Application of Finite Element Analysis and BP Neural Network to Infrared Nondestructive Testing
-
摘要: 应用有限元分析和BP神经网络分析主动加热式红外无损检测中影响缺陷定量的因素。对背部开有不同深度、不同直径沉孔缺陷的平板表面采用恒定热流加热。使用有限元分析软件Ansys计算得到表面温度分布云图和温升。将计算数据作为样本训练用于缺陷定量分析的BP神经网络，测试表明神经网络方法对于未知缺陷深度的识别相当有效。研究表明该信号处理技术对红外无损检测具有工程应用价值。Abstract: Finite element analysis combined with back propagation(BP) neural network were applied to the active heating infrared nondestructive testing for determing the influencing factors on defect quantitation. Finite element analysis software Ansys, was used to calculate the temperature field of the plates with back sinking hole defects of different depth and diameter under constant heating condition, and surface temperature cloudy map and temperature rise were obtained. The data obtained were used as the sample to train the BP neural network for quantitative evaluation of defects. It was showed by testing that neural network was very effective in predicting the depth of defects. The signal processing method was valuable for engineering application to infrared nondestructive testing.
-
-
[1] 虞和洛，宋利明.故障诊断的热像技术［M］.北京：冶金工业出版社，1992. [2] 董其国.红外诊断技术在电力设备中的应用［M］.北京：机械工业出版社，1998. [3] 陈 衡，侯善敬.电力设备故障红外诊断［M］.北京：中国电力出版社，1989. [4] 梅 林，陈自强，王裕文，等.脉冲加热红外热成像无损检测的有限元模拟及分析［J］.西安交通大学学报,2000,34(1):66-70. [5] Vavilov V, Grinzato E, Bison PG, et al. Surface transient temperature inversion for hidden corrosion characterization: theory and application\[J\]. Int J Heat Mass Transfer,1996,39(2):335-371. [6] 于 慧，关荣华.红外技术定量诊断设备内部缺陷可行性的理论证明［J］.红外与激光工程,2002,31(3):229. [7] 杨建刚.人工神经网络实用教程［M］.杭州：浙江大学出版社，2001. [8] Xavier Maldague, Yves Largouet, Jean Pierre Couturier. A study of defect depth using neural networks in pulsed phase thermography: modelling, noise, experiments［J］. Rev Gen Therm,1998,37:704-717.
计量
- 文章访问数: 6
- HTML全文浏览量: 1
- PDF下载量: 7
下载:
