红外热波技术、有限元与SVM相结合的复合材料分层缺陷检测方法

Composite layer defect detection method based on infrared heat wave technology,finite element and SVM

现有对复合材料分层缺陷的检测方法,因数据样本少导致精度低、耗时长、成本高等问题,限制了其广泛应用。提出了一种基于红外热波技术、有限元与支持向量机相结合的复合材料缺陷检测方法。通过建立有限元热分析模型获取激励面所有节点的温升数据,并与红外热波实验所得数据从单个与整体方面进行分析,结果表明,仿真数据与实验数据之间具有强相关性且仿真数据没有环境误差,推断出仿真数据可作为训练样本;建立SVM模型并用实验数据进行测试,以测试结果判断复合材料是否存在分层缺陷,该方法通过有限元仿真大大增加了样本的数量。实验结果表明,设计的4组实验中分层缺陷识别率分别为80.36%、77.99%、86.96%、81.13%,证实了该检测分层缺陷的方案的可行性。

Because of the problems of low accuracy,long time and high cost due to the lack of data samples,the existing detection methods for layered defects of composite materials limit their wide application.In this paper,a composite defect detection method based on infrared heat wave technology,finite element and support vector machine is proposed.The temperature rise data of all nodes on the excitation surface were obtained by establishing a finite element thermal analysis model.And analyze experiment data obtained by the infrared thermal wave experiment from the individual and the overall aspect.The results showed a strong correlation between simulation data and experimental data.And the simulation data has no environmental error.Inferred that the simulation data could be used as training samples.Establishing SVM model and testing with experimental data to determine whether the composite material has delamination defects based on the test results.This method greatly increases the number of samples through finite element simulation.The experimental results show that the identification rates of layered defects in the four sets of experiments are 80.36%,77.99%,86.96%,and 81.13%,respectively.It is confirmed that the scheme for detecting layered defects is feasible.