전문가시스템을 이용한 석면 및 비석면의 분류 및 확인 (Classifying and Identifying Asbestos andNon-Asbestos Fibers by a Rule Building Expert System)

서지정보

· 발행기관 : 한국대기환경학회
· 수록지 정보 : 한국대기환경학회지 / 24권 / 3호 / 346 ~ 356페이지
· 저자명 : 최영아, 이태정, 김동술

초록

Asbestos is the name of a group of minerals with long and thin bers that originate naturally in the environment.Asbestos mainly affects lungs and the membrane that surrounds the lungs. In general, PCM(phase contrast micro-scopy) and PLM (polarized light microscopy) have been used to analyze asbestos bers. However, these methodshave often problems to over-estimate number concentration when counting real asbestos bers. Moreover, there aremany difculties when separating and identifying various asbestos and non-asbestos bers. In order to determinequantitative information on brous particles, source proles for asbestos and non-asbestos bers must be initiallydeveloped on the basis of their chemical compositions and physical parameters. In our study, a SEM/EDX was used to develop source proles from known asbestos samples as reference sam-ples. We could make the source prole matrix consisting of 6 types of asbestos bers and 2 types of non-asbestosbers by analyzing 380 bers. Based on these proles, a rule building expert system was developed by using thevisual basic application(VBA). Various bers were successfully classied by 2 simple rules in the EXCEL environ-ment based on several visual steps such as inserting data, viewing results, and saving results. For a case study totest the expert system, samples from a construction materials and from various indoor environments such as a resi-dental area, a preschool classroom, and an underground store were collected and analyzed. As a result of the survey,a total of 76 individual test ber particles was well classied into 5 different types of particle classes; 9.3% of chry-sotile, 15.4% of amosite, 0.8 of crocidolite, 4.2% of tremolite, 5.8% glass ber, 21.1% of other bers, and 43.5% ofunknown bers in terms of number concentration. Even though unknown portion was high, it will be decreasedmarkedly when expanding ber source proles.

영어초록

Asbestos is the name of a group of minerals with long and thin bers that originate naturally in the environment.Asbestos mainly affects lungs and the membrane that surrounds the lungs. In general, PCM(phase contrast micro-scopy) and PLM (polarized light microscopy) have been used to analyze asbestos bers. However, these methodshave often problems to over-estimate number concentration when counting real asbestos bers. Moreover, there aremany difculties when separating and identifying various asbestos and non-asbestos bers. In order to determinequantitative information on brous particles, source proles for asbestos and non-asbestos bers must be initiallydeveloped on the basis of their chemical compositions and physical parameters. In our study, a SEM/EDX was used to develop source proles from known asbestos samples as reference sam-ples. We could make the source prole matrix consisting of 6 types of asbestos bers and 2 types of non-asbestosbers by analyzing 380 bers. Based on these proles, a rule building expert system was developed by using thevisual basic application(VBA). Various bers were successfully classied by 2 simple rules in the EXCEL environ-ment based on several visual steps such as inserting data, viewing results, and saving results. For a case study totest the expert system, samples from a construction materials and from various indoor environments such as a resi-dental area, a preschool classroom, and an underground store were collected and analyzed. As a result of the survey,a total of 76 individual test ber particles was well classied into 5 different types of particle classes; 9.3% of chry-sotile, 15.4% of amosite, 0.8 of crocidolite, 4.2% of tremolite, 5.8% glass ber, 21.1% of other bers, and 43.5% ofunknown bers in terms of number concentration. Even though unknown portion was high, it will be decreasedmarkedly when expanding ber source proles.

참고자료

· 없음