• Resources Recycling
• /
• v.4 no.1
• /
• pp.12-19
• /
• 1995

This study was carried out to remove the iron and impurities usmg hydrocyclone and HGMS for recycling of sludge from the granite stone cutting and polishing industrγ in the basic of chemi떠1 analysis and minerallogical investigation. This sludge consist of 70.9% $SiO_2$ 13.6% $Al_2O_3$ and It also contained 2.52% of $Fe_2O_7$ and 0.29% of $TiO_2$, as a main impurities to decrease the whiteness. As the result of hydrocyclone experiment, It was very good condition that are 100~150 g/l of sludge amount, 2.0~ 2.5 mm of underflow nozzle size, and 1.2~1.6 kg/$\textrm{cm}^2$ of pressure for 85% sludge product with the $-37{\mu}\textrm{m}$ size. $Fe_2O_3$ and $TiO_2$, contents by treatment of HGMS were decreased with 0.65% and 0.07% each at 10,000 gauss of magnetic field strength, and addih$\upsilon$n of Sodium tripolyphosphate as a dispersant was effected to get low grade F Fe,Ol and TiO, concentrate. PhYSIcal properties of this stone sludge product were showed 58.5% of whiteness, 1 13.4% of firing shrinkage and 3.0812 $\textrm{m}^2$/g of specific surface area.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.22 no.1
• /
• pp.60-71
• /
• 2019

Electrical power supply is needed to operate the radar system in the field. In addition, it should not cause performance deterioration under the environmental factors due to characteristics of military equipment, and should not cause malfunction due to electromagnetic waves generated in radar, and then should not cause malfunction in radar equipment. Therefore, By applying a permanent magnet to the rotor of the generator, light weighting and high efficiency of generator were achieved. As a result, electrical performance test of the generator, the rated output power was 80.8 kW, the maximum output power was 88.1 kW, and the output power efficiency was 98.1 % under the full load condition. When the load capacity of the generator was changed from no load to full load, the maximum voltage variation was 3.6 % and the frequency variation was 0.3 %. As a result of the transient response test for measuring the output power of the generator according to the load characteristics change, the maximum voltage variation of 7.9 %, frequency variation of 0.5 % were confirmed, and the transient response time was 2.1 seconds. Environmental tests were conducted in accordance with MIL-STD-810G and MIL-STD-461F to evaluate the operability of the generator groups. Normal operation of radar system generator group was confirmed under high temperature and low temperature environment conditions. Electromagnetic tests were conducted to check if electromagnetic wave generated from both radar system and generator group in operation caused any performance deterioration to each other. As a result, it was confirmed that the performance deterioration due to electromagnetic wave inflow, radiation, and conduction did not occur. It is expected that it should be possible to provide high efficiency power supply and stable power supply by applying to various military system as well as radar system.

• Journal of Intelligence and Information Systems
• /
• v.25 no.1
• /
• pp.163-177
• /
• 2019

As smartphones are getting widely used, human activity recognition (HAR) tasks for recognizing personal activities of smartphone users with multimodal data have been actively studied recently. The research area is expanding from the recognition of the simple body movement of an individual user to the recognition of low-level behavior and high-level behavior. However, HAR tasks for recognizing interaction behavior with other people, such as whether the user is accompanying or communicating with someone else, have gotten less attention so far. And previous research for recognizing interaction behavior has usually depended on audio, Bluetooth, and Wi-Fi sensors, which are vulnerable to privacy issues and require much time to collect enough data. Whereas physical sensors including accelerometer, magnetic field and gyroscope sensors are less vulnerable to privacy issues and can collect a large amount of data within a short time. In this paper, a method for detecting accompanying status based on deep learning model by only using multimodal physical sensor data, such as an accelerometer, magnetic field and gyroscope, was proposed. The accompanying status was defined as a redefinition of a part of the user interaction behavior, including whether the user is accompanying with an acquaintance at a close distance and the user is actively communicating with the acquaintance. A framework based on convolutional neural networks (CNN) and long short-term memory (LSTM) recurrent networks for classifying accompanying and conversation was proposed. First, a data preprocessing method which consists of time synchronization of multimodal data from different physical sensors, data normalization and sequence data generation was introduced. We applied the nearest interpolation to synchronize the time of collected data from different sensors. Normalization was performed for each x, y, z axis value of the sensor data, and the sequence data was generated according to the sliding window method. Then, the sequence data became the input for CNN, where feature maps representing local dependencies of the original sequence are extracted. The CNN consisted of 3 convolutional layers and did not have a pooling layer to maintain the temporal information of the sequence data. Next, LSTM recurrent networks received the feature maps, learned long-term dependencies from them and extracted features. The LSTM recurrent networks consisted of two layers, each with 128 cells. Finally, the extracted features were used for classification by softmax classifier. The loss function of the model was cross entropy function and the weights of the model were randomly initialized on a normal distribution with an average of 0 and a standard deviation of 0.1. The model was trained using adaptive moment estimation (ADAM) optimization algorithm and the mini batch size was set to 128. We applied dropout to input values of the LSTM recurrent networks to prevent overfitting. The initial learning rate was set to 0.001, and it decreased exponentially by 0.99 at the end of each epoch training. An Android smartphone application was developed and released to collect data. We collected smartphone data for a total of 18 subjects. Using the data, the model classified accompanying and conversation by 98.74% and 98.83% accuracy each. Both the F1 score and accuracy of the model were higher than the F1 score and accuracy of the majority vote classifier, support vector machine, and deep recurrent neural network. In the future research, we will focus on more rigorous multimodal sensor data synchronization methods that minimize the time stamp differences. In addition, we will further study transfer learning method that enables transfer of trained models tailored to the training data to the evaluation data that follows a different distribution. It is expected that a model capable of exhibiting robust recognition performance against changes in data that is not considered in the model learning stage will be obtained.

• Journal of Astronomy and Space Sciences
• /
• v.22 no.4
• /
• pp.419-430
• /
• 2005

We have carried out a statistical analysis on solar wind dynamic pressure pulses during geomagnetic storms. The Dst index was used to identify 111 geomagnetic storms that occurred in the time interval from 1997 through 2001. We have selected only the events having the minimum Dst value less than -50 nT. In order to identify the pressure impact precisely, we have used the horizontal component data of the magnetic field H (northward) at low latitudes as well as the solar wind pressure data themselves. Our analysis leads to the following results: (1) The enhancement of H due to a pressure pulse tends to be proportional to the magnitude of minimum Dst value; (2) The occurrence frequency of pressure pulses also increases with storm intensity. (3) For about $30\%$ of our storms, the occurrence frequency of pressure pulses is greater than $0.4\#/hr$, implying that to. those storms the pressure pulses occur more frequently than do periodic substorms with an average substorm duration of 2.5 hrs. In order to understand the origin of these pressure pulses, we have first examined responsible storm drivers. It turns out that $65\%$ of the studied storms we driven by coronal mass ejections (CMEs) while others are associated with corotating interaction regions $(6.3\%)$ or Type II bursts $(7.2\%)$. Out of the storms that are driven by CMEs, over $70\%$ show that the main phase interval overlaps with the sheath, namely, the region between CME body and the shock, and with the leading region of a CME. This suggests that the origin of the frequent pressure pulses is often due to density fluctuations in the sheath region and the leading edge of the CME body.