Pre-trained artificial neural network for prediction of long-rod penetration depth in a semi-infinite target

A pre-trained artificial neural network is provided for predicting the scaled penetration depth P/L of a rod penetrating into a semi-infinite target, based on the rod length-over-diameter ratio L/D, the impact velocity and the density and hardness of the target and projectile materials.

The tensorflow keras sequential network is stored as neural_network.h5 in hierarchical data format (HDF5). The script main.py demonstrates an application of the neural network. Alternatively, the notebook main.ipynb or its static version main.html may be consulted.

本数据集提供了一项经过预训练的人工神经网络，该网络可用于预测棒状物体穿透半无限目标时的比例穿透深度P/L。该预测基于棒状物体的长度与直径之比L/D、冲击速度以及目标物和投射物材料的密度和硬度。该tensorflow keras顺序网络以层次数据格式（HDF5）存储，具体为neural_network.h5文件。main.py脚本展示了该神经网络的实际应用。此外，可通过main.ipynb笔记本或其静态版本main.html进行查阅。