基于无损检测技术的竹木复合结构材料可靠性分析

The mechanical properties and reliability analysis of bamboo-wood composite structural materials were investigated with nondestructive testing technology combined with wavelet analysis, neural network and time series analysis. On this basis, service life prediction of bamboo-wood composite structure materials was undertaken to realize reliability and safety nondestructive evaluation. Firstly, the multiple nolinear regression analysis model and artificial intelligence model between nondestructive testing signals and dynamic and static mechanical strenght of bamboo-wood composite materials were constructed with the vibration testing and static bending test, in order to obtain the mechanical strength of the material structure by nondestructive testing method. Then, through the mechanical strength indexes and reliability annlysis theory, the numerical simulation of structural materials reliability index was conducted, combined with the random variables and failure limit state equation determined according to the practical condition of bamboo-wood composite structural materials. Finally, time series analysis and neural network technology were employed to implement service life prediction of bamboo-wood composite structure material, combined with the reliability indexes calculated by the precious reliability analysis, by set up the artificial intelligence model. The performance and reliability comprehensive evaluation system of bamboo wood composite structure were built to provide technical support for the optimization design of structure and performance of bamboo-wood composite structure materials, and to provide a theoretical basis for the reliability design of it. The safety and reliability of bamboo-wood composite structure materials in practical application were guaranteed to reduce the occurrence of major safety accidents.

利用无损检测技术结合小波分析、神经网络和时间序列分析对竹木复合结构材料的力学性能进行检测，对其可靠性进行分析，在此基础上对竹木复合结构材料的使用寿命进行预测，从而实现结构材料可靠性和安全性的无损评价。首先利用振动检测、弯曲试验和信号处理构建竹木复合结构材料无损检测信号与材料动态、静态力学强度之间的多元非线性回归模型和人工智能模型，对结构材料力学强度进行无损检测。然后根据实际工况条件确定竹木复合结构材料的随机变量和失效极限状态方程，并利用材料的力学强度和可靠性分析理论对结构材料的可靠性指标进行数值模拟。最后利用材料可靠性指标，结合时间序列和神经网络技术构建人工智能模型对竹木复合结构材料的使用寿命进行预测。从而建立竹木复合结构材料性能与可靠性综合评价体系，为材料结构与性能的优化设计及其可靠性设计提供技术支持和理论依据，保证竹木复合结构材料在实际应用中安全可靠，减少重大安全事故的发生。

大力开发和利用竹木复合材料既可以发挥木材和竹材各自的优良特性，又能缓解木材资源的短缺的现状，还可以分利用丰富的竹材资源和速生林木材资源，但竹木复合材料性能的控制是关键。本研究以竹木复合集装箱底板为例，首先利用振动无损检测方法和图像处理技术并结合小波分析、神经网络建模，研究了竹木复合材料的组坯方式、截面特征参数对竹木复合材料的力学性能和材料集中载荷的影响以及这些力学性能之间的相互关系，结果发现四组预测模型密度与压块指数、顺纹抗弯与顺纹静曲强度、顺纹抗弯与压块指数以及横纹抗弯与压块指数的线性相关性强。利用人工神经网络建立组坯参数预测竹木复合集装箱底板力学性能参数的模型，训练数据的顺纹短跨距剪切力、顺纹MOR和顺纹MOE的R值分别为0.872，0.895，0.884。.研究了基于振动应力波、经典理论、图像处理、剖面密度分布的竹木复合材料力学性能无损检测与评价，发现利用人工神经网络构建模型，纵向和横向试样纵向共振的拟合效果比弯曲振动的拟合效果好；整体而言，横向集装箱底板试件模型比纵向集装箱底板试件模型效果好。当跨高比大于20时,经典理论、一阶剪切变形理论和有限元法对弹性模量和挠度的预测值基本接近，将理论预测值与测试值进行对比，纵向弹性模量预测准确率高于横向，误差分别为﹣1.62%，5.58%以及﹣0.24%，可以用来预测竹木复合材料的力学性能，节省检测成本。剖面密度曲线波峰波谷与顺纹、横纹力学性能的模型预测中，顺纹预测准确率高于横纹，顺纹抗弯强度、静曲强度和弹性模量预测值与实际值之间的拟合系数分别为0.87、0.71和0.80，误差在20%内占比分别为84.8%、70.0%和63.6%。.还利用成组法在四级应力水平下测试竹木复合集装箱底板的疲劳寿命，根据材料应力与寿命数据拟合 S-N曲线和失效概率为参数的P-S-N曲线，发现竹木复合集装箱底板疲劳寿命服从对数正态分布，并通过理论计算得到竹木复合材料不同应力水平下任意一可靠度的疲劳寿命，实现工况条件确定条件下竹木复合材料的疲劳寿命与可靠性分析。