基于人工神经网络的路面模量反算

Backcalculation of pavement layer moduli based on artificial neural networks

1.
Department of Highway and Traffic Engineering, Changsha Communications University, Changsha 410076, China
2.
School of Highway, Chang′an University, Xi′an 710064, China

摘要: 根据三层BP神经网络模型和弹性层状体系理论, 结合JILS FWD研究了层状体系路面的模量反算。通过理论和实测弯沉盆的反算, 比较了精确网络与噪音网络的反算能力, 从而提出了人工神经网络实现模量反算的关键技术。噪音网络与国内外常用反算程序的比较结果表明, 神经网络法的反算结果具有良好的精度和可靠性
- 路面 /
- 模量反算 /
- 人工神经网络 /
- 弯沉盆 /
- 落锤式弯沉仪
Abstract: The backcalculation of pavement layer moduli for layered system is researched with JILS FWD according to the model of 3 layered BP neural network and the theory of elastic layered system.The backcalculating ability of accurate network and noise network is compared through the backcalculation of theoretical and measured deflection basins. The key techniques are presented for the backcalculation of moduli with artificial neural networks.The comparisions of the noise network to the common backcalculation programs show that the backcalculation results for the method of neural networks are good in accuracy and reliability.
- pavement /
- backcalculation of moduli /
- artificial neural network /
- deflection basin /
- falling weight deflectometer

图 1 三层BP网络结构

Figure 1. A three layer BP network architecture

下载: 全尺寸图片幻灯片

表 1 弯沉盆数据

Table 1. Data of deflection basins

下载: 导出CSV

表 2 精确网络反算模量结果

Table 2. Results of backcalculation of moduli with the accurate BP network

下载: 导出CSV

表 3 模量反算结果比较表

Table 3. Comparisons of results for backcalculation of moduli

下载: 导出CSV

[1]	Harichandran R S, Mahmood T, Raab A R, Baladi G Y. Modified newton algorithm for backcalculation of pavement layer properties[J]. Transportation Research Record 1384, TRB, National Research Council, Washington, D.C., 1993.
[2]	Scullion T, Uzan J, Paredes M. MODULUS: a microcomputer-based backcalculation system[J]. Transportation Research Record 1260, TRB, National Research Council, Washington, D.C., 1990.
[3]	Zha Xudong, Zhang Qisen. Study on backcalculation for pavement moduli using genetic algorithm[A]. Proceedings of the the Third International Conference on Road & Airfield Pavement[C]. Beijing: 1998.
[4]	Meier R W, Alexander D R, Freeman R B. Using atrificial neural networks as a forward approach to backcalculation[J]. Transportation Research Record 1570, TRB, National Research Council, Washington, D.C., 1997.
[5]	John Hertz, Anders Krogh, Richard G Palmer. Introduction to the theory of neural networks[M]. Canada, Redwood City: Addison-Wesley Publishing Company, 1991.
[6]	查旭东. 基于同伦方法的路面模量反算的研究[D]. 西安: 长安大学, 2001.