Optimization Of 3D Deformation Field Of Concrete Based On DVC

Concrete is a kind of complex multiphase composite material including cement mortar,aggregate and pore crack.Its composition determines the heterogeneity of this complex material structure and the discontinuity of physical and mechanical properties.It is of great significance to study and quantitatively express the nonlinear change of internal micro structure under the condition of concrete bearing and to establish the relationship between micro structure deformation and macro mechanical response.The heterogeneity of concrete materials determines that many methods or techniques for studying the mechanical properties of materials can not fully reflect their localized deformation characteristics.In 1999,the digital volume correlation(DVC)proposed by B.K.Bay and others has become an important technical means to characterize the internal deformation of objects.The original data of concrete is obtained by scanning the uni-axial compression concrete specimen with computer tomography(CT),and then the 3D reconstruction volume image of concrete is obtained by computer reverse reconstruction.The displacement field and strain field of concrete specimen deformation are obtained by combining the digital volume image correlation method to realize the 3D visualization of deformation.Compared with the plane digital image correlation method in displacement,strain accuracy and calculation efficiency,the digital image correlation method faces many problems,and the solutions also have many shortcomings,which need to be improved.This paper puts forward the following improvement measures for the calculation process of three-dimensional deformation field of concrete:(1)The accuracy of displacement field obtained by DVC method is affected by interpolation error and the details,noise and artifacts in the source image.The error of displacement field is usually amplified in the process of strain calculation.Accurate full field strain measurement can help us to understand the mechanical properties of materials in depth.Therefore,this paper proposes an optimized DVC method,which combines the explosion search algorithm(ESA)and the digital volume correlation method to overcome the problem that the traditional DVC method is greatly affected by the initial value of iteration.By quickly matching the coordinates of the estimated points before and after deformation,the initial displacement value can be obtained and the calculation efficiency can be improved.(2)The evaluation function of DVC is optimized globally,and the regularization is added to the objective function to suppress the inherent noise and alleviate the mismatch before and after deformation due to too little detail(high-frequency information)in some regions of the source image.In this paper,a fast iterative digital volume correlation method is proposed,which can effectively calculate the finite deformation field that can not be solved by the general DVC algorithm.The third-order polynomial Gaussian peak value fitting is used to ensure the accuracy of the subvoxel displacement calculation.(3)Based on the theory proposed in this paper,the improved DVC method is simulated by computer controlled three-dimensional speckle experiment,which can simulate imaging and rigid body translation deformation,and verify the accuracy and reliability of the algorithm.Finally,the 3D reconstruction of the data obtained from the uni-axial compression test of concrete is carried out,and the micro displacement measurement of the 3D volume image before and after deformation is carried out by the DVC method proposed in this paper,and the full field strain of concrete is obtained by the appropriate difference algorithm.Compared with the traditional DVC method,the experimental results show that the optimized DVC algorithm guarantees the accuracy and improves the calculation efficiency greatly.