Classifying and Identifying Asbestos and Non-Asbestos Fibers by a Rule Building Expert System

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