Prediction of Flash Points of Organosilicon Compounds
Flash point (FP) is the primary property to classify flammable liquids for the purpose of assessing their fire and explosion hazards. Because of the advancement of technology in discovery or synthesis of new compounds, there is often a significant gap between the demand for such data and their availability. In this work, a quantitative structure property relation (QSPR) study is presented for predicting the FP of a data set of 230-organosilicon compounds. The stepwise regression method is used to select the molecular descriptors for prediction of the FP from 3224 molecular descriptors. Depending on the p-value of accepting a descriptor to enter the model, models with different number of descriptors are obtained. A 13-descriptor model and a 6descriptor model are obtained with thep-value of 0.0005 and 0.00001, respectively. It is found that the 6-descriptor multiple linear regression model could fit the data with R2 = 0.8780. The 13descriptor multiple linear regression model could fit the data with R2 = 0.9214. The average fitting errors in percentage for these two models are less than 3.78％ and 3.08％, respectively. Thus, the proposed model offers reliable prediction of FP for the organosilicon compounds.
Flash Point organosilicon compounds Quantitative Structure Property Relation
TSAI Yi-jen CHEN Chan-cheng LIAW Homg-jang
Department of Occupational Safety and Health, China Medical University, 91 Hsueh-Shih Rd., Taichung Department of Safety, Health and Environmental Engineering, National Kaohsiung First University of S
国际会议
The 2010 International Symposium on Safety Science and Technology(2010 安全科学与技术国际会议)
杭州
英文
951-962
2010-10-26（万方平台首次上网日期，不代表论文的发表时间）