Investigation of the correlations between the field pavement in-place density and the intelligent compaction measure value (ICMV) of asphalt layers. (19th July 2021)

• Record Type: Journal Article
• Authors: Zhang, Weiguang; Raza Khan, Ali; Yoon, Soojin; Lee, Jusang; Zhang, Runhua; Zeng, Kun
• Is Part Of: Construction & building materials. Volume 292(2021)
• Publication Date: 2021-07-19
• Subjects: IC (intelligent compaction) -- ICMV's (intelligent compaction measurement values) -- MV's (measurement values) -- NNG (non-nuclear gauge density) -- CCV (compaction control value); Building materials -- Periodicals; 624.18
• DOI: 10.1016/j.conbuildmat.2021.123439
• Languages: English
• ISSNs: 0950-0618

Abstract

Highlights: There is a consistent trend observed between the NNG in-place density and the CCV value measured by the IC technique. The developed models can be used to predict the pavement in-place density by including different rolling patterns, with predictive accuracy increases with the inclusion of pass count. The newer IC technique and its corresponding ICMV's can be employed to control/monitor the compaction quality of the asphalt layers. Abstract: With the advancement in pavement technology, non-destructive testing is getting fame. Intelligent Compaction Measure Value (ICMV) provided from the intelligent compactor has been explored to indicate the pavement stiffness and the compaction quality of asphalt pavement. This study aims to analyze the advanced ICMV and traditional non-nuclear gauge (NNG) density measurements to investigate the relative correlations in compaction measuring/monitoring indicators. This research applies Sakai Compaction Control Value (CCV) to predict the non-nuclear gauge density in means of different machine learning models. For analysis, data were collected from US-52 highway in the United Sates with three passes. Relative percentage change is calculated to measure the overall and individual grid change in density and CCV values with increase of passes. Four different type of models are finally developed include the first three simple ones (pass 1, pass 2, and pass 3 models) to predict the in-place density for different rolling pattern followed by different contractors, and the fourth model (all pass data model) which is developed by joining the data of all passes along with the categorical variable pass count. All model results are reasonably good and significant. Results show that with the inclusion of pass count, model prediction accuracy increases. The pass count is the categorical variable which can affect the model prediction power and provide good results. The possibility of substituting the current practice of quality control testing, NNG, with ICMV was proved throughout this study.