Chemometrics analysis of petroleum-based accelerants in fire debris
Petroleum-based accelerants such as diesel, gasoline, kerosene and others are usually related to fire debris analysis because they are inexpensive, readily available and commonly used to enhance the burning intensity of fire. However, combustion process and the presence of pyrolysis products can lea...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2015
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/53622/1/FatinAmalinaMFS2015.pdf |
| _version_ | 1796859868264529920 |
|---|---|
| author | Ahmad Shuhaimi, Fatin Amalina |
| author_facet | Ahmad Shuhaimi, Fatin Amalina |
| author_sort | Ahmad Shuhaimi, Fatin Amalina |
| collection | ePrints |
| description | Petroleum-based accelerants such as diesel, gasoline, kerosene and others are usually related to fire debris analysis because they are inexpensive, readily available and commonly used to enhance the burning intensity of fire. However, combustion process and the presence of pyrolysis products can lead to misclassification of accelerants to the arson investigator. Furthermore, fire debris which has been exposed for several days may undergo some component lost and makes the detection more difficult. In this study, gas chromatography-mass spectrometry (GC-MS) was used to identify the accelerants present in simulated arson incidents. Total ion chromatogram and the peak area from the GC-MS data were used to perform chemometrics techniques which include principal component analysis (PCA), linear discriminant analysis (LDA), partial least square-discriminant analysis (PLS-DA) and support vector machine (SVM). The performance of these methods was further tested by analyzing samples which have been exposed for several days in the environment. Three accelerant classes were formed by these classification models which consist of gasoline, kerosene and diesel. Supervised pattern recognition technique showed satisfactory results, in terms of correctly classified samples, which were 90.4% (LDA), 85.3% (PLS-DA) and 96.7% (SVM) for training sets. A test set produced a value of 87.5% correct classification for LDA, 83.3% for PLS-DA while the best classification is 91.7% by SVM. Fire debris analysis using GC-MS with the aid of chemometrics methods give a promising result in the identification and classification of accelerants used to initiate the fire in arson cases. |
| first_indexed | 2024-03-05T19:33:36Z |
| format | Thesis |
| id | utm.eprints-53622 |
| institution | Universiti Teknologi Malaysia - ePrints |
| language | English |
| last_indexed | 2024-03-05T19:33:36Z |
| publishDate | 2015 |
| record_format | dspace |
| spelling | utm.eprints-536222020-08-23T07:45:16Z http://eprints.utm.my/53622/ Chemometrics analysis of petroleum-based accelerants in fire debris Ahmad Shuhaimi, Fatin Amalina QD Chemistry Petroleum-based accelerants such as diesel, gasoline, kerosene and others are usually related to fire debris analysis because they are inexpensive, readily available and commonly used to enhance the burning intensity of fire. However, combustion process and the presence of pyrolysis products can lead to misclassification of accelerants to the arson investigator. Furthermore, fire debris which has been exposed for several days may undergo some component lost and makes the detection more difficult. In this study, gas chromatography-mass spectrometry (GC-MS) was used to identify the accelerants present in simulated arson incidents. Total ion chromatogram and the peak area from the GC-MS data were used to perform chemometrics techniques which include principal component analysis (PCA), linear discriminant analysis (LDA), partial least square-discriminant analysis (PLS-DA) and support vector machine (SVM). The performance of these methods was further tested by analyzing samples which have been exposed for several days in the environment. Three accelerant classes were formed by these classification models which consist of gasoline, kerosene and diesel. Supervised pattern recognition technique showed satisfactory results, in terms of correctly classified samples, which were 90.4% (LDA), 85.3% (PLS-DA) and 96.7% (SVM) for training sets. A test set produced a value of 87.5% correct classification for LDA, 83.3% for PLS-DA while the best classification is 91.7% by SVM. Fire debris analysis using GC-MS with the aid of chemometrics methods give a promising result in the identification and classification of accelerants used to initiate the fire in arson cases. 2015-04 Thesis NonPeerReviewed application/pdf en http://eprints.utm.my/53622/1/FatinAmalinaMFS2015.pdf Ahmad Shuhaimi, Fatin Amalina (2015) Chemometrics analysis of petroleum-based accelerants in fire debris. Masters thesis, Universiti Teknologi Malaysia, Faculty of Science. http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:84651 |
| spellingShingle | QD Chemistry Ahmad Shuhaimi, Fatin Amalina Chemometrics analysis of petroleum-based accelerants in fire debris |
| title | Chemometrics analysis of petroleum-based accelerants in fire debris |
| title_full | Chemometrics analysis of petroleum-based accelerants in fire debris |
| title_fullStr | Chemometrics analysis of petroleum-based accelerants in fire debris |
| title_full_unstemmed | Chemometrics analysis of petroleum-based accelerants in fire debris |
| title_short | Chemometrics analysis of petroleum-based accelerants in fire debris |
| title_sort | chemometrics analysis of petroleum based accelerants in fire debris |
| topic | QD Chemistry |
| url | http://eprints.utm.my/53622/1/FatinAmalinaMFS2015.pdf |
| work_keys_str_mv | AT ahmadshuhaimifatinamalina chemometricsanalysisofpetroleumbasedaccelerantsinfiredebris |