马尾松木材管胞特征与物理力学性质的相关分析

为了了解马尾松管胞特征与物理力学性质之间的相关关系,对马尾松家系进行了管胞特征与物理力学性质的测定,通过相关分析和回归分析研究了它们之间的具体相关性,并导出了预测模型。结果表明,长宽比和壁腔比,长宽比和柔性系数在1~14 a年轮段相关性显著或极显著,15~22 a年轮段的相关性没有达到显著水平。壁腔比和柔性系数的相关性在整个年轮段都极显著。各个材性指标即气干密度、抗压强度和冲击韧性之间的相关性极显著。抗弯强度与其他材性指标间的相关性相对较小,与冲击韧性没有达到相关显著水平。3个管胞指标与各个材性指标的相关性都极显著。3个指标中壁腔比与各个材性指标的相关系数最高。以物理力学性质指标为因变量,考虑每个试件所处年轮和晚材率的管胞特征为自变量的三元三次方程回归模拟结果显示,几个材性指标中拟合度最高的是抗压强度,R~2是0.717,最低的是抗弯强度,R~2是0.362。实测值与预测值的相对误差最小的是气干密度,最大的是抗弯强度,具体误差值为7.59%和37.63%。根据回归模型,可以利用管胞特征定性预测部分材性指标。

Correlation analysis of tracheid characteristics and physical mechanics properties of Pinus massoniana timber

Tracheid characteristics and physical mechanics properties of Pinus massoniana were measured and their specific rele-vance were analyzed to understand the relationship among these indexes. The results showed that the correlation between length to width ratio and wall thickness to lumen ratio, and length to width ratio and flexibility coefficient was significant or extremely signifi-cant in the 1-14 ring segment, however, in 15-22 ring segment, the correlation did not reach a significant level. The correlation be-tween the wall thickness to lumen ratio and flexibility coefficient was extremely significant in the whole ring segments. The correlation between physical and mechanical properties was extremely significant. Among them, the bending strength had a little correlation with other indexes and no correlation with impact toughness. The correlation between tracheid characteristics and physical mechanics properties was extremely significant, and the Pearson coefficient of the wall thickness to lumen ratio was the highest one. According to the correlation relationships, physical and mechanical properties as the dependent variables, tracheid characteristics which have considered annual ring and latewood percentage as the independent variables, ternary cubic models were derived. The highest fitting degree was compression strength as the R2 was 0.717 while the lowest one was bending strength as the R2 was 0.362. The least rela-tive error between measured value and predicted value was the density while the biggest one was the bending strength since the spe-cific error value was 7.59% and 37.63% respectively. According to the regression model, some wood properties can be predicted by the tracheid characteristics.