| With the rapid economic development of Yunnan Province,the traffic volume of ordinary national highways in Yunnan province keeps growing and the maintenance demand increases year by year.However,due to the long length of ordinary national highways in Yunnan Province,weak basic conditions and limited maintenance funds,how to better invest the limited funds into pavement maintenance and achieve accurate maintenance according to scientific decision-making to maintain and improve pavement performance? In particular,how to accurately predict the decay of service performance of asphalt pavement in Yunnan province in order to solve the subsequent road maintenance decision and maintenance fund allocation problem is the concern and urgent problem of highway maintenance department at present.This paper combines the actual maintenance needs of ordinary national and provincial roads in Yunnan Province.The main research contents and achievements are as follows:First of all,taking the asphalt pavement of the common national highway in Yunnan Province as the research object,the common pavement diseases in each city and city were investigated and studied,and the common disease types of national highway in different cities and cities in Yunnan Province were obtained.The statistical analysis showed that: The dominant damage types in the central region are cracks and ruts,the dominant damage types in the northwest are massive cracks and transverse cracks,the dominant damage types in the southwest are longitudinal cracks,pits and oil flooding,the dominant damage types in the southeast are transverse cracks,ruts and oil flooding,and the dominant damage types in the northeast are loose and settlement.Then the cause of the disease,that is,the influencing factors of pavement performance are analyzed.Secondly,the comprehensive service performance of the general national and provincial trunk highway network in Yunnan Province was investigated.Considering the large differences in climate and other influencing factors among different regions in Yunnan Province,in order to make the prediction of road performance more accurate,In this paper,temperature,rainfall and traffic load are selected as the influencing factors of asphalt pavement performance of national highways in Yunnan province.Combined with existing regional studies,regional division,typical cities and representative sections of national and provincial trunk highway network root in Yunnan Province are carried out,and PCI and PQI are selected as predictive indicators.Then,grey relational degree and analytic hierarchy process were used to evaluate the service performance of the asphalt highway in Yunnan province,and the service performance evaluation system of asphalt pavement was established in different regions.The results show that: Compared with the current standard specification method,a more accurate evaluation result of asphalt pavement performance grade is obtained,which is conducive to the subsequent maintenance decision guidance.Finally,considering the influence factors of pavement performance comprehensively,Based on multiple linear regression method,LSTM(Long Short Term Memory)model and XGBoost(e Xtreme Gradient)Boosting)algorithm constructed a prediction model for the pavement performance of the general national trunk asphalt highway in Yunnan province under the sub-regional multi-factor coupling condition,and compared the predicted results of different models in2018-2022 with the actual test values,the results show that: The root-mean-square error of the prediction results of each region of the XGBoost model is about 2,and the fitting coefficient is above 0.9.The other two models have root-mean-square error of 2-4,and the fitting coefficient is between 0.6 and 0.9,which verifies that the prediction results of XGBoost are more reliable. |
