Abbasgholipour, Sina (2018-12). Developing Pavement Management Systems for Small to Medium Size Cities: Challenges, Solutions, and a Case Study. Master's Thesis.
Thesis
In this research, three problems associated with developing a Pavement Management System (PMS) for small to mid-size cities are discussed, and a solution is suggested for each of these problems. First, the comparison of condition of the road network based on PCI and IRI indicates that using IRI as a pavement condition indicator in urban areas can lead to misleading information about the condition of the road network. Additionally, the relationship between IRI and PCI was assessed. It was demonstrated that there is a weak correlation between IRI and PCI which is not enough to estimate one from another accurately. The lack of a meaningful relationship between these two pavement performance indicators in urban areas can be explained by distortions of IRI measurements in local road networks. Next, the problem with developing pavement prediction models without having a historical condition database is raised and a solution is suggested to overcome this problem. It is shown that, by having a rough estimate of the overall change in condition of the road network during one year, an iterative process can be used to estimate pavement prediction models coefficient. The methodology is tested using data from the city of College Station and one prediction model was developed for each M&R treatment. Finally, problems associated with segmentation of the road network in urban areas are discussed in more detail. An automated segmentation method based on PDA approach is developed in Python. The segmentation code is compatible with ArcGIS which provides the user with all the visualization and analytical tools in ArcGIS. This segmentation method is successfully implemented for the city of College Station's road network considering four different thresholds.
In this research, three problems associated with developing a Pavement Management System (PMS) for small to mid-size cities are discussed, and a solution is suggested for each of these problems. First, the comparison of condition of the road network based on PCI and IRI indicates that using IRI as a pavement condition indicator in urban areas can lead to misleading information about the condition of the road network. Additionally, the relationship between IRI and PCI was assessed. It was demonstrated that there is a weak correlation between IRI and PCI which is not enough to estimate one from another accurately. The lack of a meaningful relationship between these two pavement performance indicators in urban areas can be explained by distortions of IRI measurements in local road networks. Next, the problem with developing pavement prediction models without having a historical condition database is raised and a solution is suggested to overcome this problem. It is shown that, by having a rough estimate of the overall change in condition of the road network during one year, an iterative process can be used to estimate pavement prediction models coefficient. The methodology is tested using data from the city of College Station and one prediction model was developed for each M&R treatment. Finally, problems associated with segmentation of the road network in urban areas are discussed in more detail. An automated segmentation method based on PDA approach is developed in Python. The segmentation code is compatible with ArcGIS which provides the user with all the visualization and analytical tools in ArcGIS. This segmentation method is successfully implemented for the city of College Station's road network considering four different thresholds.
