| Thus wood plastic composite (WPC) is an environmental, economical, low carbon and recycle material, it has become a new and important industrial material, which has widely used to process the products with low additional values in the past but now manufacture the products with high additional values, such as building, house, pipe and floor. The research on the vibration properties and nondestructive evaluation of WPC can provide the scientifical basis for the dynamic property assessment and inspection of structure based on the WPC. The HDPE based WPC was employed to as the research objects in this study, and the vibration properties of WPC specimens such as intrinsic frequencies, dynamic modulus of elasticity (MOE), displacement and curvature mode shape were studied systematically, using the analysis of theoretical, experimental and finite element simulation respectively. Then, BP neural networks was adopted to qualitatively and quantitatively identify the inner hole defect in WPC specimens.The modal analysis theory for studying the vibration properities of WPC was described firstly in this paper. Based on this, the vibration modal experiment was conduct to the HDPE based WPC specimens, intact and defective respectively, for identifying the parameters of frequency response function (FRF), intrinsic frequencies, mode shape and dynamic MOE. Then, the finite element models of the intact and defective WPC were built using ANSYS software to make the modal analysis and validate the experiment results. The BP neural networks for qualitatively and quantitatively identifying the inner defect in WPC specimens were constructed and trained based on the intrinsic frequennncies, displacement and curvature mode shape respectively. At last, the changes of intrinsic frequencies and dynamic MOE of WPC specimens at normal and frozen states were tested and discussed.The research results show that:(1)The intrinsic frequency and mode shape are two important parameters for respresenting the vibration properties of WPC. The intrinsic frequencies of WPC are significant affected by internal defect, and the intrinsic frequencies of defective WPC are lower than that of intact WPC. The displacement mode shape is not sensitive to the small defect or damage in the WPC specimens, but the curvature mode shape is a sensitive parameter to the size and position of local damage or defect. (2)The dynamic MOE of WPC tested by longitudinal wave, longitudinal vibration and flexural vibration methods are higher than the static MOE, the ratios of dynamic MOE to static MOE are 3.02,2.94 and 2.83 respectively. The dynamic MOEs of WPC are significant affected by internal defect, which of intact specimens are higher than that of damaged specimens. In addition, central damage has more significant effect on the dynamic MOEs than other position damage. (3)Through the training and verification of BP neural networks, three parameters of intrinsic frequennncies, displacement and curvature mode shape are proved to effectively identify the location and size of local defect.(4)The intrinsic frequennncies and dynamic MOE in frozen WPC specimens are higher than that in WPC specimens at normal temperature. |