Data Mining Of Asphalt Pavement Structural Information Monitoring

Field measurement research of dynamic responses in asphalt pavement structure is one of the important means to clarify the damage mechanism of asphalt pavement in complex environment. With the rapid development of sensor monitoring technology and computer technology, long-term monitoring of the asphalt pavement structure and a large number of data accumulation become possible. Many monitoring project had accumulated a lot of pavement dynamic responses data and environmental data. However, the existing research focuses on the regularity description of monitoring data, failed to use huge amounts of data.The paper introduced the theory of data mining to the field of asphalt pavement structure information monitoring to analysis the pavement structure information data effectively. The paper researched mass data processing method, analyzed the dynamic responses of asphalt pavement structure under complex environment and forecasted the performance of asphalt pavement based on the data mining. The main research contents and results are as follows.First of all, the common methods to collect and store the asphalt pavement structure monitoring data were analyzed. The massive data processing methods, such as the way to determine key information extraction, data filtering, valid data determining were researched. The application of data mining in road engineering and other related fields were analyzed, such as bridge monitoring, tunnel monitoring and slope monitoring. The data mining concept about asphalt pavement monitoring data was put forward. The suitable methods were researched to mining the huge amounts of asphalt pavement structure information monitoring data.Secondly, mining the temperature data combined with numerical simulation method to study the dynamic responses of asphalt pavement structure in actual temperature field. An asphalt pavement structure temperature field analysis model of finite element was established. Temperature field numerical simulation method based on the measured temperature data at the key position of asphalt pavement structure was researched. The results were compared with the measured temperature to verify the effectiveness of temperature simulate method. An asphalt pavement structure dynamic analysis three-dimensional finite element model was established using ABAQUS software. The dynamic responses were calculated and compared with the responses of isothermal model to analysis the limitation of isothermal model in the mechanical analysis of asphalt pavement. The key position of asphalt pavement structure information monitoring was researched based on the early research experience, line position of the wheel track, field investigation and numerical simulation.Thirdly, time series method was used to study the strain influence factors. The axle load spectrum was researched according to the data get by dynamic weighing system. The relationship between the largest vertical strain and axial load was established according to the correlation method. The traffic volumes in the design year were classified to several grades according to the shaft load. Considering the parameters such as temperature, traffic volume and axle load, the rut forecast method based on the measured residual vertical strain was established. An indoor rutting test was designed, and a three layer composite plate was made with optical fiber grating strain sensors in it. Measuring the vertical residual strain and rut deformation under certain load times to verify the rut forecast model based on residual vertical strain.Finally, the statistic distribution of the measured longitudinal strain and transverse strain and the decay law of responses were researched. The responses were predicted using the time series autoregressive model. Looking for the method to predict the pavement performance based on the responses decay analysis.The paper focused on the mining research of huge asphalt pavement structure information monitoring data. The dynamic responses of asphalt pavement under the true temperature and the influencing factors of strains were researched with the data mining method, as well the way to predict the pavement performance. The research improves the utilization of asphalt pavement monitoring data.