基于神经网络的长杆弹侵彻能力预测模型  

Prediction Model on Long-Rod Projectile Penetrating into Steel Target Based on Neural-Network

作　　者：王海宽[1] 李文生[1] 熊飞[1] 

机构地区：[1]军械工程学院装备指挥与管理系,河北石家庄050003

出　　处：《计算机仿真》2015年第2期1-5,66,共6页Computer Simulation

摘　　要：装备毁伤效应快速评定与预测,一直是装备使用人员、战场抢修人员所关注的重要问题,针对上述问题,利用ANSYS/LS-DYNA有限元软件对钨合金长杆弹侵彻中厚钢靶板全过程进行了数值模拟并进行了详细分析,同时通过均匀试验与数值仿真得到了弹体侵彻钢靶板的终点效应样本数据,然后利用BP神经网络对样本数据进行训练,从而构建了不同速度钨合金长杆弹侵彻不同厚度钢靶板后的剩余质量与剩余动能预测模型,实现了对钨合金长杆弹毁伤效应的快速预测,预测结果较为理想,为装备毁伤效应快速评定提供了基础。It＇s a key point to predict the equipment damage for equipment user and Battlefield Repair. Based on the problem, the process of Long - rod projectile of tungsten alloy penetrating into moderately thick steel target was analyzed with the ANSYS/LS -DYNA software. The sample dates were obtained by simulation and uniform testing design method, and then sample dates were trained by artificial neural network to build a prediction mode that can forecast the Long- rod projectile＇s remaining kinetic and remaining mass after it penetrated different thickness steel targets with different velocity. The prediction results are ideal, and it provides a basis for equipment damage assessment rapidly.

关 键 词：有限元 侵彻 终点效应 均匀试验 人工神经网络 

分 类 号：TJ012.4[兵器科学与技术—兵器发射理论与技术]