Pavement Material Properties Research Based On Gene Expression Programming Algorithm

As the important basis of pavement structure design and construction, the pavement material performance research has always been the hot spot of the direction of the road engineering research. Based on gene expression programming algorithm for dynamic modulus of asphalt mixture, the autogenous shrinkage of concrete, the dynamic modulus of recycled asphalt mixture pavement material performance research, analysis of the results obtained have important practical implications for road engineering, structural engineering and other fields.This paper systematically analyzes the asphalt dynamic modulus, concrete autogenous shrinkage, recycled asphalt mixture dynamic modulus research status. Then, based on the previous research,by using the basic theory of gene expression programming algorithm, for asphalt dynamic modulus, concrete autogenous shrinkage, recycled asphalt mixture dynamic modulus were more in-depth analysis. The main research contents can be summarized as follows:(1) Using gene expression programming algorithm to predict asphalt dynamic modulus, with eight main factors of asphalt dynamic modulus: air void(aV),effective binder content(b effV), viscosity of binder(?),loading frequency( f),per cumulative retained on The aggregate gradation variables include percent retained 19 mm sieve(3 4?),percent retained 9.5mm sieve(3 8?),percent retained 4.75 mm sieve(4?) and percent passing 0.075 mm sieve(2 0 0?), which constitute the main parameters to predict dynamic modulus of asphalt mixture model. Gene expression programming algorithm can create a dynamic modulus of asphalt mixes prediction model by discrete 8 parameters. The results show that between the dynamic modulus predicted and measured values has obtained a high correlation and by compared with Witczak1999 function model, Korea dynamic modulus prediction model, function model and artificial neural network model, gene expression programming algorithm prediction model has some of superiority than other models.(2) Gene expression programming algorithm research dynamic modulus of recycled asphalt mixture prediction model. Based on the results of the hot mix asphalt dynamic modulus, with reference to adopt Witczak 1999 Asphalt dynamic modulus prediction models eight factors, additional recycling asphalt mixture ratio doped, using as recycled asphalt mixture dynamic modulus of nine input parameters. Using gene expression programming algorithm create a dynamic modulus of asphalt mixes prediction model by discrete 9 parameters. The goodness of fit is analyzed between the predicted and measured values, and the correlation between dynamic modulus and mixing ratio. The sensitivity of between each influence factors and dynamic modulus is analyzed. The results showed that: there is a better fitting between predicted values and measured values, prediction model of dynamic modulus of recycled asphalt mixture material has a certain reference value.(3) From autogenous shrinkage of concrete test, research different water cement ratio and dosage of silicon powder on the effect of autogenous shrinkage. Application of gene expression programming algorithm of concrete self relation between shrinkage and main influence factors(water cement ratio, mineral admixtures, aggregate content, cement paste content, high efficiency water reducing agent, curing temperature and curing age), get autogenous shrinkage of concrete prediction model, analysis the autogenous shrinkage predicted and the measured values fitting, and analysis the main influence factors and the correlation between the autogenous shrinkage. Studies have shown that: Autogenous shrinkage of concrete on the accuracy of prediction models meet the requirements of engineering design, for the design of concrete structures have some significance.Finally, in a comprehensive summary of the full-text basis of their work, made some suggestions and prospects for further research.