沥青路面动态响应的随机性分析模型构建

Dynamic responses monitoring is an important tool for asphalt pavement performance research. Because of the huge variability, the valuable information of the monitoring data usually can’t be fully utilized. The variability of responses was considered as the abnormal error of the data in past researches. Factually, modern stochastic theory has already been capable to reveal its regulations in nowadays. To mining the dynamic responses data effectively, the deterministic and stochastic characteristics will be analyzed from two aspects, which are spectral analysis and time domain analysis; the stochastic models will be chosen and fitted to simulate the long-term trends for mainly responses, which refers to stress, strain, temperature and moisture. The main steps of this research are: 1) describing the basic statistical characteristics of responses, such as cumulative frequency distribution, means, variances and covariances; 2) implementing Fourier transform for response waves, to estimate the spectrum of deterministic component and stochastic component; 3) decomposing the deterministic component into long-term trends and seasonal trends, and analyzing the models for each of them; 4) performing autocorrelation test for stochastic component, to choose the reasonable regression models and estimate the parameter, and fitting the stochastic trends with the criterion of white noise. The inherent patterns of dynamic responses for asphalt pavement could be explained through the proposed method from a new perspective. The results will be helpful for long-term pavement performance research.

动态响应监测是研究沥青路面使用性能的重要技术手段。受监测数据变异性大的影响，实测响应的丰富信息往往不能被充分利用。以往研究常将响应变异性当作是“纯随机性”简单抛弃，而现代随机理论实际已可对其进行充分挖掘。本研究拟从频域和时域两个方面，分析实测沥青路面响应（应力、应变、温度、湿度）的确定性和随机性特征，构建随机性分析模型，揭示动态响应的变化规律。主要步骤为：1）描述各种响应的频数分布、离散与集中趋势等统计学特征；2）对响应波形进行傅里叶变换，揭示其中的确定性成分及随机性成分的频谱构成；3）将确定性成分分解为长期波动及周期性波动两种趋势，分析其变化规律；4）对随机性成分进行自相关性检验，构建符合其随机特征的随机模型，并以“白噪声”为判断准则，拟合响应的随机性变化趋势。通过上述研究，可以从全新角度揭示沥青路面动态响应的变化规律，为分析沥青路面长期服役行为机理奠定坚实基础。

本项目对道路响应数据进行了统计分析、时域分析及频域分析，从中获知了温度、湿度、应力及应变响应在不同荷载、状态、车速下的响应规律。并从中总结出了响应数据的滤波方法及数据处理与分析的流程与方法。同时，根据长期结构服役行为数据的分析结果，本研究还对响应数据进行了随机性分析。.本项目通过对道路响应数据的随机性分析，获知了道路结构内部随外界响应的变化规律。通过对数据的时频域分析以及不同滤波算法的对比，获得了对响应数据可靠的处理分析流程。为后续的全寿命周期响应数据的分析及道路结构优化设计提供了数据基础及方法支撑。.本项目的主要结论有：由于外界噪声的干扰，道路响应数据在分析前需要进行必要的滤波处理；响应数据的基线存在周期性的变化规律，且在外界扰动较大时会产生随机变化；响应数据的波形、峰值、周期等会受到结构深度、荷载重量、车速等多种因素共同影响。