Measurement And Characterization Of Coarse Aggregate Morphology Based On Predicted Void Content

As the main component of concrete,coarse aggregates are in great demand.A large number of domestic and foreign researchers have found that the close accumulation of coarse aggregates will affect various properties of concrete,and their accumulated density is mainly affected by the particle size(grading)and particle shape of themselves.At present,the commonly used grading measurement method is sieving.Coarse aggregate samples are screened step by step through various standard sieves to determine the proportion of each particle size.This method is very time-consuming and inefficient,and particle shape detection cannot be carried out when measuring coarse aggregate grading.With the development of computer and image recognition technology,online detection of coarse aggregate particle size and shape has become feasible.In this paper,the coarse aggregate images collected in different states are studied:For coarse aggregates with ideal distribution,a detection system based on image method is designed to measure the particle size and shape of coarse aggregate.The three-dimensional measurement of irregular aggregate is realized by adding a linear laser.A regression model of coarse aggregate morphologies and void content is established by using the method of multiple linear regression analysis,and it is proved that there is a significant relationship between them by F-test.Through the verification test,it is found that the error between the predicted and the real void content is less than 0.5%,and the error is small.The rapid prediction of void content can be realized by using the particle size and shape of coarse aggregate.The significance of particle shape parameters is verified by partial regression coefficient t-test,and the most suitable particle shape characterization methods are determined,including Maximum projected sphericity,Angularity,Flatness and Elongation.For the coarse aggregate in the stacking state,the segmentation results of traditional segmentation method and deep learning algorithm are compared and analyzed in this paper.In the segmentation of coarse aggregate with different particle sizes,the gradation proportion of coarse aggregate with particle size of 5 ～ 10 mm is increased by 5.93% compared with the result of watershed algorithm,while the proportion of coarse aggregate with particle size of 10 ～ 20 mm and 20 ～ 31.5mm is increased by more than 20%;In the coarse aggregate segmentation of different materials,compared with watershed algorithm,the proportion of cumulative gradation in the coarse aggregate image segmentation of limestone,granite and cobble is increased by 28.62%,37.68% and 12.78% respectively;Compared with the coarse aggregate segmentation of different gradation,the calculation result of deep learning algorithm is closer to the screening method.In order to verify the adaptability of the developed deep learning segmentation algorithm in practical engineering,this paper applies the coarse aggregate online detection system based on deep learning to commercial concrete mixing plant and cone crusher.After segmenting the complex stacked aggregate image collected on site,the effect is good,which proves that the deep learning algorithm can adapt to complex engineering conditions.The work of this paper is helpful to quickly predict the void content through the morphological parameters of coarse aggregate and determine the appropriate morphological characterization parameters.The practical engineering application results of the detection system can prove that this paper has good theoretical research significance and practical application value.