• 한국지하수토양환경학회지:지하수토양환경
• /
• 제11권6호
• /
• pp.83-91
• /
• 2006

회분식 반응조에서 마이크로(mZVI) 및 나노(nZVI) 크기의 영가 철을 환원물질로 이용하여 고폭화약물질 3종에 대한 환원동역학을 측정하였다. 각 화약류를 이용하여 어미물질에 대한 nZVI와 mZVI의 비표면적 환원상수 $k_{SA}$과 비중량 환원상수 $k_{M}$을 측정한 후, 중간산물의 거동을 비교하였다. 그 결과 두 상수를 사용해서는 nZVI 반응조내 어미 물질과 중간환원산물들의 거동을 완전히 설명할 수 없었다. 화약물질을 mZVI로 처리한 반응조에서는 초기 환원물질인 nitroso-RDXs, nitroso-HMXs 및 hydroxylamino-TNT가 주로 축적되었으나, nZVI로 처리한 반응조에서는 동일한 겉보기 반응속도임에도 불구하고 환원말기물질인 극성중간산물들과 TAT가 축적되었다. 그러므로 중간산물들의 환원까지 고려하는 새로운 매계변수의 개발이 필요한 것으로 판단된다.

• 공업화학
• /
• 제4권1호
• /
• pp.82-93
• /
• 1993

염소화 공정을 통해 batch-boat-system에서 회중석으로부터 텅스텐 성분을 추출하기 위해, 환원제인 탄소의 무게비, 반응온도, 반응시간, $Cl_2$ gas의 유량 그리고 시료입도와 같은 주요 반응 변수들에 대한 영향을 조사하였다. 이 염소화 공정들에 대한 적정조건들은 반응온도 $700^{\circ}C$ 이상, 광물시료에 대한 탄소의 무게비 0.08, 반응시간 20분, $Cl_2$ gas의 유량 $0.6{\ell}/min$, 광물시료의 입도 -200mesh였으며 위 조건하에서 광물중 99%의 tungsten성분이 추출되었다. 반응속도는 고온에서는 $Cl_2$ gas의 확산단계가, 저온에서는 화학반응단계가 속도결정단계로 보이며 각각의 단계에서 활성화 에너지는 고온 부분에서는 7.98kcal/mol이며 저온 부분에서는 31.2kcal/mol이었다.

• Journal of Microbiology
• /
• 제40권2호
• /
• pp.119-124
• /
• 2002

Experiments for the characterization of inactivation were performed in a series of batch processes with the total coliform used as a general indicator organism based on the chlorine residuals as a disinfectant. The water samples were taken from the outlet of a settling basin in a conventional surface water treat- ment system that is provided with the raw water drawn from the mid-stream of the Han River, The inactivation of total coliform was experimentally analysed for the dose of disinfectants contact time, filtration and mixing intensity. The curves obtained from a series of batch processes were shaped with a general tailing-off and biphasic mode of inactivation, i.e. a sharp loss of bacterial viability within 15 min followed by an extended phase. In order to observe the effect of carry-over suspended solids on chlorine consumption and disinfection efficiency, the water samples were filtered, prior to inoculation with coliforms, with membranes of both 2.5$\mu$m and 11.0 $\mu$m pore size, and with a sand tilter of 1.0 mm in effective size and of 1.4 in uniformity coefficient. As far as the disinfection efficiency is concerned, there were no significant differences. The parameters estimated by the models of Chick-Wat-son, Hom and Selleck from our experimental data obtained within 120 min are: log(N/N$\_$0/)=-0.16CT with n=1, leg(N/N$\_$0/)=-0.71C$\^$0.87/ with n 1 for the Chick-Watson model, log (N/N$\_$0/)=-1.87C$\^$0.47/ T$\^$0.36/ for the Hom model, log (MHo)=-2.13log (1+CT/0.11) for the Selleck model. It is notable that among the models reviewed with regard to the experimental data obtained, the Selleck model appeared to most closely resemble the total coliform survival curve.

• 한국지반공학회논문집
• /
• 제20권1호
• /
• pp.29-36
• /
• 2004

암모늄으로 오염된 불량 매립지의 주변지반 지하수를 정화함에 있어 반응벽체를 적용할 경우, 고려해야 할 주요 설계인자를 평가하기 위해 회분식 실험, 투수시험, 주상실험을 수행하였다. 반응물질로는 높은 양이온교환능(CEC)를 가지는, 천연 제올라이트의 일종인 Clinoptilolite를 사용하였다. 회분식 실험의 경우, 암모늄 오염액의 초기농도와 Clinoptilolite 입자크기를 변화시키며 Clinoptilolite의 암모늄에 대한 제거율을 평가하였다. 암모늄의 초기농도가 80ppm으로 고농도인 경우를 제외하고는 단위 량의 Clinoptilolite로 약 80% 암모늄을 제거할 수 있었다. Clinoptilolite의 입자크기에 의한 영향은 뚜렷하지 않았다. 투수시험은 Clinoptilolite와 주문진사를 무게비 20 : 80으로 혼합한 후 시편을 성형하여 수행하였다. 투수시험에는 연성벽체 투수기를 사용하였다. 시험결과, 세척된 0.42∼0.85mm의 크기를 가지는 Clinoptilolite를 포함하는 시편이, 약 $10^{-3}$cm/s의 투수계수로 가장 높은 값을 보였다. 주상실험에서는 실제 매립지 침출수를 이용해, 유동 상태에서 암모늄 이온을 포함한 침출수와 Clinoptilolite 혼합토의 반응성을 검토하였다. 본 연구를 통해 Clinoptilolite는 암모늄을 정화하고자 하는 반응벽체에 적용 가능한 반응물질로 판단되었다.

• 대한원격탐사학회지
• /
• 제35권2호
• /
• pp.279-288
• /
• 2019

본 연구에서는 항공정사영상을 이용하여 SegNet 기반의 의미분할을 수행하고, 토지피복분류에서의 그 성능을 평가하였다. 의미분할을 위한 분류 항목을 4가지(시가화건조지역, 농지, 산림, 수역)로 선정하였고, 항공정사영상과 세분류 토지피복도를 이용하여 총 2,000개의 데이터셋을 8:2 비율로 훈련(1,600개) 및 검증(400개)로 구분하여 구축하였다. 구축된 데이터셋은 훈련과 검증으로 나누어 학습하였고, 모델 학습 시 정확도에 영향을 미치는 하이퍼파라미터의 변화에 따른 검증 정확도를 평가하였다. SegNet 모델 검증 결과 반복횟수 100,000회, batch size 5에서 가장 높은 성능을 보였다. 이상과 같이 훈련된 SegNet 모델을 이용하여 테스트 데이터셋 200개에 대한 의미분할을 수행한 결과, 항목별 정확도는 농지(87.89%), 산림(87.18%), 수역(83.66%), 시가화건조지역(82.67%), 전체 분류정확도는 85.48%로 나타났다. 이 결과는 기존의 항공영상을 활용한 토지피복분류연구보다 향상된 정확도를 나타냈으며, 딥러닝 기반 의미분할 기법의 적용 가능성이 충분하다고 판단된다. 향후 다양한 채널의 자료와 지수의 활용과 함께 분류 정확도 향상에 크게 기여할 수 있을 것으로 기대된다.

• 치위생과학회지
• /
• 제20권4호
• /
• pp.206-212
• /
• 2020

Background: A Teachable Machine is a kind of machine learning web-based tool for general persons. In this paper, the feasibility of Google's Teachable Machine (ver. 2.0) was studied in the diagnosis of the tooth-marked tongue. Methods: For machine learning of tooth-marked tongue diagnosis, a total of 1,250 tongue images were used on Kaggle's web site. Ninety percent of the images were used for the training data set, and the remaining 10% were used for the test data set. Using Google's Teachable Machine (ver. 2.0), machine learning was performed using separated images. To optimize the machine learning parameters, I measured the diagnosis accuracies according to the value of epoch, batch size, and learning rate. After hyper-parameter tuning, the ROC (receiver operating characteristic) analysis method determined the sensitivity (true positive rate, TPR) and specificity (false positive rate, FPR) of the machine learning model to diagnose the tooth-marked tongue. Results: To evaluate the usefulness of the Teachable Machine in clinical application, I used 634 tooth-marked tongue images and 491 no-marked tongue images for machine learning. When the epoch, batch size, and learning rate as hyper-parameters were 75, 0.0001, and 128, respectively, the accuracy of the tooth-marked tongue's diagnosis was best. The accuracies for the tooth-marked tongue and the no-marked tongue were 92.1% and 72.6%, respectively. And, the sensitivity (TPR) and specificity (FPR) were 0.92 and 0.28, respectively. Conclusion: These results are more accurate than Li's experimental results calculated with convolution neural network. Google's Teachable Machines show good performance by hyper-parameters tuning in the diagnosis of the tooth-marked tongue. We confirmed that the tool is useful for several clinical applications.

• International Journal of Computer Science & Network Security
• /
• 제22권9호
• /
• pp.175-182
• /
• 2022

In our daily life, we come across different types of information, for example in the format of multimedia and text. We all need different types of information for our common routines as watching/reading the news, listening to the radio, and watching different types of videos. However, sometimes we could run into problems when a certain type of information is required. For example, someone is listening to the radio and wants to listen to jazz, and unfortunately, all the radio channels play pop music mixed with advertisements. The listener gets stuck with pop music and gives up searching for jazz. So, the above example can be solved with an automatic audio classification system. Deep Learning (DL) models could make human life easy by using audio classifications, but it is expensive and difficult to deploy such models at edge devices like nano BLE sense raspberry pi, because these models require huge computational power like graphics processing unit (G.P.U), to solve the problem, we proposed DL model. In our proposed work, we had gone for a low complexity model for Audio Event Detection (AED), we extracted Mel-spectrograms of dimension 128×431×1 from audio signals and applied normalization. A total of 3 data augmentation methods were applied as follows: frequency masking, time masking, and mixup. In addition, we designed Convolutional Neural Network (CNN) with spatial dropout, batch normalization, and separable 2D inspired by VGGnet [1]. In addition, we reduced the model size by using model quantization of float16 to the trained model. Experiments were conducted on the updated dataset provided by the Detection and Classification of Acoustic Events and Scenes (DCASE) 2020 challenge. We confirm that our model achieved a val_loss of 0.33 and an accuracy of 90.34% within the 132.50KB model size.

• Journal of the Korean Wood Science and Technology
• /
• 제34권2호
• /
• pp.30-36
• /
• 2006

Wood fuel must be dried before combustion to minimize the energy loss. Sawdust of Japanese red pine was dried in a batch type fluidized-bed to investigate the drying characteristics of sawdust as a raw material for bio-fuel. The minimum fluidization air velocity was increased as particle size was increased. It took about 21 minutes and 8 minutes to dry 0.08 m-deep bed of particles with average particle size of 1.3 mm from 100% to 10% moisture content at air temperature of $20^{\circ}C$ and $50^{\circ}C$, respectively.

• 열처리공학회지
• /
• 제15권1호
• /
• pp.16-20
• /
• 2002

In conventional metallothermic reduction(MR) for obtaining tantalum powder in batch-type operation, it is difficult to control morphology and location of deposits because the reaction occurs by direct physical contact between reductants and feed materials. On the other hand, a electronically mediated reaction(EMR) is capable to overcome these difficulties through the reaction by electron transfer and have a merit of continuous process. In this study an MR and EMR method has been applied to the production of a tantalum powder by sodium reduction of $K_2TaF_7$. As the reduction temperature increases, the particle size and yield of tantalum powder obtained by MR and EMR method is increased.

• 상하수도학회지
• /
• 제18권5호
• /
• pp.689-695
• /
• 2004

Electrocoagulation has been suggested as a promising alternative to conventional coagulation. The process is characterized by reduced sludge production, no requirement for chemical use, and ease of operation. However, this coagulation has scarcely been studied in water purifying process. This study was performed several batch experiments to compare turbidity removal between electrocoagulation and chemical coagulation. In addition, characteristics of floe were evaluated with zeta potential and particle size distributions. Electrocoagulation showed a relatively higher removal of turbidity (approximately 5%) with the same aluminum amount than conventional chemical coagulation. In addition, turbidity removal by electrocoagulation was less sensitive to pH and was greater for more extensive pH range than chemical coagulation. The results of zeta potential and floc size distributions illustrated that electrocoagulation provided the preferable conditions for coagulation such as zeta potential close to zero millivolt and increased portions of particles in the range of 40 and $100{\mu}m$.