人工神经网络预测液体叠氮燃料密度

Prediction on density of liquid azido-fuel with artificial neural network

采用人工神经网络进行了液体叠氮燃料密度预测。通过将已知的特征分子结构编码作为输入参数,设计了神经网络代码,并得到了预测的密度数据。结果表明,密度预估值和文献值相比偏差为–1.8%~2.69%,人工神经网络对液体叠氮燃料的密度预测结果具有一定参考价值。

The density of the liquid azido-fuel was predicted by using the artificial neural network. Using the known characteristic molecular structure code as the input parameter, the neural network code was designed, and the predicted density data were obtained. The results show that the relative derivation between the predicted density values and the literature values is within-1.8% ~ 2.69%. The artificial neural network possesses certain reference value for prediction results of liquid azido-fuel density.