论文部分内容阅读
粘接类复合材料在航天领域得到应用广泛,但是粘接质量的判断一直是航天复合材料发展的一个重要课题。敲击检测在应用于复合材料粘接质量的检测时,受复合材料自身特性的影响较小,因此受到越来越多的关注;但是,敲击检测方法的数据处理问题一直没有得到解决,因此这也制约了其进一步发展。本文以敲击检测的数据处理为出发点,提出了聚类分析技术用以解决这一难题。经过选取数据样本进行验证,发现基于自组织竞争神经网络的聚类分析技术,可以很好地解决敲击检测的数据处理问题。
Adhesive composite materials are widely used in aerospace field, but the judgment of bonding quality has been an important issue in the development of aerospace composite materials. Punch detection is more and more concerned when it is applied to the testing of the quality of the composite material. Therefore, the data processing problem of the percussion detection method has not been solved. Therefore, This also restricts its further development. In this paper, percussion detection data processing as a starting point, proposed clustering analysis technology to solve this problem. After selected data samples for verification, it is found that clustering analysis technology based on self-organizing competitive neural network can solve the data processing problem of percussive detection well.