复合材料低速冲击损伤面积神经网络估算方法

损伤面积是复合材料层合板在受低速冲击后其损伤严重程度主要表征参数之一。本文基于3层拓扑结构的BP人工神经网络,以冲击能量和凹坑深度作为输入参数,建立了损伤面积的快速估算模型。利用试验样本数据对BP神经网络模型训练后,选取样本数据进行仿真验证。对比分析说明该损伤面积估算模型具有良好的试验数据内在联系发掘能力,估算准确性与效率较高。本文研究为复合材料层合板低速冲击损伤面积估算提供了一种新的有效方法。

Estimation Method for Damage Area After Low-Velocity Impact of Composite Material Based on Neural Network

The damage area is one of the main characterization parameters of the extent of the damage in the composite laminates after the low-velocity impact. Based on the BP artificial neural network of three-layer topological structure, this paper establishes a rapid estimation model of damage area with the impact energy and the depth of indentation as input parameters. After training the BP neural network model with the test sample data, other sample data are simulated and verified. Through the comparative analysis, it is considered that the damage area estimation model has sufficient generalization ability of experimental data, and its estimation accuracy and efficiency can fulfill the requirements. It provides a new and effective method for estimating the damage area of composite laminates after low-velocity impact.