• Journal of Intelligence and Information Systems
• /
• v.25 no.3
• /
• pp.43-62
• /
• 2019

At one time, the anomaly detection sector dominated the method of determining whether there was an abnormality based on the statistics derived from specific data. This methodology was possible because the dimension of the data was simple in the past, so the classical statistical method could work effectively. However, as the characteristics of data have changed complexly in the era of big data, it has become more difficult to accurately analyze and predict the data that occurs throughout the industry in the conventional way. Therefore, SVM and Decision Tree based supervised learning algorithms were used. However, there is peculiarity that supervised learning based model can only accurately predict the test data, when the number of classes is equal to the number of normal classes and most of the data generated in the industry has unbalanced data class. Therefore, the predicted results are not always valid when supervised learning model is applied. In order to overcome these drawbacks, many studies now use the unsupervised learning-based model that is not influenced by class distribution, such as autoencoder or generative adversarial networks. In this paper, we propose a method to detect anomalies using generative adversarial networks. AnoGAN, introduced in the study of Thomas et al (2017), is a classification model that performs abnormal detection of medical images. It was composed of a Convolution Neural Net and was used in the field of detection. On the other hand, sequencing data abnormality detection using generative adversarial network is a lack of research papers compared to image data. Of course, in Li et al (2018), a study by Li et al (LSTM), a type of recurrent neural network, has proposed a model to classify the abnormities of numerical sequence data, but it has not been used for categorical sequence data, as well as feature matching method applied by salans et al.(2016). So it suggests that there are a number of studies to be tried on in the ideal classification of sequence data through a generative adversarial Network. In order to learn the sequence data, the structure of the generative adversarial networks is composed of LSTM, and the 2 stacked-LSTM of the generator is composed of 32-dim hidden unit layers and 64-dim hidden unit layers. The LSTM of the discriminator consists of 64-dim hidden unit layer were used. In the process of deriving abnormal scores from existing paper of Anomaly Detection for Sequence data, entropy values of probability of actual data are used in the process of deriving abnormal scores. but in this paper, as mentioned earlier, abnormal scores have been derived by using feature matching techniques. In addition, the process of optimizing latent variables was designed with LSTM to improve model performance. The modified form of generative adversarial model was more accurate in all experiments than the autoencoder in terms of precision and was approximately 7% higher in accuracy. In terms of Robustness, Generative adversarial networks also performed better than autoencoder. Because generative adversarial networks can learn data distribution from real categorical sequence data, Unaffected by a single normal data. But autoencoder is not. Result of Robustness test showed that he accuracy of the autocoder was 92%, the accuracy of the hostile neural network was 96%, and in terms of sensitivity, the autocoder was 40% and the hostile neural network was 51%. In this paper, experiments have also been conducted to show how much performance changes due to differences in the optimization structure of potential variables. As a result, the level of 1% was improved in terms of sensitivity. These results suggest that it presented a new perspective on optimizing latent variable that were relatively insignificant.

• Korean Journal of Fisheries and Aquatic Sciences
• /
• v.12 no.3
• /
• pp.125-130
• /
• 1979

In this paper, the noise pressure propagated in the air on account of the engine revolution of a stern trawler, Sae-Ba-Da(G. T. 2275.71) was measured at the check points No.1 through No.43 when the vessel was cruising, towing nets, and drifting. The experiment was carried out in the period from August 23 to October 22, 1978 at the locations of lat. $33^{\circ}$ 47'N, long. $127^{\circ}$ 34'E; lat. $34^{\circ}$ 24'N, long. $128^{\circ}$ 23'E; and lat. $6^{\circ}$ 01'N, long. $108^{\circ}$ 04'E. In case of cruising, noise on the weather deck came from funnel noise. The highest noise pressure was 92dB at observation point No.9 where tile noise pressure from main engine was 105dB when the engine was operated at 730rpm and $12^{\circ}$ sorely propeller pitch. The noise measured was reduced to 90dB at observation point No.9 when the screw propeller pitch was changed to $8^{\circ}$ that resulted in reduction of engine to 103dB. In case of towing net, the main engine revolution and screw propeller pitch was fixed at 730rpm and $8^{\circ}$ respectively. But the engine noise pressure was increased up to 106dB due to the towing resistance by 14 tons of the nets, and the noise pressure was 90dB at No.9 point. A hight noise was also generated from screw because of the towing reoistance and could be measurable even in the wall of the insulated freezing room. When the vessel was drifting: the noise pressure from the generator operated, at 720rpm was 100dB. This caused 87dB noise pressure at No.9 point. The noise pressure in the boarding or residence sections was 45 to 60dB in each case of cruisinrg towing net or drifting but it was so high as 82dB on the open deck that voice could hardly be heap.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.25 no.6
• /
• pp.346-361
• /
• 2013

This article reviews the papers published in the Korean Journal of Air-Conditioning and Refrigeration Engineering during 2012. It is intended to understand the status of current research in the areas of heating, cooling, ventilation, sanitation, and indoor environments of buildings and plant facilities. The conclusions are as follows : (1) The research works on thermal and fluid engineering have been reviewed as groups of fluid machinery, pipes and valves, fuel cells and power plants, ground-coupled heat pumps, and general heat and mass transfer systems. Research issues are mainly focused on new and renewable energy systems, such as fuel cells, ocean thermal energy conversion power plants, and ground-coupled heat pump systems. (2) Research works on the heat transfer area have been reviewed in the categories of heat transfer characteristics, pool boiling and condensing heat transfer, and industrial heat exchangers. Researches on heat transfer characteristics included the results for natural convection in a square enclosure with two hot circular cylinders, non-uniform grooved tube considering tube expansion, single-tube annular baffle system, broadcasting LED light with ion wind generator, mechanical property and microstructure of SA213 P92 boiler pipe steel, and flat plate using multiple tripping wires. In the area of pool boiling and condensing heat transfer, researches on the design of a micro-channel heat exchanger for a heat pump, numerical simulation of a heat pump evaporator considering the pressure drop in the distributor and capillary tubes, critical heat flux on a thermoexcel-E enhanced surface, and the performance of a fin-and-tube condenser with non-uniform air distribution and different tube types were actively carried out. In the area of industrial heat exchangers, researches on a plate heat exchanger type dehumidifier, fin-tube heat exchanger, an electric circuit transient analogy model in a vertical closed loop ground heat exchanger, heat transfer characteristics of a double skin window for plant factory, a regenerative heat exchanger depending on its porous structure, and various types of plate heat exchangers were performed. (3) In the field of refrigeration, various studies were executed to improve refrigeration system performance, and to evaluate the applicability of alternative refrigerants and new components. Various topics were presented in the area of refrigeration cycle. Research issues mainly focused on the enhancement of the system performance. In the alternative refrigerant area, studies on CO2, R32/R152a mixture, and R1234yf were performed. Studies on the design and performance analysis of various compressors and evaporator were executed. (4) In building mechanical system research fields, twenty-nine studies were conducted to achieve effective design of mechanical systems, and also to maximize the energy efficiency of buildings. The topics of the studies included heating and cooling, HVAC system, ventilation, renewable energy systems, and lighting systems in buildings. New designs and performance tests using numerical methods and experiments provide useful information and key data, which can improve the energy efficiency of buildings. (5) In the fields of the architectural environment, studies for various purposes, such as indoor environment, building energy, and renewable energy were performed. In particular, building energy-related researches and renewable energy systems have been mainly studied, reflecting interests in global climate change, and efforts to reduce building energy consumption by government and architectural specialists. In addition, many researches have been conducted regarding indoor environments.

• Journal of Korean Society of Occupational and Environmental Hygiene
• /
• v.11 no.3
• /
• pp.229-234
• /
• 2001

The EDXRF(Energy Dispersive X-ray Fluorescence Spectrometer) technique was applied to the determination of heavy metals in welding fume. The EDXRF method designed in this study was a non-destructive analysis method. Samples were analyzed directly by EDXRF without any pre-treatment such as digestion and dilution. The samples used to evaluate this method were laboratory samples exposed in a chamber connected with a welding fume generator. The samples were first analyzed using a non-destructive EDXRF method. The samples subsequently were analyzed using AAS method to verify accuray of the EDXRF method. The purpose of this study was to evaluate the possibility of the non-destructive analysis of heavy metals in welding fume by EDXRF. The results of this study were as follow: 1.When the samples were collected under the open-face sampling condition, a surface distribution of welding fume particles on sample filters was uniform, which made non-destructive analysis possible. 2. The method was statistically evaluated according to the NIOSH(National Institute for Occupational Safety and Health) and HSE(Health and Safety Executive) method. 3. The overall precision of the EDXRF method Was calculated at 3.45 % for Cr, 2.57 % for Fe and 3.78 % for Mn as relative standard deviation(RSD), respectively. The limits of detection were calculated at $0.46{\mu}g$/sample for Cr, $0.20{\mu}g$/sample for Fe and $1.14{\mu}g$/sample for Mn, respectively. 4. A comparison between the results of Cr, Fe, Mn analyzed by EDXRF and AAS was made in order to assess the accuracy of EDXRF method. The correlation coefficient between the results of EDXRF and AAS was 0.9985 for Cr, 0.9995 for Fe and 0.9982 for Mn, respectively. The overall uncertainty was determined to be ${\pm}12.31%$, 8.64 % and 11.91 % for Cr, Fe and Mn, respectively. In conclusion, this study showed that Cr, Fe, Mn in welding fume were successfully analyzed by the EDXRF without any sample pre-treatment such as digestion and dilution and a good correlation between the results of EDXRF and AAS was obtained. It was thus possible to use the EDXRF technique as an analysis method of working environment samples. The EDXRF method was an efficient method in a non-destructive analysis of heavy metals in welding fume.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.22 no.2
• /
• pp.84-87
• /
• 2018

Purpose The radiopharmaceutical used in the nuclear medicine department is used only for the specific patient according to the prescription or instruction of the doctor without selling, so it is dispensed and it is distributed and used for the examination. Radiopharmaceuticals administered to patients should be managed appropriately as well as radiation safety management during dispensation. The purpose of this study is to investigate microbial contamination during dispensation of radiopharmaceuticals Materials and Methods This study distinguished between general workbench and clean workbench and performed three tests. First, microbial cultivation test of radiopharmaceutical prepared and dispensed in general workbenches and sterile workbenches were carried out five times, respectively. The second test was performed settle plate method three times before and after the use of the exhaust filter. Finally, Adenosine Triphosphate (ATP) measurement was performed in each workbench to measure bacterial counts. In addition, ATP measurement were carried out by designating locations and items that may be contaminated during dispensation. Results In the microbial culture test, no microorganisms were detected in both samples. In the settle plate method, it was detected without using of the exhaust filter in a general workbench once. In the ATP measurement test, it was measured at the level of 400 RLU or less, which is the standard value of contamination, in both workbenches surface. In additional ATP measurement test, the refrigerator handle in the distribution room was measured above the reference value of 1217 RLU, the vacuum vial shield of the Tech Generator at 435 RLU, and the syringe holder at 1357 RLU. After environmental disinfection, the results were reduced to 311 RLU, 136 RLU, and 291 RLU. Conclusion No contamination by bacteria was found in both workbenches. However, microbial contamination may occur if the use of an exhaust filter or proper hand hygiene is not achieved. Regular inspections and management for aseptic processing themselves will be necessary.