红外脉冲相位复调制细化检测算法

Algorithm of infrared pulse phase thermography multiple modulation zoom spectrum inspection

试件发生损伤时会在频谱上表现为一系列特有的频率形式,该频率随材料热物性、损伤深度和尺寸的变化而变化。传统红外脉冲相位法(PPT)基于快速傅里叶变换(FFT)算法,在采样率固定的情况下,因栅栏效应、频率分辨率不足,无法准确识别峰值频率和"盲频",难以满足准确的定量分析。基于复调制的频率细化方法,首先分析全景频谱,再进行复调制移频,对特定的局部频带进行频谱细化,并通过零相位滤波器的改进,避免了附加相位和弥散现象,提高了频率分辨率,校正了频率、幅值和相位。通过铝制样件的实验,测试结果在定性分析时,可得到效果更优的幅值和相位图像;定量分析时,可得到更可靠的特征频率、更好的定量化损伤参数,能满足工程误差的要求。

Specimen damages induce a series of characteristic frequencies in the frequency spectrum, these frequencies change along with the variety of the material thermal properties, the damage depths and sizes. The traditional infrared Pulse Phase Thermography （PPT） is based on the Fast Fourier Transform （FFT） algorithm, with a fixed sampling rate, the frequency resolution is insufficient to accurately identify the peak frequency and the blind-frequency due to the barrier effect. Therefore, it is difficult to fulfill the accurate quantitative analysis. A frequency refinement algorithm based on the complex frequency modulation was proposed. The panoramic view of spectrum was analyzed, and the selected specific frequency band was deployed to local refine. With the improvement of zero phase filter, it was able to avoid the signal additional phase and dispersion, and the frequency resolution was improved, whilst the frequency, amplitude and phase were corrected. The experimental results show that this algorithm can obtain better amplitude and phase images, in addition, the reliable characteristic frequencies can be extracted, thus better quantitative damage parameters can be achieved in quantitative analysis. It meets the requirements of engineering errors.