声-超声参数与木材解剖特征的关系研究

Correlation of wood anatomical properties acousto-ultrasonic nondestructive evaluation of Larch

为探索利用声-超声(AU)技术预测落叶松木材解剖性质的可行性,测量了落叶松无疵小试样的波速(V)、峰值电压(A)、均方根电压(RMS)、上升时间(RT)、频率形心(FC)等五个声-超声参数和年轮宽度(RW)、晚材率(LR)、微纤丝角(MFA)、管胞长度(TL)和管胞宽度(TW)等5个解剖特征。分别利用一元线性回归分析方法和多元线性回归分析方法研究了声-超声参数与解剖特征之间的相关性。结果表明:1)不同解剖特征和单个声-超声参数之间的相关性存在差异。其中,管胞长度与波速呈显著正相关,其相关性决定系数较高(R^2=0.31),与其它4个声-超声参数的相关性较低(R^2≤0.16);微纤丝角与波速呈显著负相关,其R^2=0.46,与其它4个声-超声参数的相关性较低(R^2≤0.14);年轮宽度、晚材率、管胞宽度与5个声-超声参数相关性都较差(R^2≤0.11)。2)利用多元回归分析方法建立的微纤丝角线性预测模型的效果高于仅以波速为自变量建立的微纤丝角线性预测模型的效果,决定系数从0.46提高到0.63。

In order to explore the feasibility of using acoustic-ultrasonic(AU) technique to predict the anstomical properties of Larch wood, measured the ?ve acoustic-lutrasonic parameters and ?veanatomical features which are wave velocity(V), peak voltage(A), root mean square voltage(RMS), rise time(RT), frequency centroid(FC) and ring width(RW), late wood rate(LR), micro?lament angle(MFA), tracheid length(TL). the univariate linear regression analysis method and multivariate linear regression analysis method are used respectively to study the correlation between a cousto-ultrasonic parameters and anatomical features. The results indicated that :(1) there were differences in the correlation between different anatomical features and single AU parameters. Among them, the tracheid length was signi?cantly positively correlated with the wave velocity with a high determination coef?cient(R^2=0.31), it is lower correlations with the other four AU parameters(R^2≤0.16). There was a significant negative correlation between the micro?ber angle and the wave velocity, with a high correlation coef?cien(R^2=0.46), it is lower correlations with the other four AU parameters(R^2≤0.14). The correlations almong the growth ring width, ratio of latewood, tracheid width and all of all the ?ve AU parameters were poor(R^2≤0.11);(2) The effect of the linear prediction model of micro?ber angle established by multiple regression analysis is higher than that of establishing only with wave velocity as the independent variable, and the determination coef?cient increased from 0.46 to 0.63.