QSPR studies of impact sensitivity of nitro energetic compounds using three-dimensional descriptors

The quantitative structure–property relationship (QSPR) studies were performed between molecular structures and impact sensitivity for a diverse set of nitro energetic compounds based on three-dimensional (3D) descriptors. The entire set of 156 compounds was divided into a training set of 127 compounds and a test set of 29 compounds according to Kennard and Stones algorithm. Multiple linear regression (MLR) analysis was employed to select the best subset of descriptors and to build linear models; while nonlinear models were developed by means of artificial neural network (ANN). The obtained models with ten descriptors involved show good predictive power for the test set: a squared correlation coefficient (R2) of 0.7222 and a standard error of estimation (s) of 0.177 were achieved by the MLR model; while by the ANN model, R2 and s were 0.8658 and 0.130, respectively. Therefore, the proposed models can be used to predict the impact sensitivity of new nitro compounds for engineering.

Figure optionsDownload high-quality image (109 K)Download as PowerPoint slideHighlights
► New QSPR models were presented to predict impact sensitivity of nitro compounds.
► Dataset was split into training and test set by Kennard and Stones algorithm.
► Three-dimensional descriptors were used to represent molecular structures.
► QSPR methodologies used multiple linear regression and neural networks.
► The nonlinear model demonstrated better predictive ability than the linear model.