• 한국연소학회지
• /
• 제20권4호
• /
• pp.10-18
• /
• 2015

Combustion instability is a major issue in design and maintenance of gas turbine combustors for efficient operation with low emissions. With the thermoacoustic view point the instability is induced by the interaction of the unsteady heat release of the combustion process and the change in the acoustic pressure in the combustion chamber. In an effort to study the combustion dynamics of gas turbine combustors, Morgans et al (2014) have developed OSCILOS (open source combustion instability low order simulator) code and it is currently available online. In this study the code has been utilized to predict the combustion instability of a reported case for lean premixed gas turbine combustion, and then its prediction results have been compared with the corresponding experimental data. It turned out that both the predicted and the experimental combustion instability results agree well. Further the effects of some typical inlet acoustic boundary conditions on the prediction have been investigated briefly. It is believed that the validity and effectiveness of the open source code is reconfirmed through this benchmark test.

• 유통과학연구
• /
• 제14권9호
• /
• pp.25-29
• /
• 2016

Purpose - Finding top K persuaders in consumer network is an important problem in marketing. Recently, a new method of computing persuasion scores, interpreted as fixed point or stable distribution for given persuasion probabilities, was proposed. Top K persuaders are chosen according to the computed scores. This research proposed a new definition of persuasion scores relaxing some conditions on the matrix of probabilities, and a method to identify top K persuaders based on the defined scores. Research design, data, and methodology - A new method of computing top K persuaders is computed by singular value decomposition (SVD) of the matrix which represents persuasion probabilities between entities. Results - By testing a randomly generated instance, it turns out that the proposed method is essentially different from the previous study sharing a similar idea. Conclusions - The proposed method is shown to be valid with respect to both theoretical analysis and empirical test. However, this method is limited to the category of persuasion scores relying on the matrix-form of persuasion probabilities. In addition, the strength of the method should be evaluated via additional experiments, e.g., using real instances, different benchmark methods, efficient numerical methods for SVD, and other decomposition methods such as NMF.

• Spatial Information Research
• /
• 제6권2호
• /
• pp.125-132
• /
• 1998

지리정보시스템을 위한 공간 데이터베이스는 공간 데이터의 특성에 의해 다차원의 대용량 데이터를 다루기 때문에 공간 질의의 I/O성능이 매우 중요한 역할을 한다. 따라서 본 논문에서는 공간 질의의 I/O성능을 높이기 위해서 복잡한 공간 객체들을 다루는 대표적인 접근기법들인 Z-변형을 이용한 B 트리, KDB트리, R트리, MAX트리에 대해 기술하였다. 또한, 다양한 실제 데이터와 질의 집합을 사용해서 여러 공간 색인 기법들의 성능을 측정하였다. 벤치마크 실험을 해본 결과, MAX 트리는 삽입, 영역 질의, 공간 조인 둥의 연산에 대해 다른 색인 기법들보다 상대적으로 좋은 성능을 나타냈다. MAX 트리는 향후 GIS의 하부 저장시스템을 구성하는 색인기법으로 사용될 것으로 기대된다.

• Nuclear Engineering and Technology
• /
• 제24권3호
• /
• pp.252-262
• /
• 1992

시간 및 공간 종속형 중성자 수송 방정식으로부터 비균질 원자로 노심해석의 효율적인 방법을 개발하였다. 이를 위해 계산 시간을 단축하고 각 집합체 크기의 소격격자(coarse mesh)에 대한 평균 중성자속을 정확히 예측할 수 있도록 노달방법(nodal method)을 도입하였고, 노드 별 평균 중성자속과 노드 각 경계면의 평균 중성자속 및 유속(flux and current)과의 관계식을 얻기 위하여 조정 인자( correction factor)로서 불연속인자(discontinuity factor)를 사용하였으며, 이 인자는 이전 시간대(previous time step)의 노드 평균 중성자속, 확산계수, 그리고 불연속인자 등에 따라 새로이 계산(updating)된다. 본 논문에서 개발된 방법을 시간에 따라 비교적 단순히 변하는 과도 노심(TWIGL)과 급격한 중성자 거동의 변화를 모사하는 과도 노심(LRA)에 적용한 결과 정확성 및 효율성이 입증되었 다.

• 대한전기학회논문지:전력기술부문A
• /
• 제52권11호
• /
• pp.632-638
• /
• 2003

Modern power systems are very complex and to model them completely is impractical for electromagnetic transient studies. Therefore areas outside the immediate area of interest must be represented by some form of frequency dependent equivalent. The s-domain rational function form of frequency dependent equivalent does not need refitting if the simulation time-step is changed in the electromagnetic transient program. This is because the s-domain rational function coefficients are independent of the simulation time-step, unlike the z-domain rational function coefficients. S-domain rational function fitting techniques for representing frequency dependent equivalents have been developed using Least Squares Fitting(LSF). However it does not suffer the implementation error that exited in this work as it ignored the instantaneous term. This paper Presents the formulation for developing 1 Port Frequency Dependent Network Equivalent(FDNE) with the instantaneous term in S-domain and illustrates its use. This 1 port FDNE have been applied to the CIGRE Benchmark Rectifier test AC system. The electromagnetic transient package PSCAD/EMTDC is used to assess the transient response of the 1 port (FDNE) developed with Thevenin and Norton Equivalent network. The study results have indicated the robustness and accuracy of 1 port FDNE for electromagnetic transient studies.

• Spatial Information Research
• /
• 제7권2호
• /
• pp.271-282
• /
• 1999

본 논문에서느 다중-속성 데이터와 공간 조인 연산을 효율적으로 수행하는 색인기법인 SJ(Spatial Join) 트리를 제안한다. 또한, 다중-속성 데이터를 다루기 위한 기존의 다양한 알고리즘들을 계산 복잡도와 I/O 연산의 복잡도와 함께 설명한다. 우리는 이 논문을 통해서 제안된 SJ 트리가 기존의 데이터베이스 시스템에서 색인 기법으로 많이 사용되는 B-트리를 일반화한 것이라는 것을 보여준다. 이것은 SJ 트리가 기존의 대부분의 B-트리를 이용하는 저장구조에 쉽게 구현될 수 있다는 것을 의미한다. 공간 출력을 갖는 공간 조인 연산은 R-트리, B-트리, K-D-B 트리, SJ 트리에 대해서 성능평가를 수행한다. 성능평가 결과 제안된 SJ 트리가 점 데이터를 갖는 공간 조인 연산에 대해서 다른 색인 기법들보다 상대적으로 우수한 결과를 보여준다.

• 전자공학회논문지C
• /
• 제35C권3호
• /
• pp.79-89
• /
• 1998

The Error Back-Propagation(EBP) algorithm is widely applied to train a multi-layer perceptron, which is a neural network model frequently used to solve complex problems such as pattern recognition, adaptive control, and global optimization. However, the EBP is basically a gradient descent method, which may get stuck in a local minimum, leading to failure in finding the globally optimal solution. Moreover, a multi-layer perceptron suffers from locking a systematic determination of the network structure appropriate for a given problem. It is usually the case to determine the number of hidden nodes by trial and error. In this paper, we propose a new algorithm to efficiently train a multi-layer perceptron. OUr algorithm uses stochastic perturbation in the weight space to effectively escape from local minima in multi-layer perceptron learning. Stochastic perturbation probabilistically re-initializes weights associated with hidden nodes to escape a local minimum if the probabilistically re-initializes weights associated with hidden nodes to escape a local minimum if the EGP learning gets stuck to it. Addition of new hidden nodes also can be viewed asa special case of stochastic perturbation. Using stochastic perturbation we can solve the local minima problem and the network structure design in a unified way. The results of our experiments with several benchmark test problems including theparity problem, the two-spirals problem, andthe credit-screening data show that our algorithm is very efficient.

• Structural Engineering and Mechanics
• /
• 제17권1호
• /
• pp.113-140
• /
• 2004

A mixed eight-node hexahedral element formulated via the Hu-Washizu principle as well as the field extrapolation technique is presented. The mixed element with only three translational degrees of freedom at each node can provide extremely accurate and reliable performance for popular benchmark problems such as spacial beams, plates, shells as well as general three-dimensional elasticity problems. Numerical calculations also show that when extremely skewed and coarse meshes and nearly incompressible materials are used, the proposed mixed element can still possess excellent behaviour. The mixed formulation starts with introduction of a parallelepiped domain associated with the given general eight-node hexahedral element. Then, the assumed strain field at the nodal level is constructed via the Hu-Washizu variational principle for that associated parallelepiped domain. Finally, the assumed strain field at the nodal level of the given hexahedral element is established by using the field extrapolation technique, and then by using the trilinear shape functions the assumed strain field of the whole element domain is obtained. All matrices involved in establishing the element stiffness matrix can be evaluated analytically and expressed explicitly; however, a 24 by 24 matrix has to be inverted to construct the displacement extrapolation matrix. The proposed hexahedral element satisfies the patch test as long as the element with a shape of parallelepiped.

• Nuclear Engineering and Technology
• /
• 제48권1호
• /
• pp.33-42
• /
• 2016

Spacer grids play an important role in maintaining the proper form of the fuel assembly structure and ensuring the safety of reactor core design. This study applies the Monte Carlo method to the analysis of the neutronics effects of spacer grids in a typical pressurized water reactor (PWR). The core problem used to analyze the neutronics effects of spacer grids is a modified version of Korea Advanced Institute of Science and Technology benchmark problem 1B, based on an Advanced Power Reactor 1400 (APR1400) core model. The spacer grids are modeled and added to this test problem in various ways. Then, by running MCNP5 for all cases of spacer grid modeling, some important numerical results, such as the effective multiplication factor, the spatial distributions of neutron flux, and its energy spectrum are obtained. The numerical results of each case of spacer grid modeling are analyzed and compared to assess which type has more advantages in accuracy of numerical results and effectiveness in terms of geometry building. The conclusion is that the most realistic modeling for Monte Carlo calculation is the "volume-preserving" streamlined heterogeneous spacer grids, but the "banded" dissolution spacer grids modeling is a more practical yet accurate model for routine (deterministic) analysis.

• Nuclear Engineering and Technology
• /
• 제49권4호
• /
• pp.675-691
• /
• 2017

Jets in cross flow are of fundamental industrial importance and play an important role in validating turbulence models. Two jet configurations related to thermal fatigue phenomena are investigated: ${\bullet}$ T-junction of circular tubes where a heated jet discharges into a cold main flow and ${\bullet}$ Rectangular jet marked by a scalar discharging into a main flow in a rectangular channel. The T-junction configuration is a classical test case for thermal fatigue phenomena. The Vattenfall T-junction experiment was already subject of an OECD/NEA benchmark. A LES modelling and calculation strategy is developed and validated on this data. The rectangular-jet configuration is important for basic physical understanding and modelling and has been analyzed experimentally at CEA. The experimental work was focused on turbulent mixing between a slightly heated rectangular jet which is injected perpendicularly into the cold main flow of a rectangular channel. These experiments are analyzed for the first time with LES. The overall results show a good agreement between the experimental data and the CFD calculation. Mean values of velocity and temperature are well captured by both RANS calculation and LES. The range of critical frequencies and their amplitudes, however, are only captured by LES.