二维内部缺陷的红外瞬态定量识别算法

Transient quantitative infrared thermographic identification of two-dimensional inner defect

内部缺陷的反问题定量识别是红外无损检测技术的重要内容,而红外稳态识别方法存在受误差影响大等缺点。文中对具有矩形内部缺陷的试件建立了二维物理和数学传热模型,通过有限体积法求解瞬态温度分布,分析了检测表面的温度分布规律,运用Levenberg-Marquardt法研究了瞬态情况下对缺陷的尺寸和方位进行定量识别的方法。数值算例证明了算法的有效性;初始假设等对识别结果的影响不大;最大检测温差越大,识别结果越准确;瞬态测温法便于得到较大的检测温差,有利于识别结果的准确性。

The inverse quantitative identification of the inner defect is an important part of infrared nondestructive testing, but steady-state infrared testing technique is always affected by error and other factors. A two-dimensional physical and mathematical heat transfer model of a test piece with an inner rectangle defect was built. Through solving the transient equations with the finite volume method, the temperature distribution on the inspection surface was analyzed, and the Levenberg-Marquardt method was used to study the method of evaluating size and depth of the defect based on transient thermography. The effectiveness of the method is validated by numerical experiment. The effect of the initial guess is negligible. The larger the maximum temperature difference is, the more accurate the calculational results becomes. The larger temperature difference is easy to get with the transient thermography method, which is conducive to improve the accuracy of the results.