Internal optimization of the texture component approximation method

17th International Conference on Textures of Materials

材料科学;物理学

Nikolayev, D.^1 ; Lychagina, T.^1 ; Rusetsky, M.^2 ; Ulyanenkov, A.^3 ; Sasaki, A.^4

Frank Laboratory of Neutron Physics, Joint Institute for Nuclear Research, Dubna, Russia^1

Belarusian State University, Minsk, Belarus^2

Rigaku Europe SE, Ettlingen, Germany^3

Rigaku Corporation, Tokyo, Japan^4

关键词: Approximation methods;    Bell-shaped functions;    Mean square deviation;    Minimization procedures;    Non-linear optimization problems;    Orientation distribution function;    Quantitative measures;    Weighted linear combinations;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/82/1/012007/pdf
DOI  :  10.1088/1757-899X/82/1/012007

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

The component approximation method for the reconstruction of orientation distribution function (ODF) is based on the assumption that the texture could be presented as a weighted linear combination of distributions depending on the parameters, which are related to the position of bell shaped function in orientation space and to the dispersion. The method uses a minimization procedure to obtain the values of ODF parameters. Traditionally, the mean- square deviation of the measured and recalculated pole figures is minimized. However, the quantitative measure of the fit is RP value which differs from the mean-square deviation. In the present work it is suggested to minimize the RP value to obtain ODF parameters. We are using Trust Region method for solving a non-linear optimization problem. The convergences of the proposed method for different minimized functional are compared. We also illustrate a usage of the different objective function on modeling data for the cubic crystalline symmetry. This study is fulfilled using new RIGAKU software for quantitative texture analysis.