摘要
路面行驶质量指数(Riding Quality Index,RQI)可以反映路面平整度变化以及行驶质量舒适性程度。针对现有数学模型和机器学习模型对特定路段小样本且分布规律性不强的数据在分析时存在较大的局限性,基于检测数据和养护历史对广州西二环高速公路RQI展开分析。首先根据历年路面状况检测结果整合2012年~2015年这4年间的RQI指数,基于时空关联特性对异常数据进行修复。其次结合养护历史,提取有效数据,分析不同养护情况下的RQI时空变化过程。最后基于灰色模型对整个路段以公里桩为空间粒度对未进行养护工程的路段展开分析,挖掘不同路域环境下每公里RQI的变化规律。结果表明,与真实情况相比,所提出方法的分析精度R2达到0.902。与分析精度低于0.5的随机森林、CatBoost(Categorical Boosting)等机器学习模型相比,基于公里粒度分析的灰色模型具有更强的适应性和鲁棒性。
Riding Quality Index(RQI)can reflect the change of pavement roughness and riding comfort.In view of the limitations of the existing mathematical model and machine learning model in the analysis of small samples and poorly distributed data of specific sections,the RQI of Guangzhou West Second Ring Highway is analyzed based on the detection data and maintenance history.Firstly,the RQI index of 2012—2015is integrated according to the pavement condition detection results over the years,and the abnormal data are repaired based on the spatio-temporal correlation characteristics.Secondly,combined with the curing history,the effective data are extracted to analyze the spatio-temporal variation process of RQI under different curing conditions.Finally,based on the Grey model,the whole road section is analyzed with the kilometer pile as the spatial grain size for the road section without maintenance,and the variation rule of RQI per kilometer under different road domain environment is excavated.The results show that compared with the real situation,the analytic accuracy of the proposed method reaches 0.902.Compared with other machine learning models such as Random Forest and CatBoost(Categorical Boosting)with R2 less than 0.5,the Grey model based on kilometer-size analysis has better adaptability and robustness.
作者
张艳红
郭伟彤
张斌
姜宏维
ZHANG Yan-hong;GUO Wei-tong;ZHANG Bin;JIANG Hong-wei(China Highway Engineering Consulting Corporation,Beijing 100097,China;Research and Development Center of Transport Industry of Technologies,Materials and Equipment of Highway Construction and Maintenance,Beijing 100097,China;Research and Development Center on Highway Pavement Maintenance Technology,CCCC,Beijing 100097,China;Shanxi Communications Holding Group Co.Ltd.,Taiyuan 030024,China)
出处
《公路》
北大核心
2023年第12期22-28,共7页
Highway
基金
中交集团重点研发项目,项目编号zzkj-2018-005。
关键词
路面性能
路面行驶质量指数
灰色模型
时空特性
pavement performance
Riding Quality Index
Grey model
spatio-temporal characteristics
