Research on the self-accelerating decomposition temperature of organic peroxides by neural network
In order to study the quantitative structure-activity relationship between the self-accelerating decomposi-tion temperature(SADT)and the structure of organic peroxides,molecular shape index(nK)and electrical distance vector index(M,)of 46 organic peroxides were calculated and 8 structural parameters,namely 2K,3K and 4K of the molecular shape indices,and M9,M21,M26,M27 and M33 of the electrical distance vector indices were obtained by screening and optimization.With the eight parameters as input variables of neural network and the SADT as output variable,the 8:4:1 network structure was adopted and BP neural network method was used to establish a satisfying quantitative structure-property relationship(QSPR)prediction model.The total correlation coefficient was 0.997.The predicted values of SADT by the model were in satisfactory agreement with the experiment values,with an average er-ror of only 0.9℃,which was better than literature methods.The results showed that there was a good nonlinear relation-ship between the self-accelerating decomposition temperature and the eight molecular structure parameters.
万方数据
