基于动态弯沉分析的路面结构模量反算方法研究

Back-calculation Research of Pavement Structure Modulus Based on Dynamic Deflection Analysis

落锤式弯沉仪作为一种动态无损检测方法,在世界范围内得到了广泛应用。弯沉盆包含了大量路面结构信息,可据此对结构层模量进行评价。该文利用谱元法路面计算程序,通过逐步回归和多层感知元神经网络 2 种方法,对路面模量反算模型进行了研究,并提出了基于弯沉盆的动态模量反算方法。对半刚性沥青路面的实测弯沉数据进行反算后,发现由于实测和理论弯沉盆不完全相同,因此基于理论弯沉盆拟合得到的回归公式虽然相关性很好,但在应用于实测数据时却无法得到正确结果;而数值验证显示神经网络方法的反算值稳定性较好。反算数据显示沥青混合料模量受温度影响很大,水泥稳定碎石基层模量与龄期有关。通过弯沉反算得到的材料模量值可以为路面结构设计提供参考。

The falling weight deflectometer has been applied worldwide as a non-destructive test for quality evaluation of pavements.The deflection bowl contains a lot of information about the structure and can be used for modulus evaluation.In this contribution two methods were used in the back-calculation procedure,respectively as pace regression and multilayer perceptron network method,based on the program developed from spectral element method.A new dynamic back-calculation model for pavement modulus was developed.Numerical implementations were carried out with measured deflection data of the semi-rigid asphalt pavements.It was found that because of the difference between calculated and measured deflection data,it is difficult to achieve the correct modulus results with formula exactly derived from the theoretical deflection bowls.On the other hand the multilayer perceptron network method was proven to have good stability.The back-calculated results show that the temperature has a great effect on the modulus of asphalt mixture and the stiffness of cement treated aggregate base grows with time.The modulus obtained in this research can provide reference for pavement design.