考虑层间接触和材料非线性的道路参数反演

Road Parameter Inversion Considering Interlayer Contact and Material Nonlinearity

为进一步完善道路结构评价体系,现充分考虑道路材料非线性特征和层间非连续接触状态对道路参数反演的影响,建立了FWD荷载作用下的道路动力有限元模型,构建了路面结构参数-路表弯沉盆数据库。在对数据库进行加噪处理的基础上,应用神经网络模型和遗传算法优化神经网络模型,实现了弯沉盆参数指标与路面结构参数的映射和学习,最终达到道路结构参数反演的目的。结果表明:在道路响应计算中,有必要考虑材料非线性和层间接触状态的影响;模型对路面模量和路基非线性参数的反演效果较好,但对层间摩擦因数的反演效果不理想;加入噪声处理,可使模型获得良好的鲁棒性和泛化能力;相比于神经网络模型,遗传算法优化神经网络模型的反演性能更优,具有较高的计算效率和反演精度。

In order to further improve the road structure evaluation system,this paper fully considers the influence of the nonlinear characteristics of road materials and the discontinuous contact state between layers on the inversion of road parameters.Finite element model under FWD load is established,and database of pavement structure parameters-deflection basin was constructed.On the basis of denoising,neural network model and genetic algorithm are used to optimize the neural network model to realize the mapping and learning of the deflection basin parameter index and the pavement structure parameter,and finally achieve the purpose of inversion of the road structure parameter.The results show that it is necessary to consider the influence of material nonlinearity and interlayer contact state in the calculation of road response;The inversion effect of the model on pavement modulus and subgrade nonlinear parameters is good,but the inversion effect on interlayer friction coefficient is not ideal;Adding noise processing can make the model have good robustness and generalization ability;Compared with the neural network model,the genetic algorithm optimized neural network model has better inversion performance,higher computational efficiency and inversion accuracy.