Study On Backcalculation Of Pavement Layer Stress Nonlinear Properties

There has been about 20 years for the study of pavement layer properties backanalysis based on Falling Weight Deflectometer (FWD) deflection testing, from which some researches have been used progressively in actual engineering. At the same time, most of present backanalysis theories are based on multi-layered linear elastic hypothesis and quasi-static analysis method, which nonlinear characteristic and dynamic effect are ignored. At present, multi-leveled load regression analysis is used to reflect the nonlinear performance approximately, but biggish error may be introduced for great difference between the method and actual pavement layer properties. So it is necessary to develop dynamic backanalysis method by simulating nonlinear characteristic rationally. In this paper, pavement layer properties backanalysis is studied systemically integrated with definite optimization technique. The main research results are as follows:1.The nonlinear least square optimization technique is improved to offer better performance on stability and speed of convergence. Also, it is utilized in the backcalculation of pavement layer properties.2.Three nonlinear constitutive models including volume stress model, k1 ?k3 model and bilinear model are chosen to simulate different materials. Then quasi-static deformation and stress under FWD load is computed. Furthermore, nonlinear static backanalysis theory of pavement layer properties combining with nonlinear least square optimization technique is put forward. Through backanalysis of theoretic and actual deflection basin, rationality and numerical stability are proved. Thus, the backanalysis method of this paper can give satisfactory result in identifmg pavement layer nonlinear properties.3.Dynamic equilibrium equation of pavement structure response under FWD pulse load is established by means of volume stress model and bilinear model. Surface deflection history is obtained through step-by-step integration procedure. Nonlinear dynamic backanalysis theory of pavement layer properties is researched in which improved nonlinear least square optimization technique is integrated. Analysis of examples demonstrates that more rational and reliable result can be reached for nonlinear characteristic and dynamic effect being considered simultaneously.4.FWD deflection testing and linear static, nonlinear static, linear dynamic and nonlinear dynamic backanalysis are conducted on each layer of one selected pavement during construction. Actual application indicates that in some cases linear static backanalysis may result in unreasonably values, whereas nonlinear dynamic backanalysis can give rational result. Finally, applicability of each method is analyzed, which can be used as reference and foundation in actual application.