養(yǎng)護干預下的瀝青路面平整度預測
發(fā)布時間：2019-06-27 07:53
【摘要】：平整度作為使用性能系統(tǒng)中的綜合性指標,不僅可以評估路面的行駛質(zhì)量,也可以為養(yǎng)護決策提供依據(jù)。本文依托交通運輸部建設科技項目《瀝青路面長期使用性能研究(LAPP)》(編號：2014 318 223 010),以平整度為出發(fā)點,建立在不同截面信息及養(yǎng)護活動干預下瀝青路面使用性能的演化規(guī)律,從而為養(yǎng)護規(guī)劃中確定養(yǎng)護時機及最佳養(yǎng)護方案提供現(xiàn)實依據(jù),也為進一步完善科學化養(yǎng)護規(guī)劃體系提供了思路。本文在平整度影響因素機理分析及大量已有研究的基礎上,總結(jié)出平整度發(fā)展的影響因素可以分為七類,分別為：路面病害、路齡、交通荷載、路面結(jié)構、環(huán)境、初始平整度和養(yǎng)護活動。在此基礎上,結(jié)合我國瀝青路面長期使用性能數(shù)據(jù)庫中數(shù)據(jù)類型,提取出所有與平整度有關的數(shù)據(jù)項,并對所有數(shù)據(jù)進行預處理,完成瀝青路面平整度數(shù)據(jù)庫的重構。最后,采用統(tǒng)計學方法對不同的檢測數(shù)據(jù)和屬性數(shù)據(jù)進行分析,得到其統(tǒng)計學規(guī)律與對比分析。對影響因素進行逐項分析,選擇其中的三大類截面信息：路面病害、交通荷載和路面結(jié)構作為變量分析的候選變量,并細化為11個具體變量。通過因子分析與聚類分析的對比研究,確定采用SAS統(tǒng)計分析軟件對候選變量進行變量聚類分析,最終確定平整度預測模型的輸入變量為破損率、交通量和路面結(jié)構指數(shù)。通過各種預測方法及理論的對比分析,選擇適合于面板數(shù)據(jù)建模的混合效應模型作為平整度預測模型的統(tǒng)計學理論基礎,以Logistic模型為基礎模型采用SAS中的NLMIXED模塊建立新建路面平整度的非線性混合效應模型。模型中的隨機效應能夠體現(xiàn)不同路段的差異性,而在參數(shù)中引入?yún)f(xié)變量不僅能夠進一步解釋各路段的差異性,還可以在一定程度上提高模型擬合精度。同時,模型的建立驗證了非線性混合效應模型在平整度預測中的適用性,為接下來的養(yǎng)護活動干預研究提供了理論基礎。最后,基于非線性混合效應模型研究養(yǎng)護活動干預下瀝青路面的平整度衰變規(guī)律,并分析不同養(yǎng)護措施、養(yǎng)護前平整度值及養(yǎng)護路段的路面結(jié)構指數(shù)對養(yǎng)護活動后平整度演化規(guī)律的影響；綜合以上分析,結(jié)合養(yǎng)護閾值的確定、費用效益分析等,最終提出基于平整度的瀝青路面養(yǎng)護規(guī)劃體系。
[Abstract]:As a comprehensive index in the performance system, smoothness can not only evaluate the driving quality of pavement, but also provide the basis for maintenance decision. Based on the construction science and technology project of Ministry of Transportation (LAPP) (No. 2014 318223 010), this paper establishes the evolution law of asphalt pavement performance under the intervention of different cross section information and maintenance activities, which provides a realistic basis for determining the maintenance time and the best maintenance scheme in the maintenance planning, and also provides an idea for further perfecting the scientific maintenance planning system. Based on the mechanism analysis of the influencing factors of smoothness and a large number of existing studies, this paper summarizes that the influencing factors of flatness development can be divided into seven categories: pavement disease, road age, traffic load, pavement structure, environment, initial smoothness and maintenance activities. On this basis, combined with the data types in the long-term performance database of asphalt pavement in China, all the data items related to smoothness are extracted, and all the data are preprocessed to complete the reconstruction of asphalt pavement smoothness database. Finally, the statistical method is used to analyze the different detection data and attribute data, and the statistical law and comparative analysis are obtained. The influencing factors are analyzed item by item, and three kinds of cross section information are selected: pavement disease, traffic load and pavement structure as candidate variables for variable analysis, and refined into 11 specific variables. Through the comparative study of factor analysis and cluster analysis, it is determined that SAS statistical analysis software is used to cluster the candidate variables, and finally the input variables of the smoothness prediction model are damage rate, traffic volume and pavement structure index. Through the comparative analysis of various prediction methods and theories, the mixed effect model suitable for panel data modeling is selected as the statistical theoretical basis of the smoothness prediction model. Based on the Logistic model, the nonlinear mixed effect model of the smoothness of the new pavement is established by using the NLMIXED module of SAS. The random effect in the model can reflect the difference of different road sections, and the introduction of covariables into the parameters can not only further explain the differences of each road section, but also improve the fitting accuracy of the model to a certain extent. At the same time, the establishment of the model verifies the applicability of the nonlinear mixed effect model in smoothness prediction, and provides a theoretical basis for the next study of conservation intervention. Finally, based on the nonlinear mixed effect model, the smoothness decay law of asphalt pavement under the intervention of maintenance activities is studied, and the influence of different maintenance measures, leveling value before maintenance and pavement structure index on the evolution law of smoothness after maintenance activities is analyzed. Based on the above analysis, combined with the determination of maintenance threshold, cost-benefit analysis and so on, the asphalt pavement maintenance planning system based on smoothness is put forward.
【學位授予單位】：東南大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：U416.217
本文編號：2506626
[Abstract]:As a comprehensive index in the performance system, smoothness can not only evaluate the driving quality of pavement, but also provide the basis for maintenance decision. Based on the construction science and technology project of Ministry of Transportation (LAPP) (No. 2014 318223 010), this paper establishes the evolution law of asphalt pavement performance under the intervention of different cross section information and maintenance activities, which provides a realistic basis for determining the maintenance time and the best maintenance scheme in the maintenance planning, and also provides an idea for further perfecting the scientific maintenance planning system. Based on the mechanism analysis of the influencing factors of smoothness and a large number of existing studies, this paper summarizes that the influencing factors of flatness development can be divided into seven categories: pavement disease, road age, traffic load, pavement structure, environment, initial smoothness and maintenance activities. On this basis, combined with the data types in the long-term performance database of asphalt pavement in China, all the data items related to smoothness are extracted, and all the data are preprocessed to complete the reconstruction of asphalt pavement smoothness database. Finally, the statistical method is used to analyze the different detection data and attribute data, and the statistical law and comparative analysis are obtained. The influencing factors are analyzed item by item, and three kinds of cross section information are selected: pavement disease, traffic load and pavement structure as candidate variables for variable analysis, and refined into 11 specific variables. Through the comparative study of factor analysis and cluster analysis, it is determined that SAS statistical analysis software is used to cluster the candidate variables, and finally the input variables of the smoothness prediction model are damage rate, traffic volume and pavement structure index. Through the comparative analysis of various prediction methods and theories, the mixed effect model suitable for panel data modeling is selected as the statistical theoretical basis of the smoothness prediction model. Based on the Logistic model, the nonlinear mixed effect model of the smoothness of the new pavement is established by using the NLMIXED module of SAS. The random effect in the model can reflect the difference of different road sections, and the introduction of covariables into the parameters can not only further explain the differences of each road section, but also improve the fitting accuracy of the model to a certain extent. At the same time, the establishment of the model verifies the applicability of the nonlinear mixed effect model in smoothness prediction, and provides a theoretical basis for the next study of conservation intervention. Finally, based on the nonlinear mixed effect model, the smoothness decay law of asphalt pavement under the intervention of maintenance activities is studied, and the influence of different maintenance measures, leveling value before maintenance and pavement structure index on the evolution law of smoothness after maintenance activities is analyzed. Based on the above analysis, combined with the determination of maintenance threshold, cost-benefit analysis and so on, the asphalt pavement maintenance planning system based on smoothness is put forward.
【學位授予單位】：東南大學
【學位級別】：碩士
【學位授予年份】：2016
【分類號】：U416.217
【參考文獻】
相關期刊論文 前10條
1 何振宇;余武軍;;AASHTO瀝青混凝土路面設計方法探討[J];交通科技;2014年03期
2 袁捷;蘇爾好;杜先照;滕力鵬;;基于線性混合效應模型的道面使用性能預測[J];同濟大學學報(自然科學版);2014年05期
3 王雪如;張金喜;孔祥杰;;基于行車舒適性的路面破損評價問題探討[J];市政技術;2014年03期
4 劉強;劉化學;張寧波;;基于平整度的高速公路路面設計加鋪厚度驗證[J];中外公路;2014年01期
5 周鵬飛;溫勝強;康海貴;;基于馬爾可夫鏈與神經(jīng)網(wǎng)絡組合的路面使用性能預測[J];重慶交通大學學報(自然科學版);2012年05期
6 陳興偉;;力學-經(jīng)驗路面設計指南(MEPDG)簡介[J];上海公路;2011年03期
7 杜二鵬;馬松林;景海民;;基于灰色系統(tǒng)理論的瀝青路面使用性能預測[J];同濟大學學報(自然科學版);2010年08期
8 王爽;黃曉明;朱湘;;應用神經(jīng)網(wǎng)絡-時間序列預測路面平整度[J];現(xiàn)代交通技術;2007年03期
9 喻翔;毛敏;;高速公路路面使用性能預測方法研究[J];交通標準化;2006年08期
10 倪富健;方昱;薛智敏;;時間序列在路面平整度預測中的應用[J];東南大學學報(自然科學版);2006年04期
相關博士學位論文 前1條
1 曾峰;瀝青路面預防性養(yǎng)護決策方法的研究與技術應用[D];哈爾濱工業(yè)大學;2009年
相關碩士學位論文 前6條
1 嚴濤;基于變參數(shù)和混合模型的醫(yī)學縱向數(shù)據(jù)研究[D];南昌大學;2014年
2 李莉娟;基于預防性養(yǎng)護的山區(qū)高等級公路路面使用性能預測和養(yǎng)護決策研究[D];重慶交通大學;2012年
3 崔文博;基于路面使用性能衰變規(guī)律的預防性養(yǎng)護時機研究[D];長安大學;2011年
4 楊亮;遺傳算法在路面管理系統(tǒng)中的應用研究[D];西南交通大學;2008年
5 邵東野;基于國際平整度指數(shù)（IRI）的瀝青路面養(yǎng)護評價體系[D];長安大學;2007年
6 薛智敏;現(xiàn)代高速公路瀝青路面路面管理系統(tǒng)框架及其預測模型研究[D];東南大學;2006年
，本文編號：2506626
本文鏈接：http://www.sikaile.net/kejilunwen/jiaotonggongchenglunwen/2506626.html
教材專著
