View : 1087 Download: 0

Full metadata record

DC Field Value Language
dc.contributor.advisor윤정호-
dc.contributor.author양효선-
dc.creator양효선-
dc.date.accessioned2017-08-27T12:08:29Z-
dc.date.available2017-08-27T12:08:29Z-
dc.date.issued2017-
dc.identifier.otherOAK-000000142304-
dc.identifier.urihttp://dcollection.ewha.ac.kr/jsp/common/DcLoOrgPer.jsp?sItemId=000000142304en_US
dc.identifier.urihttps://dspace.ewha.ac.kr/handle/2015.oak/236520-
dc.description.abstractThis dissertation aims at developing non-linear approximation methods for piecewise smooth data based on exponential polynomials and applying them for numerical solutions of some partial differential equations, especially hyperbolic conservation laws. Given a discrete set of data sampled from a piecewise smooth function, the main challenge is to develop approximation methods capable of obtaining high approximation order in smooth areas and capturing non-smooth solutions without generating artifacts such as oscillations. To this end, this study introduces a novel high order weighted essentially non-oscillatory (WENO) finite difference scheme. First, the WENO interpolations are constructed by using exponential (or trigonometric) polynomials instead of using algebraic polynomials. Especially a tension parameter introduced in exponential basis function can be regulated to the local features of given data, which yields better approximation near steep gradients without spurious oscillations, compared to the WENO schemes based on algebraic polynomials at lower computational cost. Second, to design new WENO weights, we present two important measurements: a discontinuity detector (at the cell boundary) and a smoothness indicator. A detailed analysis is performed to verify that proposed scheme provides the required convergence order of accuracy. Finally, we suggest to combine the WENO scheme with adaptive mesh strategy: a moving mesh technique which is developed by a new monitor function. Some numerical experiments are presented and compared with other WENO schemes to demonstrate the new algorithm's ability.;이 논문은 지수 함수를 기반으로 하여 구간적으로 매끄러운 데이터를 대상으로 하는 비선형 근사 기법을 개발하고, 특히 Conservation Laws와 같은 일부 편미분 방정식의 수치 해법에 적용하도록 하는 데에 그 목적을 둔다. 이를 위해 본 연구에서는 고차 Weighted Essentially Non-Oscillatory (WENO) 유한 차분 기법을 소개하고, 다음의 세 가지 방향으로 발전시킨다. 첫째, 다항함수를 기반으로 하던 기존의 WENO 방법과 다르게, 주어진 데이터의 특성에 맞춰질 수 있도록 매개 변수를 갖는 지수 함수를 이용하여 보간법을 구현한다. 이는 기존의 방법들과 비교하여 낮은 계산 비용으로 급격한 기울기 근처에서 수치해의 소실 또는 진동이 없는 정밀한 근사 결과를 산출한다. 둘째, 셀 경계에서 불연속점을 판별하는 인자(Discontinuity detector)와 국지적 매끄러운 정도를 측정하는 인자(Smoothness indicator)를 구성하고, 이 두 가지 측정치를 통해 새로운 WENO 가중치를 제시한다. 셋째, 균등한 그리드에서 정의되었던 WENO 방법에 국지적 데이터의 특성에 따라 메쉬를 움직이는 기술(Moving mesh)을 결합하여 효율적인 수치 방법을 구성해 낸다. 위의 각 단계에서 제안된 기법이 최적의 근사 정확도를 제공하는지를 검증하기 위해 상세한 분석 및 증명을 수행하고, 마지막으로 다양한 수치적 실험을 통해 이 학위 논문에서 제안한 새로운 알고리즘의 우수성을 확인한다.-
dc.description.tableofcontents1 Introduction 1 1.1 Approximation by non-polynomial spaces 4 1.2 Adaptive mesh strategy 6 1.3 Outline of the proposed scheme 7 2 Review of WENO schèmes 11 2.1 Flux and numerical flux 12 2.2 Fifth-order WENO scheme 15 2.3 Sixth-order WENO scheme 17 3 WENO scheme based on exponential polynomials 20 3.1 Space of exponential polynomials 21 3.2 Numerical flux based on using exponential functions 25 3.2.1 Construction of numerical flux 25 3.2.2 Approximation order of numerical flux 32 3.3 Formulations for WENO with exponential basis 37 3.3.1 Optimal weights based on exponential polynomials 38 4 Construction of WENO-NE method 42 4.1 WENO schemes with new non-linear weights 42 4.1.1 Generalized undivided-differences 43 4.1.2 Discontinuity detector 46 4.1.3 Non-linear weights 48 4.2 Convergence order of WENO-NE 50 4.2.1 Convergence order in smooth region 50 4.2.2 Convergence order at critical points 52 4.3 Numerical results 53 4.3.1 Scalar test problems 54 4.3.2 One-dimensional Euler systems 57 4.3.3 Two-dimensional Euler systems 66 5 WENO scheme with moving mesh technique 73 5.1 A moving mesh WENO scheme 73 5.1.1 WENO for transformed conservation laws 73 5.1.2 De Boor's algorithm 75 5.2 Construction of a new moving mesh WENO method 76 5.2.1 Monitor functions 76 5.2.2 The moving mesh WENO algorithm 79 5.3 Numerical results 82 5.3.1 Scalar test problems 83 5.3.2 Hyperbolic systems 90 5.3.3 Weakly hyperbolic systems 96 6 Concluding remarks 101 References 102 Appendix: Optimal Tension Parameter 111 Korean Abstract 115-
dc.formatapplication/pdf-
dc.format.extent6201243 bytes-
dc.languageeng-
dc.publisher이화여자대학교 대학원-
dc.subject.ddc500-
dc.titleDevelopment of Weighted Essentially Non-Oscillatory Schemes Based on Exponential Polynomials-
dc.typeDoctoral Thesis-
dc.title.translated지수 함수를 기반으로 한 WENO 수치 방법 구성-
dc.format.pagevi, 115 p.-
dc.contributor.examiner김창호-
dc.contributor.examiner민조홍-
dc.contributor.examiner윤정호-
dc.contributor.examiner이준엽-
dc.contributor.examiner하영수-
dc.identifier.thesisdegreeDoctor-
dc.identifier.major대학원 수학과-
dc.date.awarded2017. 8-
Appears in Collections:
일반대학원 > 수학과 > Theses_Ph.D
Files in This Item:
There are no files associated with this item.
Export
RIS (EndNote)
XLS (Excel)
XML


qrcode

BROWSE