A Study On the Comparison Of Two Receptor Models:Chemical Mass Balance Model and Positive Matrix Factorization Model

• Main Authors: Jyh-Feng Liang, 梁志鋒
• Other Authors: 望熙榮
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/94883354832951295833

碩士 === 中興大學 === 環境工程學系所 === 94 === Two receptor models, Chemical Mass Balance (CMB) and Positive Matrix Factorization (PMF), are applied to estimate the source contributions of TaChia area in this study. There are thirty three samples of PM2.5 and PM2.5~10 , respectively. And four samples are impacted by Asian Dust Storm. This research will analysis the source contribution of Asian Dust Storm by receptor model. CMB model is written by Matlab program language. Condition Index and π matrix are used to identify the collinearity of source profiles by CMB model. Their advantages are that collinearity of source profiles are defined definitely and source profiles can be chosen flexible. PMF model is used the EPA PMF 1.1 version developed by USEPA. The results of two models are compared. Vehicle emissions, vegetative burning, ammonium sulfate, ammonium nitrate, crustal materials, incinerator, oil-fired boiled are analyzed to the source contributions of PM2.5 by two receptor models. Vehicle emissions are the major source contributions of PM2.5, and it was estimated about 57 % and 35 % of PM2.5 by CMB model and PMF model. The second to fourth source contributions are vegetative burning, ammonium sulfate, ammonium nitrate, and they accounts for 44.3 % and 48.4 % of total source contributions to CMB model and PMF model, respectively. Six sources include vehicle emissions, crustal materials, marine spray, ammonium nitrate, incinerator, vegetative burning are resolved to the source contribution of PM2.5~10 by two receptor models. The results show that vehicle emission and crustal materials are primary and secondary source contributions of PM2.5~10. They accounts for 74 % and 61 % of total source contributions to PM2.5~10 according to the results obtained from CMB model and PMF model, respectively. Vehicle emissions estimated by CMB model are still 52 % of total source contribution higher than 35 % estimated by PMF model. The contribution of Asian dust storm is only resolved by CMB model, and it accounts for 3.4 % of total source contributions. Instead, PMF model can’t resolve the contribution of Asian dust storm. In conclusion, the major sources identified by the two receptor models are the same. The reason why high differences of contributions to vehicle emissions may be the source profile collected from the foreign area and it is not proper for the characteristics of vehicle emissions for TaChia area. Incinerator and crustal materials are low percentages of total contributions and their regression coefficient(r2) are low of the results between CMB model and PMF model. The reasons may be due to the incompleteness of profiles and a lack of local-specific profiles. In addition, a lack of samples to Asian dust storm, PMF model can’t resolve the source contribution of Asian dust storm.