• 화약ㆍ발파
• /
• 제13권3호
• /
• pp.19-30
• /
• 1995

NATM as method of tunneling has been applied to construction of domestic subway, roads, rail way, water way etc. Accordingly, we have NATM's many drafts and constructional results, but many problems and accidents have occurred under construction of tunnel using NATM for shortage of technical data. Poorness of constructional improvement, systemic inconsistency etc. Especially, everyone was shocked at Gupo's train wrecking accident lately. The purpose of this thesis is presentation of means for setting technical problems, by looking into Gupo's train wrecking accident and home records that applying NATM in tunneling failed, to minimize future safety accidents we find that the general problems of home fifteen sites having occured accidents is badly geological survey, nonconfirmationi of base rock's state, formal measuring management, shortage of specialists, systemical discrepancy and that disregarding NATM's rules makes general problems. The results of this study are Summarised as follows : 1. We advise repletion of design standards to practice crosshole test for confirming connected rock base on vertical section of tunnel. 2. We advise to practice pre0boring and pre-grouting for a weak layer difficult in applying NATM. 3. We advise systemic improvements that field servicer can construct tunnel of his own free will considering base rock's state at tunnel . 4. We advise that specialist, who can make a conduct and supervise above mentioned items as well as measuring management, should be posted at field.

• 한국안전학회지
• /
• 제9권4호
• /
• pp.92-102
• /
• 1994

NATM as method of tunneling has been applied to construction of domestic subway, roads, rail way, water way etc. Accordingly we have NATM's many drafts and constructional results, but many problems and accidents have occurred under construction of tunnel using NATM for shortage of technical data, poorness of constructional improvement, systemic inconsistency etc. Especially, everyone was shocked at Gupo's train wrecking accident lately. The purpose of this thesis is presentation of means for settling technical problems, by looking into Gupo's train wrecking accident and home records that applying NATM in tunneling failed, to minimize future safety accidents we find that the general problems of home fifteen sites haying occured accidents is badly geological survey, nonconfirmation of base rock's state, formal measuring management, shortage of specialists, systemical discrepancy and that disregarding NATM's rules makes general problems. The results of this study are summarised as follows ; 1. We advise repletion of design standards to practice crosshole test for confirming connected rock base on vertical section of tunnel. 2. We advise to practice pre-boring and pre-grouting for a weak layer difficult in applying NATM. 3. We advise systemic improvements that field servicer can construct tunnel of his own free will considering base rock's state at tunnel. 4. We advise that specialist, who can make a conduct and supervise above mentioned items as well as measuring managements, should be posted at field.

• 대한교통학회지
• /
• 제25권5호
• /
• pp.91-100
• /
• 2007

본 연구는 KTX의 고속운행에 의한 통행시간단축과 일반열차(새마을, 무궁화)의 다수 정차에 의한 접근 편의성 등 복합 장점을 지닌 직결운행이란 새로운 대안을 제시하고 도입 타당성을 검토하고자 하였으며, 이에 경부선의 과거 철도 수송수요 Data를 근거로 KTX 개통전 후 수요변화 및 경부선 2단계(동대구${\sim}$부산) 개통에 의한 유발수요 등을 고려하여 예측하였다. 분석 결과 직결운행 도입(본 연구 2011년 가정)전 후 이용자 구간통행시간은 일반열차 이용시에 비해 $70{\sim}100$분 이상 절감될 것으로 나타났으며, 1일 열차 총 수송수요의 증가로 인해 운영수입은 1억 3천여만원/일 증가할 것으로 예측되었다. 또한 새마을(디젤차량)을 KTX(전기차량)로 대체함에 따라 연간 이산화탄소 배출량이 $20{\sim}50%$수준으로 감소되어 보다 환경 친화적 효과가 클 것으로 기대된다. 단, 직결운행 수요를 위한 고가의 차량구입비용(본 연구에서는 KTXⅡ 도입 가정, 1조 3천여억원)과 연간 운영비용을 감안해 볼 때 운영자입장에서의 수익/비용분석결과는 매우 낮은 것으로(PI=0.54) 나타났다. 따라서, 보다 정확한 직결운행 도입 타당성을 객관적으로 검토하기 위해서는 차량구입 재원확보 방안 및 운행효율 극대화를 위한 제약조건 개선 등이 수반되어야 할 것이다.

• 한국IT서비스학회지
• /
• 제19권5호
• /
• pp.83-91
• /
• 2020

Language Models such as BERT has been an important factor of deep learning-based natural language processing. Pre-training the transformer-based language models would be computationally expensive since they are consist of deep and broad architecture and layers using an attention mechanism and also require huge amount of data to train. Hence, it became mandatory to do fine-tuning large pre-trained language models which are trained by Google or some companies can afford the resources and cost. There are various techniques for fine tuning the language models and this paper examines three techniques, which are data augmentation, tuning the hyper paramters and partly re-constructing the neural networks. For data augmentation, we use no-answer augmentation and back-translation method. Also, some useful combinations of hyper parameters are observed by conducting a number of experiments. Finally, we have GRU, LSTM networks to boost our model performance with adding those networks to BERT pre-trained model. We do fine-tuning the pre-trained korean-based language model through the methods mentioned above and push the F1 score from baseline up to 89.66. Moreover, some failure attempts give us important lessons and tell us the further direction in a good way.

• Industrial Engineering and Management Systems
• /
• 제12권3호
• /
• pp.281-287
• /
• 2013

In this paper, we introduce a roots method that uses the roots inside the unit circle of the associated characteristics equation to evaluate the steady-state system-length distribution at three epochs (pre-arrival, arbitrary, and post-departure) and sojourn-time distribution in GI/PH/1 queueing model. It is very important for an air base to inspect airplane oil because low-quality oil leads to drop or breakdown of an airplane. Since airplane oil inspection is composed of several inspection steps, it sometimes causes train congestion and delay of inventory replenishments. We analyzed interarrival time and inspection (service) time of oil supply from the actual data which is given from one of the ROKAF's (Republic of Korea Air Force) bases. We found that interarrival time of oil follows a normal distribution with a small deviation, and the service time follows phase-type distribution, which was first introduced by Neuts to deal with the shortfalls of exponential distributions. Finally, we applied the GI/PH/1 queueing model to the oil train congestion problem and analyzed the distributions of the number of customers (oil trains) in the queue and their mean sojourn-time using the roots method suggested by Chaudhry for the model GI/C-MSP/1.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.2198-2202
• /
• 2003

The data transmission of MVB(Multifunction Vehicle Bus) of TCN(Train Network Communication) is divided into the periodic transmission phase and the sporadic transmission phase. TCN standard defines the event-polling method for the message transfer in the sporadic phase. However, since the event-polling method does not use pre-scheduling to the priority of the messages to be transmitted, it is inefficient for the real-time systems. To schedule message transmission, a master node should know the priority of message to be transmitted by a slave node prior to the scheduling the sporadic phase, but the existing TCN standard does not support any protocol for this. This paper proposes the slave frame bit-stuffing algorithm, with which a master node gets the necessary information for transmission scheduling and includes the simulation results of the event-polling method and the proposed algorithm.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2010년도 춘계학술대회 논문집
• /
• pp.165-170
• /
• 2010

Recently, Interior Permanent Magnet Machine(IPM) is widely used for traction motor in the high speed train. Higher efficiency and power density are the superb performance of IPM. Due to the high power density, however, it has lots of heat source which are originated from copper losses and core losses. These heat source can cause the permanent demagnetization in magnet and the loss of torque and power. To prevent the undesirable loss in the traction motor, the accurate loss calculation and the thermal analysis should be preceded. Especially, the end-winding area and permanent magnet area should be examined correctly. In this paper, the electromagnetic fields were examined by finite element method to analyze the electromagnetic properties of IPM and thermal analysis are carried out with pre-calculated losses. To validate the analysis result, the experiment set with forced air cooling system is manufactured.

• 전기학회논문지
• /
• 제63권11호
• /
• pp.1602-1607
• /
• 2014

This paper deals with the design and property analysis of 7kW-class small-scaled superconducting Linear Synchronous Motor (LSM) and testing equipment for a number of performance pre-tests prior to the development of coreless-type superconducting LSM suitable for 600km/h very high speed train. First, the basic design and property analysis are conducted before developing a small-scaled superconducting LSM model with 2-pole superconducting electromagnets, and additionally the cost-down design of the superconducting electromagnets is conducted to use less high-Tc superconducting wire. Finally, the superconducting magnet coil span is selected at 120mm, and input ground armature current of 670Aturns is required to produce 44.7N of thrust based on research findings.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2004년도 추계학술대회 논문집
• /
• pp.634-639
• /
• 2004

Under light collision accidents, the energy absorption strategy for the coupler and expansion tube of the TTX(Tilting Train Express) initial design is established in the paper. Also, 1st shearing bolts are designed. When the absorbed energy of the coupler reaches its maximum, the connecting bolts between the coupler and the car body are sheared off not to transmit the impact force to the car body structure. To absorb more energy after the lst shearing bolts work, a expansion tube is designed conceptually and installed at the rear part of the coupler. Using Hyper-Mesh and LS-DYNA, pre/post processing and light collision analyses are preformed, respectively.

• Journal of Information Processing Systems
• /
• 제18권4호
• /
• pp.524-534
• /
• 2022

In this paper, we produce Trash Object Detection (TOD) dataset to solve trash detection problems. A well-organized dataset of sufficient size is essential to train object detection models and apply them to specific tasks. However, existing trash datasets have only a few hundred images, which are not sufficient to train deep neural networks. Most datasets are classification datasets that simply classify categories without location information. In addition, existing datasets differ from the actual guidelines for separating and discharging recyclables because the category definition is primarily the shape of the object. To address these issues, we build and experiment with trash datasets larger than conventional trash datasets and have more than twice the resolution. It was intended for general household goods. And annotated based on guidelines for separating and discharging recyclables from the Ministry of Environment. Our dataset has 10 categories, and around 33K objects were annotated for around 5K images with 1280×720 resolution. The dataset, as well as the pre-trained models, have been released at https://github.com/jms0923/tod.