Higher-order Chebyshev Rational Approximation Method (CRAM)

Maria Pusa

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

Abstract

The burnup equations can in principle be solved by computing the exponential of the burnup matrix. However, due to the difficult numerical characteristics of burnup matrices, the problem is extremely stiff, and the matrix exponential solution was long considered infeasible for an entire burnup system containing over a thousand nuclides. After discovering that the eigenvalues of burnup matrices are generally confined to a region near the negative real axis, the Chebyshev rational approximation method (CRAM) was introduced as a novel method to solve the burnup equations. It can be characterized as the best rational function on the negative real axis and it has been shown to be capable of simultaneously solving an entire burnup system both accurately and efficiently. The main difficulty in using CRAM for computing the matrix exponential is determining the coefficients of the rational function for a given approximation order. Some polynomial CRAM coefficients have been published in 1984, and based on these literature values, CRAM approximations up to the order 16 have been thus far applied in burnup calculations. The topic of this paper is the computation of CRAM approximations and their application to burnup equations. A Remez-type method utilizing the equioscillation property of best approximations is used to construct the CRAM approximants for approximation orders 1, ..., 50. Numerical results are presented for a large burnup system and for a decay system. It is demonstrated that higher-order CRAM can be used to accurately solve the burnup equations even with time steps of the order of millions of years.
Original languageEnglish
Title of host publicationProceedings of the international conference on physics of reactors
Subtitle of host publicationPHYSOR 2014
Number of pages11
VolumeJAEA-Conf 2014-003
DOIs
Publication statusPublished - 2015
MoE publication typeA4 Article in a conference publication
EventInternational Conference on the Physics of Reactors, PHYSOR 2014: The Role of Reactor Physics toward Sustainable Future - Kyoto, Japan
Duration: 28 Sep 20143 Oct 2014

Conference

ConferenceInternational Conference on the Physics of Reactors, PHYSOR 2014
Abbreviated titlePHYSOR2014
CountryJapan
CityKyoto
Period28/09/143/10/14

Fingerprint

Rational functions
Isotopes
Polynomials

Cite this

Pusa, M. (2015). Higher-order Chebyshev Rational Approximation Method (CRAM). In Proceedings of the international conference on physics of reactors: PHYSOR 2014 (Vol. JAEA-Conf 2014-003). [1119422] https://doi.org/10.11484/jaea-conf-2014-003
Pusa, Maria. / Higher-order Chebyshev Rational Approximation Method (CRAM). Proceedings of the international conference on physics of reactors: PHYSOR 2014. Vol. JAEA-Conf 2014-003 2015.
@inproceedings{54fa452b755640359b61504df4d66f1e,
title = "Higher-order Chebyshev Rational Approximation Method (CRAM)",
abstract = "The burnup equations can in principle be solved by computing the exponential of the burnup matrix. However, due to the difficult numerical characteristics of burnup matrices, the problem is extremely stiff, and the matrix exponential solution was long considered infeasible for an entire burnup system containing over a thousand nuclides. After discovering that the eigenvalues of burnup matrices are generally confined to a region near the negative real axis, the Chebyshev rational approximation method (CRAM) was introduced as a novel method to solve the burnup equations. It can be characterized as the best rational function on the negative real axis and it has been shown to be capable of simultaneously solving an entire burnup system both accurately and efficiently. The main difficulty in using CRAM for computing the matrix exponential is determining the coefficients of the rational function for a given approximation order. Some polynomial CRAM coefficients have been published in 1984, and based on these literature values, CRAM approximations up to the order 16 have been thus far applied in burnup calculations. The topic of this paper is the computation of CRAM approximations and their application to burnup equations. A Remez-type method utilizing the equioscillation property of best approximations is used to construct the CRAM approximants for approximation orders 1, ..., 50. Numerical results are presented for a large burnup system and for a decay system. It is demonstrated that higher-order CRAM can be used to accurately solve the burnup equations even with time steps of the order of millions of years.",
author = "Maria Pusa",
note = "Papers in zip file.",
year = "2015",
doi = "10.11484/jaea-conf-2014-003",
language = "English",
volume = "JAEA-Conf 2014-003",
booktitle = "Proceedings of the international conference on physics of reactors",

}

Pusa, M 2015, Higher-order Chebyshev Rational Approximation Method (CRAM). in Proceedings of the international conference on physics of reactors: PHYSOR 2014. vol. JAEA-Conf 2014-003, 1119422, International Conference on the Physics of Reactors, PHYSOR 2014, Kyoto, Japan, 28/09/14. https://doi.org/10.11484/jaea-conf-2014-003

Higher-order Chebyshev Rational Approximation Method (CRAM). / Pusa, Maria.

Proceedings of the international conference on physics of reactors: PHYSOR 2014. Vol. JAEA-Conf 2014-003 2015. 1119422.

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

TY - GEN

T1 - Higher-order Chebyshev Rational Approximation Method (CRAM)

AU - Pusa, Maria

N1 - Papers in zip file.

PY - 2015

Y1 - 2015

N2 - The burnup equations can in principle be solved by computing the exponential of the burnup matrix. However, due to the difficult numerical characteristics of burnup matrices, the problem is extremely stiff, and the matrix exponential solution was long considered infeasible for an entire burnup system containing over a thousand nuclides. After discovering that the eigenvalues of burnup matrices are generally confined to a region near the negative real axis, the Chebyshev rational approximation method (CRAM) was introduced as a novel method to solve the burnup equations. It can be characterized as the best rational function on the negative real axis and it has been shown to be capable of simultaneously solving an entire burnup system both accurately and efficiently. The main difficulty in using CRAM for computing the matrix exponential is determining the coefficients of the rational function for a given approximation order. Some polynomial CRAM coefficients have been published in 1984, and based on these literature values, CRAM approximations up to the order 16 have been thus far applied in burnup calculations. The topic of this paper is the computation of CRAM approximations and their application to burnup equations. A Remez-type method utilizing the equioscillation property of best approximations is used to construct the CRAM approximants for approximation orders 1, ..., 50. Numerical results are presented for a large burnup system and for a decay system. It is demonstrated that higher-order CRAM can be used to accurately solve the burnup equations even with time steps of the order of millions of years.

AB - The burnup equations can in principle be solved by computing the exponential of the burnup matrix. However, due to the difficult numerical characteristics of burnup matrices, the problem is extremely stiff, and the matrix exponential solution was long considered infeasible for an entire burnup system containing over a thousand nuclides. After discovering that the eigenvalues of burnup matrices are generally confined to a region near the negative real axis, the Chebyshev rational approximation method (CRAM) was introduced as a novel method to solve the burnup equations. It can be characterized as the best rational function on the negative real axis and it has been shown to be capable of simultaneously solving an entire burnup system both accurately and efficiently. The main difficulty in using CRAM for computing the matrix exponential is determining the coefficients of the rational function for a given approximation order. Some polynomial CRAM coefficients have been published in 1984, and based on these literature values, CRAM approximations up to the order 16 have been thus far applied in burnup calculations. The topic of this paper is the computation of CRAM approximations and their application to burnup equations. A Remez-type method utilizing the equioscillation property of best approximations is used to construct the CRAM approximants for approximation orders 1, ..., 50. Numerical results are presented for a large burnup system and for a decay system. It is demonstrated that higher-order CRAM can be used to accurately solve the burnup equations even with time steps of the order of millions of years.

U2 - 10.11484/jaea-conf-2014-003

DO - 10.11484/jaea-conf-2014-003

M3 - Conference article in proceedings

VL - JAEA-Conf 2014-003

BT - Proceedings of the international conference on physics of reactors

ER -

Pusa M. Higher-order Chebyshev Rational Approximation Method (CRAM). In Proceedings of the international conference on physics of reactors: PHYSOR 2014. Vol. JAEA-Conf 2014-003. 2015. 1119422 https://doi.org/10.11484/jaea-conf-2014-003