Research On Microstructure Characterization And Reconstruction Of Multiphase Heterogeneous Materials

Microstructure Characterization and Reconstruction(MCR)of materials can be used to establish the Processing-Structure-Property(PSP)relationship,which has become the research focus of computational materials science and material design.Multiphase heterogeneous materials are widely used in aviation,electrical,geology,medicine and other fields.Using MCR method for this material can obtain a large number of statistically equivalent models,which provides an effective way to construct the PSP relationship of multiphase heterogeneous materials.This article takes aero-engine aluminum silicon-polyster(AlSi-PHB)sealing coating as the research object.The content of AlSi in the coating is51.5%?52.5%,and the total content of PHB and pores is 47.5%?48.5%.The rate is 1% to 7%,and the PHB and pores are randomly distributed on the AlSi matrix.The coating composition has a wide range of phase sizes,irregular morphology,rough boundaries,and random distribution.It also contains a unique honeycomb network structure,and its microstructure characterization and reconstruction are difficult to characterize and reconstruct.In this paper,three characterization and reconstruction methods of random field,texture synthesis,and physical descriptor are carried out for AlSi-PHB seal coating.The improved random field method has a large error in the autocorrelation parameter of the reconstructed image,and solves the problem of the texture synthesis method.Accurately control the problem of reconstructed image components,and propose a new MCR method by combining the physical descriptor method and the texture synthesis method.Finally,the two-point correlation function and the two-point cluster correlation function are used to characterize and evaluate the reconstruction model.The main research results are as follows:(1)For the random field method,threshold segmentation is used to convert continuous random media into discrete random media,which leads to the problem of errors between the autocorrelation parameters of the reconstructed image and the actual parameters.By iteratively optimizing the autocorrelation parameters a and b,a multi-phase heterogeneous model with accurate parameters and high microstructure consistency is reconstructed.In order to verify the effectiveness of the improved method,the random field methods before and after the improvement were used to perform characterization and reconstruction under the same autocorrelation parameters a and b.Compared with the original image,the results show that the method before and after the improvement reconstructs a and b of the image.The relative errors are 39.0% and 40.1%,respectively.The improved relative errors are reduced to 8.6%and 0%,respectively,which proves that the improved random field method can effectively and accurately reconstruct the multiphase heterogeneous model.(2)Aiming at the problem that the texture synthesis method cannot control the components of each phase of the reconstructed image,the contour recognition algorithm is used to search the edges of each phase and adjust the pixels to achieve component control and iterative optimization,and a texture synthesis method with controllable components is proposed.The research shows that compared with the original image,the composition error of AlSi phase in the reconstructed image before optimization method is 3.2%,PHB phase is8.2%,and the porosity is 5.0%.After optimization,the errors of each phase are reduced to within 0.01%.It shows that the texture synthesis method proposed in this paper can realize the accurate reconstruction of the composition of heterogeneous multiphase materials.(3)In view of the limitation of the physical descriptor method that it is difficult to reconstruct the microstructure with high content and irregular morphology,combining it with the optimized texture synthesis method,it is proposed to be suitable for multi-phase heterogeneous AlSi-PHB sealing.Strictly coated MCR new method.The physical descriptor method is used to reconstruct the pores with low content and relatively regular geometric structure,and the optimized texture synthesis method is used to reconstruct the PHB phase with high content and complex morphology,and the final reconstructed image is obtained after combination.This method combines the advantages of the two MCR methods of physical descriptor and texture synthesis,and can accurately reconstruct multi-phase heterogeneous materials on the basis of physical meaning parameterization.(4)Use the two-point correlation function and the two-point cluster correlation function to evaluate the MCR results.First,evaluate the consistency of the reconstruction results of each method.For the reconstruction of the pore phase,the physical descriptor combined with the texture synthesis method is compared with the random field method.For the PHB phase,the random field method is better but it is difficult to reconstruct its aggregation characteristics.Secondly,the parameterized evaluation of the size,uniformity and periodicity of the different phases in the reconstruction results.Studies have shown that the texture synthesis method can accurately describe the aggregation and distribution of PHB phases,and the combined new method is better than the random field method in describing the pore phase as a whole.The random field method is relatively simple to control the composition of the reconstructed object,and is suitable for studying the relationship between the content of each phase of the coating and the overall performance.The texture synthesis method is better for modeling the complex network structure of PHB phase,and is suitable for studying the influence of PHB relative to coating properties.The physical descriptor combined with texture synthesis method is more suitable for constructing the relationship between pore size and quantity and other structural features and coating performance.