Back-Calculation of Subgrade Modulus Based on Genetic Algorithm and Viscoelasticity Theory
To improve the applicability of the PFWD detection method for the back-calculation of subgrade modulus,taking the subgrade splicing project of the west loop line in Nairobi as an example,this paper adopted 3D Kelvin viscoelasticity model and quasi-static response analysis to deduce the displacement function of viscoelastic subgrade under long-term fixed loads,and combined the principle of liner superposition of viscoelasticity with genetic algorithm to propose a new method based on multi-population genetic algorithm for the modulus back-calculation of viscoelastic subgrade.Among them,the corresponding program for back-calculation was written by Matlab.The comparison between theoretical derivation and field tests shows that the peak value of vertical displacement on the top of subgrade under PFWD loads lags behind the load peak value.However,selecting only the displacement and load peak value for back-calculation,the traditional method based on the elastic half-space model yields a calculation result 20% larger than the result by the new method.At the same time,the correlation coefficient between the displacement time-history curve calculated by the new method and the real curve is basically greater than 0.9,and compared with the results by other methods,the calculation difference is about 10% .Moreover,with fast convergence speed,good adaptability,and accurate back-calculation results of subgrade modulus,the new method can be the reference for subgrade design and construction quality control.
NETL
NSTL
万方数据
