﻿ X射线粉末衍射仪智能辅助校正系统
 计算机系统应用  2018, Vol. 27 Issue (8): 75-80 PDF
X射线粉末衍射仪智能辅助校正系统

1. 中国科学院大学, 北京 100049;
2. 中国科学院 沈阳计算技术研究所, 沈阳 110168

Intelligent Auxiliary Correction System for X-Ray Powder Diffractometer
LIU Feng, YANG Fan, YU Bi-Hui
University of Chinese Academy of Sciences, Beijing 100049, China;
Shenyang Institute of Computing Technology, Chinese Academy of Sciences, Shenyang 110168, China
Abstract: X-Ray powder Diffraction (XRD) is an exact instrument for material research with complex physical structure. Thus, a small hardware deviation will affect the quality of diffraction data, and the deviation correction of its hardware equipment will effectively guarantee the correctness of the data. Therefore, the deviation type recognition should be carried out before correcting the deviation. In general, XRD equipment is calibrated by the the equipment manufacturer expert using the special instruments to inspect item by item. This calibration highly depends on personnel experience, and the process is complicated and unefficient. This paper discusses the XRD error types and causes, and develops a intelligent auxiliary correction system for XRD based on classification and prediction, to improve the efficiency of error recognition. The system uses the general computer hardware, analyzes the standard sample diffraction data, then identifies the instrument status and assists correction automatically by feature extraction, model training, and other steps. The results show that the system can quickly identify the deviation and achieve the purpose of auxiliary correction.
Key words: X-Ray powder Diffraction (XRD)     auxiliary correction     classification

1 数据偏差现象及原因分析

1.1 强度偏差

1.2 角度偏差

1.3 偏心偏差

2 偏差判断

X射线粉末衍射仪是十分精密的科学仪器, 偏差数据出现概率小, 如果给每份数据进行全部的系统操作是不经济的, 偏差判断部分主要是筛选可能存在偏差的数据, 大大降低系统工作量. 偏差判断流程如图1所示. 另外本文实验数据为设备出厂数据, 偏差率大于应用中数据.

 图 1 偏差判断流程

2.1 数据描述

2.2 寻峰

 图 2 硅粉试样标准衍射值

2.3 BP神经网络模型

BP神经网络实质上实现了一个从输入到输出的映射功能, 而数学理论已证明它具有实现任何复杂非线性映射的功能. 这使得它特别适合于求解内部机制复杂的问题. 该功能输入特征为峰位、峰强, 维度达到44, 44维特征综合影响结果, 内部机制复杂, 适合使用神经网络模型; BP神经网络能通过学习带正确答案的实例集自动提取“合理的”求解规则, 具有自学习能力, 正好适合此处输出0到1的连续值用于判断此时状态存在偏差的可能性.

 $E(i) = \frac{1}{2}{\left[ {t(i) - y(i)} \right]^2}$ (1)

 $f\left( x \right) = \frac{1}{{1 + {e^{ - x}}}},\;\;{f^,}\left( x \right) = f\left( x \right)\left( {1 - f\left( x \right)} \right)$ (2)

 ${N_j} = \sum\nolimits_{p = 1}^{44} {{w_{jp}} * {x_p} + {b_j}} ,f({N_j})$ (3)
 ${n_k} = \sum\nolimits_{p = 1}^{20} {{w_{kp}} * f({N_p})} + {b_k},f({n_k})$ (4)

 图 3 偏差判断神经网络模型

3 偏差识别

 图 4 偏差识别流程

3.1 基于知识的特征处理

 \left\{ \begin{aligned}&\Delta {S_s} = {S_s} - S_s'\\&\Delta {S_p} = {S_p} - S_p'\\&\Delta {S_e} = {S_e} - S_e'\end{aligned}\right. (5)

 $\Delta {a_i} = {A_{0(i + 1)}} - {A_{ei}}$ (6)

 图 5 硅粉试样位置偏差衍射值

 $\Delta H = \frac{{\left( {{H'} - H} \right)}}{H}$ (7)

 $X = {C_0} + {C_1} * \sin ({C_2} * T + {C_3}) + {C_4} * T$ (8)

 图 6 硅粉试样强度偏差衍射值

3.2 分类树模型

 $split\_inf {o_A}(D) = - \sum\nolimits_{j = 1}^x {\frac{{|{D_j}|}}{{|D|}}} {\log _2}(\frac{{|{D_j}|}}{{|D|}})$ (9)

 $gain\_ratio(A) = \frac{{gain(A)}}{{split\_\inf o(A)}}$ (10)

4 结论

 [1] 马礼敦. X射线粉末衍射仪性能的评估. 上海计量测试, 2007, 34(2): 10-15. [2] 胡丽华, 张勇, 唐娟, 等. X射线粉末衍射仪的测试及使用. 化学工程师, 2012, 26(1): 16-17, 20. [3] Gozzo F, Cervellino A, Leoni M, et al. Instrumental profile of MYTHEN detector in Debye-Scherrer geometry. Zeitschrift für Kristallographie, 2010, 225(12): 616-624. DOI:10.1524/zkri.2010.1345 [4] Du R, Cai Q, Chen ZJ, et al. Mythen detector for X-ray diffraction at the Beijing synchrotron radiation facility. Instrumentation Science & Technology, 2016, 44(1): 1-11. [5] 柳小桐. BP神经网络输入层数据归一化研究. 机械工程与自动化, 2010(3): 122-123, 126. [6] 禹建丽, 卞帅. 基于BP神经网络的变压器故障诊断模型. 系统仿真学报, 2014, 26(6): 1343-1349. [7] 邓万红, 籍艳, 王平. BP神经网络在诊断气象仪故障的研究. 科技创新与生产力, 2014(9): 68-69. [8] 王继强. C4.5算法在信号设备故障诊断中的应用研究. 电子世界, 2012(9): 107-109.