Research On Two-Dimensional Full-Field Deformation And Strain Measurement Technology

With the development of instrumentation science and experimental mechanics,full-field displacement and strain measurement methods have also been advanced.Among them,digital image correlation method is an emerging optical measurement method,which is widely used in various disciplines and engineering fields because of its advantages such as simple device,wide measurement range and low environmental requirements.However,the search speed and accuracy of digital image correlation methods are still lacking in some specific scenarios,so it needs to be further studied and improved.Based on the existing theories of digital images correlation,this paper proposes improvement methods for the shortcomings.The main research work and conclusions of this paper include:(1)To solve the problem of slow integer pixel search,the paper proposes a method to accelerate integer pixel localization by using the grayscale difference of central pixel points,and demonstrates that the method has higher computational efficiency than traditional integer pixel search through simulation experiments.The speckle quality is evaluated by considering the gray gradient and entropy of the image,and the method of one-parameter relative change ratio is used to refine the analysis.An adaptive subregion selection method based on MIG and subregion image entropy is proposed,which realizes the selection of equivalent subregion sizes of different speckle patterns,and the simulation results show that the calculation precision of the algorithm is higher than that of fixed selection of subregion size.(2)The Inverse Composed Gauss Newton algorithm(IC-GN)relies on accurate initial values,and this paper proposes a method for estimating the initial values of IC-GN based on genetic algorithm and swam optimization algorithm(GA-PSO).Firstly,three valuation points are arbitrarily selected around the calculation point,then the deformation of the valuation point is searched by the GA-PSO algorithm,and then the coordinates before and after the deformation of the valuation point are substituted into the affine transformation matrix to find the deformation coefficient of the calculation point,which is used as the iterative initial value to solve for the subpixel displacement in the IC-GN matching method.The algorithm is compared with the conventional IC-GN through simulated non-linear deformation and large deformation experiments.The results show that the algorithm is more accurate and the time of calculation is reduced by 51.04%.The three-point bending experiment of pipeline steel verifies the effectiveness of the algorithm in practical application.(3)The full-field displacement and image sequences displacement solution methods are studied,and a seed point diffusion method based on breadth first search method of graph is proposed,and simulation experiments verify the accuracy and speed advantage of the method compared with multi-point cumulative calculation.A segmented measurement method combined with path-independent DIC(PiDIC)is proposed to avoid the error transfer of image sequences displacement calculation,and the stability of the method is verified by the experiment of hole materials displacement measurement.The influencing factors of strain accuracy are analyzed,and it is pointed out that different strain window sizes should be selected in the two modes of uniform deformation and non-uniform deformation.(4)Based on the algorithms of this paper,the software interface is developed,the hardware platform is built,and a two-dimensional speckle strain measurement system is designed,which is verified by tensile experiment of aluminum alloy plate.The experimental results show that the maximum relative error between the strain obtained by the measurement system and the measurement result of the extensometer is 0.73%,which proves the reliability of the system in practical measurement.