| Along with the development of road construction, it becomes more important to the nondestructive evaluation of pavement structure. Now, the inverse analysis methods are applied in the nondestructive evaluation. Based on the study of the domestic and overseas inverse analysis of pavement, the paper brings forward the new methods of the inverse analysis .Its main contents include:1. According to the method of system identification, this paper introduces improving genetic algorithm into the inverse analysis methods. The best solution for the inverse analysis is gained. Genetic algorithm can resolve the problem that initial value influences the result of the inverse analysis, which is hard in the traditional inverse analysis.2. The time of the inverse analysis can be sharply shortened, if the neural network is applied in the inverse analysis. Because the process of inverse analysis is very complex, it is difficult to get good training result by using traditional 3-layer BP network. The paper improves the network by the ways of preprocessing and post-processing and uses the network with 2 middle layers. The examples show that the inverse analysis can get satisfactory result by using neural network.3. Falling Weight Deflectomoter produces the impulse load to pavement, but the mechanics model of current inverse analysis is the assumption of static linear elastic behavior of pavement. The paper establishes the dynamic mechanics model with Ansys software and analyses the effect of dynamic load on the results of the inverse analysis.4. The paper studies how the variations of the parameters of pavement influence the result of the inverse analysis by using Monte Carlo method, and discusses the effect of the distress conditions of pavement on the result of the inverse analysis. |
