遗传-神经网络方法在炸药性能预测中的应用研究  

Application Research of Genetic-Neural Network Method in the Performance Prediction of Explosive

作　　者：于国强[1] 董璐阳 王国强[1] YU Guo-qiang;DONG Lu-yang;WANG Guo-qiang(Shaanxi Institute of Applied Physical Chemistry, Shaanxi Xi'an 710061, China)

机构地区：[1]陕西应用物理化学研究所

出　　处：《辽宁化工》2019年第7期672-675,共4页Liaoning Chemical Industry

摘　　要：采用炸药的装药粒度、相对密度等因素建立炸药爆轰临界直径的人工神经网络预测模型,并采用遗传算法对预测模型的权值与阈值进行遗传优化,采用遗传-神经网络模型预测炸药的爆轰临界直径。预测结果表明,遗传-神经网络模型预测结果更加准确。11 Kinds of macroscopy and microscopy factors of explosive fabrication including grain size, relative density,polarizability and etc. were studied to establish the artificial neural network(ANN) prediction model of critical diameter. Genetic algorithm was used to optimize the threshold value and weight value of ANN model, in order to realize the accurately prediction of explosive performance. Smaller error and better effect was obtained with the optimized GA-BP model. According to the predicting results, it can be concluded that the artificial neural network model is feasible to predict critical diameter of explosives.

关 键 词：炸药 临界直径 神经网络 遗传算法 

分 类 号：TQ564[化学工程—炸药化工]