基于声发射信号信息熵的木材损伤断裂过程研究

为获得木材在弯曲破坏过程中的声发射（acoustic emission，AE）信号特征，从AE信号的随机性出发，利用AE信号信息熵辨识木材的损伤过程，并研究木材在不同损伤断裂水平下的AE信号分布特性。首先，对气干状态的榉木和樟子松试件进行三点弯曲试验，并通过谐振频率为150 kHz的AE传感器采集原始AE信号，采样频率设置为500 kHz。然后，采用小波变换重构AE信号波形，依据无AE发生时的信号幅值确定AE阈值，统计每秒内超过阈值的次数并作为AE活动计数，再以活动计数为随机变量定义AE信息熵。最后，依据信息熵值确定应变能释放的转折点，并结合三点弯曲试验的载荷-时间曲线，将木材损伤断裂过程划分为线性变形、非线性变形、宏观断裂3个阶段。以10 ms为间隔分析并统计AE信号的频率，获得木材弯曲破坏过程的AE信号频率分布情况，从而揭示不同损伤阶段的AE信号特征。结果表明，线性变形阶段，AE信号表现为低幅值、低频率，主要集中在30~55 kHz频段内；非线性变形和宏观断裂阶段，AE信号中既存在大量的30~55 kHz低频信号成分，又存在100~110 kHz和115~130 kHz的高频信号。研究提出的基于AE活动数信息熵能够准确反映应变能释放的集中程度，为木材损伤断裂水平评价提供了客观依据。

Research on Wood Damage Fracture Process Based on Acoustic Emission Signal Information Entropy

In order to obtain the acoustic emission (AE) signal characteristics of wood during bending failure,the AE signal information entropy was used to identify the damage process of wood from the randomness of AE signal,and the distribution characteristics of AE signal under different damage and fracture levels were studied.Firstly,three-point bending test was carried out on the air-dried beech and Pinus sylvestris var.mongolica specimens,and the original AE signal was collected by AE sensor with resonance frequency of 150 kHz,and the sampling frequency was set to 500 kHz.Then,the AE signal waveform was reconstructed by wavelet transform,the AE threshold was determined according to the signal amplitude when no AE occurred,the number of times exceeding the threshold per second was counted as the AE activity counted,and then the AE information entropy was defined with the activity count as the random variable.Finally,according to the information entropy,the turning point of strain energy release was determined,and combined with the load-time curve of three-point bending test,the wood damage and fracture process was divided into three stages:linear deformation,nonlinear deformation and macroscopic fracture.The frequency of AE signal was analyzed and counted at 10 ms intervals,and the frequency distribution of AE signal in the process of wood bending failure was obtained,thus revealing the characteristics of AE signal in different damage stages.The results showed that during the linear deformation stage,AE signals were characterized by low amplitude and low frequency,mainly concentrated in the frequency range of 30~55 kHz.In the stage of nonlinear deformation and macroscopic fracture,there were not only a large number of 30~55 kHz low frequency signals,but also 100~110 kHz and 115~130 kHz high frequency signals in AE signals.The information entropy based on AE activity number can accurately reflect the concentration degree of strain energy release,which provides an objective basis for the evaluation of wood damage and fracture level.