Calibration free LIBS using multivariate hybrid chemometrics: An approach for precise quantitative elemental analysis of solid samples. PhD thesis, King Fahd University of Petroleum and Minerals.
|
PDF (PhD thesis)
final_pdf_copy.pdf Download (8MB) | Preview |
Arabic Abstract
هذف هذا العول لذراست وتحسين اداء الطزق الوستخذهت في التحليل الكوي لطيف التكسيز الوستحج بالليشر بىاسطت طيف هن الوصفىفاث الوختلفت والتي تشول عيناث صلبت هخل العيناث البزونشيت المياسيت وعيناث شبه السىائل هخل العنب وعيناث هتىافمت حيىيا هخل السوك. تن التزاح وتطىيز حلاث تمنياث جذيذة لتحسين اداء الطزق الوستخذهت للتحليل الكوي لطيف التكسيز الوستحج بالليشر. التحمك التجزيبي للتمنياث الومتزحت أظهز نتائج هوتاسة لوجوىعت هن العيناث الوختلفت. سيادة دلت التحليل الكوي للعناصز الكيويائيت في العيناث الصلبت بىاسطت هطياف التكسيز الوستحج بالليشر سىف يسهل استخذام الوطياف للعول الويذاني ولابليت تطبيمه في نطاق أوسع
English Abstract
Performance enhancement techniques for the improvement of chemometrics that are employed for quantitative analysis of LIBS spectra have been comprehensively studied in this work using spectral from different matrices which include solid samples such as standard bronze samples, semi-fluid samples such as grapes and biocompatible samples such as fish. Three novel techniques for performance enhancement are proposed and developed for hybrid support vector regression (SVR) based chemometrics as well as hybrid extreme learning machine (ELM) based chemometrics which are used for quantitative analysis of LIBS spectra. Specifically, the proposed techniques are internal reference preprocessing (IRP), homogenous hybridization and hybrid fusion. Prior to the implementation of the proposed techniques, sufficiency of single emission line for quantitative analysis of LIBS spectra using the developed chemometrics was investigated by comparing the elemental concentrations obtained from the developed sensitivity based linear learning method (SBLLM) based chemometrics using single, double and three emission lines. Experimental validation of the proposed techniques was carried out using seven standard bronze samples and excellent results are obtained. In addition, hybrid xix support vector regression and hybrid extreme learning machine chemometrics are also developed and implemented for quantitative analysis of crayfish and grape samples. The obtained results from the two chemometric models were verified and compared with the result obtained from standard analytical technique such as inductively coupled plasma mass spectrometry (ICP-MS). Implementation of the developed performance enhancement techniques for the investigated chemometrics tools employed for quantitative analysis of LIBS spectra would definitely enhance the precision of quantitative analysis of LIBS spectra, especially for in situ applications, and ultimately widen the applicability of the technique
| Item Type: | Thesis (PhD) |
|---|---|
| Subjects: | Physics |
| Department: | College of Engineering and Physics > Physics |
| Committee Advisor: | Gondal, Mohammed |
| Committee Members: | Nasser, Ibraheem and Albasheer, Watheq and Naqvi, Akthar and Chanbasha, Basheer |
| Depositing User: | TAOREED OLAKUNLE OWOLABI (g201203020) |
| Date Deposited: | 12 Aug 2018 05:58 |
| Last Modified: | 31 Dec 2020 09:03 |
| URI: | http://eprints.kfupm.edu.sa/id/eprint/140775 |