激光杂志
激光雜誌
격광잡지
LASER JOURNAL
2014年
1期
35-36
,共2页
杨友良%李俊香%马翠红%孟凡伟
楊友良%李俊香%馬翠紅%孟凡偉
양우량%리준향%마취홍%맹범위
激光诱导击穿光谱技术%多元素辨识%多元统计%特征谱线
激光誘導擊穿光譜技術%多元素辨識%多元統計%特徵譜線
격광유도격천광보기술%다원소변식%다원통계%특정보선
laser-induced breakdown spectroscopy%multi-element identification%multivariate statistics%spectral lines
激光诱导击穿光谱技术具有多元素同时检测的优点,基于钢液成分比较复杂,这一技术越来越多的应用于冶金行业。如果检测到某元素的特征谱线,即可以定性判断激发样品内存在该元素;但因为光谱检测仪器、激发条件与元素探测浓度极限的影响,某元素的谱线在光谱中不存在,并不能说明该元素在激发样品中就不存在,为提高钢液中多元素同时测量的精密度和灵敏度的研究,提出了将数理统计的多元统计分析与激光诱导击穿光谱技术定性分析相结合的优化方法。利用多元中心极限定理,无论总体分布如何,许多多元统计量的分布是近似正态的,选取多元素的若干特征谱线,建立多元素同时辨识的模型与方法,通过实验验证了此方法的可行性。
激光誘導擊穿光譜技術具有多元素同時檢測的優點,基于鋼液成分比較複雜,這一技術越來越多的應用于冶金行業。如果檢測到某元素的特徵譜線,即可以定性判斷激髮樣品內存在該元素;但因為光譜檢測儀器、激髮條件與元素探測濃度極限的影響,某元素的譜線在光譜中不存在,併不能說明該元素在激髮樣品中就不存在,為提高鋼液中多元素同時測量的精密度和靈敏度的研究,提齣瞭將數理統計的多元統計分析與激光誘導擊穿光譜技術定性分析相結閤的優化方法。利用多元中心極限定理,無論總體分佈如何,許多多元統計量的分佈是近似正態的,選取多元素的若榦特徵譜線,建立多元素同時辨識的模型與方法,通過實驗驗證瞭此方法的可行性。
격광유도격천광보기술구유다원소동시검측적우점,기우강액성분비교복잡,저일기술월래월다적응용우야금행업。여과검측도모원소적특정보선,즉가이정성판단격발양품내존재해원소;단인위광보검측의기、격발조건여원소탐측농도겁한적영향,모원소적보선재광보중불존재,병불능설명해원소재격발양품중취불존재,위제고강액중다원소동시측량적정밀도화령민도적연구,제출료장수리통계적다원통계분석여격광유도격천광보기술정성분석상결합적우화방법。이용다원중심겁한정리,무론총체분포여하,허다다원통계량적분포시근사정태적,선취다원소적약간특정보선,건립다원소동시변식적모형여방법,통과실험험증료차방법적가행성。
The advantage of laser-induced breakdown spectroscopy is simultaneous multi-element identification. Based on liq-uid steel composition is very complex. This technology is widely used in metallurgical industry. If the detected some elements of the spectral lines, which can determine the excitation of the sample in the memory element; but because of the influence of spec-trum detection instrument, excitation conditions and element concentration limit of detection. if spectral line of an element does not exist in the spectrum, which could not explain the element does not exist in the excitation of the sample. To improve the mul-tielement measuring precision and sensitivity for liquid steel, this paper presents a new method. It is combining the multivariate statistical analysis and mathematical statistics, qualitative analysis of laser induced breakdown spectroscopy. Using the multivariate central limit theorem, no matter how many multivariate statistics of population distribution, the distribution is approximately nor-mal, select the background spectral line number of point, model and method to establish the multielement simultaneous identifica-tion. It is through the experiment to verify the feasibility of this method.