Research On Pavement Performance Evaluation And Prediction Of Beijing Expressways Based On Data Mining

Asphalt pavement in China exhibits many patterns of decay with the increase of road service life and traffic load.In China,expressway maintenance mileage has recently increased,and the task remains challenging.Over time,China began to pay attention to adopting preventative maintenance technology in order to delay pavement deterioration and minimize maintenance costs.At the same time,reliable pavement performance prediction and evaluation based on local features is an important aspect of expressway maintenance management.On this basis,data on pavement structure,pavement performance indexes,and the general state of expressway preventative maintenance technologies in Beijing are gathered.The decay features of pavement performance indexes,the efficiency of preventative maintenance,and the evaluation and predict of complete service performance in Beijing are investigated using data mining and mathematical-statistical analysis.The following is the specific research:Firstly,data on pavement structure,performance index,and preventive maintenance are gathered for the years 2016 to 2020.The pavement thickness characteristics,pavement performance indices decay characteristics,and main preventative maintenance technologies of Beijing expressways are summarized utilizing mathematical-statistical methods.Secondly,the scope of performance indexes suitable for preventive maintenance in Beijing is suggested based on changes in four performance indicators following three types of preventive maintenance.On this foundation,a correlation analysis of the changes in the four indexes is performed,and PCA+K-Means is employed to assess the effectiveness.The results suggest that this strategy is effective at classifying data while also decreasing redundancy.Ultra-thin overlay has the best effectiveness.There is little difference between micro surfacing and fog seal,and there isn’t much of a difference in terms of effectiveness.The efficiency of ultra-thin overlay in increasing the RDI and SRI of pavement is the best;the fog seal largely retains the original pavement level in improving the SRI of pavement.Thirdly,the pavement performance grade of expressways in 2020 is assessed using the SVM classification evaluation model.The PQI is no longer calculated using a weight combination in the SVM model,which eliminates subjectivity and improves classification accuracy.The micro average accuracy is 89 percent,which is near 90percent;the weighted average accuracy is about 90 percent,and the evaluation model’s accuracy is relatively high.A few sample points’ evaluation findings are slightly lower than the standard evaluation grade,which is in line with Beijing’s existing road conditions.Finally,the GBDT model is used to predict expressway PQI in 2020,taking into account characteristics such as surface layer thickness,base layer thickness,and preventative maintenance technology,and it outperforms the KNN and linear regression models.The results show that the MAE of GBDT model is 1.314,the RMSE is 1.750,and R2 is 0.887.The MAE of GBDT model is 20.6% lower than that of KNN model and 29.2% lower than that of a linear regression model;The RMSE is reduced by 12.0% compared with KNN model and 23.8% compared with the linear regression model.The thesis refines the scope of application of pre-conservation technology for expressways in Beijing,proposes and validates a model for evaluating and predicting pavement performance in conjunction with Beijing’s characteristics,which can provide new methods and ideas for evaluating and forecasting experessway pavement performance,as well as serve as a reference for scientific and reasonable road management and maintenance.