The Application Study Of The United Prediction Model Of The Volumetric Parameters And Rutting Resistance Performance Of Asphalt Mixtures

In the actual production, When designing the mixing proportion of bituminous mixtures, the gradation, the type of mineral aggregate and the type of bituminous are determinate, but we are not sure that if the asphalt aggregate ratio, the volume parameters and the performance of anti-rutting meet our requests in advance. For instance, if the testing results of the anti-rutting performance couldn't meet our requests after the mixing proportion design of bituminous mixtures was accomplished, it is need to re-adjusting the mixing proportion design of bituminous mixtures by time-consuming and laborious tests. In order to guide the mixing proportion design of bituminous mixtures, the uniform model which can predict the VMA and performance of anti-rutting was established. The following three aspects of the research were undertaken center around the main topic:Firstly, The factors that affect VMA and the performance of anti-rutting are analyzed. The results of tests show that VMA was influenced by aggregate gradation and the type of aggregate greatly; The factors have important effects on the performance of anti-rutting, such as the gradation, the type of aggregate and bitumen, asphalt aggregate ratio and the percent of the voids etc..On this basis, the mathematic formula of the relation between VMA and the percentage of passing of each sieve, compound bulk density of aggregates and compound apparent density of aggregates has better correlation base on the factors that affect VMA by multiple linear regression. And compared to the VMA predicted and by multiple linear regression by using HOUDSON method, There is also a better correlation among the percentage of passing, type of bitumen, compound bulk density of aggregates, compound apparent density, asphalt aggregate ratio and the percent of the voids which can be worked out by multiple regression according to the influence of dynamic stability. The dynamic stability can be predicted by using the constitutional relationship of Burgers model base on the visco-elastic principle of bituminous mixtures.Secondly, the improved back -propagation (BP) neutral network was adopted in this paper. The VMA and the dynamic stability model were trained by some algorithms on MATLAB. we got our prediction data needed by using the generalized ability of neutral network.