Research On Mix Design And Performance Prediction Of Composite Recycled Powder Mortar Based On Artificial Neural Network

The Communist Party of China’s 20th National Congress report made the suggestion to"promote the greening of economic and social development" after the nation recommended the construction of the " double carbon" aim.Accelerate the construction of waste recycling system,which pointed out the direction for the establishment of solid waste recycling system in China.With China’s modernization progressing,urban renewal and industrial production have caused the production of solid waste to increase year by year and it is difficult to consume,bringing various environmental pollution problems.Waste red brick and coal gangue are important components of solid waste,and their open-air stacking will cause great harm to water and soil.Because the red brick has the characteristics of large crushing value and high water absorption,it is not suitable for use as coarse aggregate in concrete.Therefore,the red brick fine aggregate produced by waste red brick could be used in the production of recycled mortar.Because the physical and chemical properties of coal gangue are greatly affected by the location of the ground,and the waste red brick and coal gangue have potential pozzolanic effect,so this study prepares the waste red brick and coal gangue into red brick recycled powder and coal gangue recycled powder by mechanical grinding.Under a certain substitution rate and compound mixing ratio,the composite recycled powder is used as an auxiliary cementitious material to produce recycled mortar.Combined with the compressive strength test data of composite recycled powder mortar,the artificial neural network technology was used to accurately predict the compressive strength of mortar.Finally,the optimal compressive strength of composite recycled powder mortar was used as the optimization target to find the optimal mix ratio of mortar.This study illustrates the difference of technical indexes between red brick fine aggregate and recycled fine aggregate.Compared with recycled sand red brick fine aggregate,red brick sand has sharper and more irregular particle shape,lower density and larger crushing index.With the incorporation of red brick fine aggregate into recycled mortar,the basic properties of recycled mortar have also changed to some extent,but the impact is small.On the basis of satisfying the consistency and fluidity of recycled mortar,the compressive strength of recycled mortar produced by appropriate amount of red brick fine aggregate has been improved.This is because red brick fine aggregate has internal curing effect and volcanic ash effect of brick powder.Based on studying the influence of red brick sand on the performance of recycled mortar,this study explores the performance of composite recycled fine powder mortar composed of red brick recycled fine powder and coal gangue recycled fine powder.The characteristics of physical and chemical red brick recycled powder and coal gangue recycled powder were tested.The results show that the particle size of coal gangue recycled powder is finer than that of red brick recycled powder,which indicates that coal gangue has better grindability.The microscopic morphology test of the particles shows that the shape of the red brick recycled powder particles is relatively sharp and rough,mainly in the form of blocks and flakes,while the coal gangue recycled powder is oval.The test proves that the red brick recycled powder has a certain ability to react with volcanic ash,and the coal gangue recycled powder can fill the micro aggregate.When the ratio of red brick recycled fine powder to coal gangue recycled fine powder is 5:5,the mechanical properties of recycled mortar are the best.This is because the synergistic effect of two kinds of recycled fine powder can be maximized under this ratio.The compressive strength of mortar decreases with the increase of the replacement rate of composite recycled powder.When the replacement rate of red brick sand is 20%,the compressive strength of recycled mortar is the best,and the activity of composite recycled powder needs a long curing time to play,and the later strength of recycled mortar is better.The variation of flexural strength of recycled mortar is like that of compressive strength.Through the drying shrinkage performance test,hydration product microstructure test and three-dimensional digital image correlation measurement system(DIC)test of recycled mortar,The drying shrinkage rate of mortar is found to decrease with an increase in the replacement rate of composite recycled powder.The micro-morphology test of hydration products shows that the composite recycled micro-powder with a ratio of 5:5 under a certain substitution rate can make the microstructure of mortar uniform and dense,so the composite recycled micropowder can be used to inhibit the cracking of mortar during curing.DIC test shows that the composite recycled powder will not change the failure mode of mortar,but will aggravate the fracture failure crack propagation.This study uses artificial neural network technology to accurately predict the compressive strength of composite recycled micro-powder mortar under various mix ratios and ages,based on the study of the compressive strength of recycled mortar by composite recycled powder coupled red brick fine aggregate,and based on the GA-BP double hidden layer prediction model,the neural network genetic algorithm function extremum optimization research is carried out.The research proves that the BP neural network prediction model has certain fitting and generalization ability for the compressive strength of recycled mortar,but the error is large.Therefore,this paper introduces the method of increasing the number of hidden layers and genetic algorithm to optimize the BP neural network prediction model.The results show that the statistical parameters of the GA-BP double hidden layer prediction model are reduced by more than 70%compared with the BP neural network prediction model,and the stability and efficiency of the prediction error are improved,with strong prediction ability and generalization ability.Based on the trained prediction model,this study carried out the neural network genetic algorithm function extremum optimization research,and obtained the mix design of red brick fine aggregate and composite recycled powder under the optimal condition of recycled mortar compressive strength.Compared with the test results in the third chapter,it is proved that the optimization results are consistent with the actual test conclusions,which provides a more complete theoretical support and a more novel design scheme for the application of artificial neural network technology to the recycled mortar produced by waste red brick and coal gangue,and offers significant significance for engineering applications.This study takes recycled mortar as the research direction,and aims at the resource utilization of various solid wastes.It provides a theoretical basis for the application of waste red brick fine aggregate and composite recycled powder in recycled mortar,and provides a theoretical and practical basis for the establishment of genetic algorithm to optimize the prediction of compressive strength of BP neural network recycled mortar.The accuracy and reliability of neural network genetic algorithm function extremum optimization applied to the optimization of recycled mortar mix ratio are verified.This study provides a novel practical scheme for realizing the comprehensive utilization of various solid wastes and realizing green development.