鉴于钛合金材料的声各向异性和高声波衰减系数使疲劳试验过程中的声发射监测十分困难。提出了小波包与AR谱相结合对钛合金材料声发射信号进行分析的方法。即对输出信号进行小波包分解，然后分频段对信号进行重构得到包含不同频率成分的时域信号，再进行AR谱分析，从而提取出频谱特征。分析表明，该方法对利用声发射信号分析钛合金材料部件的损伤和裂纹扩展是有效的。可为钛合金部件损伤和裂纹扩展的识别提供可靠的依据。

Titanium alloy was largely used in the joint of central wing and outer panels of some new type military aircraft. Acoustic emission (AE) monitoring of the material might face great difficulty in the process of fatigue test of some pivotal parts because of its acoustic anisotropy and high acoustic attenuation coefficient of titanium alloys. So the understanding of characteristics of AE of titanium alloys and providing effective AE signal identification means was obviously very important. This paper introduced the method of analyzing AE signal of titanium alloys by applying wavelet packet and AR spectrum. With the help of these discussions, it was seen that wavelet packet was able to extract specific frequency segment effectively, when detecting crack expending of titanium alloys and their components was very difficult by means of traditional parameter analysis. The combination of wavelet packet with AR spectrum provided credible basis for identifying the crack expending of titanium alloys parts and it showed that signal processing had great potential in AE monitoring of structure integration.