Development Of Color Cement And Performance Of Mixture Based On BP Neural Network

In recent years,the aging problem of asphalt pavement in China has become increasingly prominent,and the initial damage problem needs to be solved urgently.In this regard,our experts and scholars have actively explored,and found that in actual production and application,emulsified asphalt,especially color emulsified asphalt,develops rapidly,and is applied in sealing,permeable and viscous layers,and color emulsified asphalt has a good effect.In the development of color emulsified asphalt,the development of color asphalt（alias color binder）is both the foundation and the key.Based on BP neural network,this paper explores the development of color binder and the performance of its mixture,and finds out some rules.Specifically from the following aspects:Firstly,based on the wide application prospects of color asphalt pavement at present,through reading a large number of documents and patents,this paper carries out trial-and-error tests continuously.And on the basis of the predecessors to improve,blend out their own formula,improve their own preparation process.Four kinds of raw material components were selected:base oil A,resin B,modifier C and decolorizer D.Then,by inquiring the mechanism of polymer miscibility reaction,combining with the test operation,through continuous tests,the problems needing attention in the test process were summarized.At the same time,some evaluation indexes of color binder at home and abroad are summarized.Secondly,L₁₆（4⁵）orthogonal table was used to arrange the experiment,in which base oil component A,resin component B and modifier component C were selected at four levels.For the convenience of visualization in the later stage,decolorant component D was fixed horizontally.At the same time,BP neural network is used to fit and genetic algorithm is used to optimize the model.Taking the three major indexes of color binder（softening point,penetration and 5 C ductility）as the research objective,80%data training network is randomly selected from the orthogonal test table,20%data testing network fitting performance,preserving the trained network,drawing relevant images,fitting the functional relationship between each component and index,and correlating them with each other.Systematic four-dimensional visualization can grasp macroscopically the change rule of cementitious material performance index under different component content.Ten new groups of group ratios（three groups of genetic algorithms to optimize the recommended group ratios,seven groups not included in the orthogonal test group ratios）were added to carry out the experiment.The actual data of the indicators obtained were compared with the network prediction data to verify the accuracy of the model prediction established in this paper.Finally,based on the principle of BP neural network,the software of predicting asphalt index by neural network is developed to predict asphalt index more intuitively.Thirdly,the mixtures of four kinds of color binders Cb-A,Cb-B,Cb-C and Cb-D developed by ourselves are designed,and four kinds of color asphalt mixtures CAM-A,CAM-B,CAM-C and CAM-D are prepared,referring to"Test Rules for Asphalt and Asphalt Mixture in Highway Engineering"（JTG E20-2011）,"Color Asphalt Concrete"（GB/T 32984-2016）and"Color Asphalt Mixture".The color asphalt binder（JT/T 1128-2017）verifies the applicability of the self-developed color binder to mixtures.In addition,taking CAM-D as an example,this paper emphatically explores the changes of relative density of gross volume,void fraction,void fraction,void fraction,void fraction,saturation,stability and flow value under different pigments（red,yellow,green and blue）,different ratio of pigments to raw mineral powder（the ratio of pigments to raw mineral powder is 15%as the starting point and 5%as the incremental difference is increased to35%）.Law.Finally,this paper explores the dynamic stability of color asphalt mixture under different pigments and cements through road performance tests of 20 kinds of color asphalt mixture（high temperature stability,low temperature performance,water stability）in five colors（red,yellow,green,blue,colorless）.The variation of flexural strength,maximum flexural strain,modulus of flexural stiffness,residual stability and splitting strength ratio.