating that the predictive ability of the SVM is much superior than PLS.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.373-376
• /
• 2018

Traditional paprika smart farm systems are often harmful to paprika growth because they are set to follow the values of several sensors to the reference value, so the system is often unable to make optimal judgement. Using decision tree techniques, the expert system for the paprika smart farm is designed to create a control system with a decision-making structure similar to that of farmers using data generated by factors that depend on their surroundings. With the current smart farm control system, it is essential for farmers to intervene in the surrounding environment because it is designed to follow sensor values to the reference values set by the farmer. To solve this problem even slightly, it is going to obtain environmental data and design controllers that apply decision tree method. The expert system is established for complex control by selecting the most influential environmental factors before controlling the paprika smart farm equipment, including criteria for selecting decisions by farmers. The study predicts that each environmental element will be a standard when creating smart farms for professionals because of the interrelationships of data, and more surrounding environmental factors affecting growth.