| Project Title | Use of Machine Learning to Predict Long-Term Skid Resistant of Concrete Pavements |
| University | Penn State |
| Principal Investigator(s) | Farshad Rajabipour |
| PI Contact Information | fxr10@psu.edu |
| Funding Source(s) and Amounts Provided (by each agency or organization) | Federal Share — $23,964 Match — $24,582 |
| Total Project Cost | $48,546 |
| Start and End Dates | 05/11/2020 — 03/31/2023 |
| Brief Description of Research Project | This project intends to use laboratory-measured concrete frictional properties data and build a machine learning (ML) model to predict the friction performance of concrete pavements. The model uses aggregate mineralogy, concrete mixture proportions, and concrete surface finish (tined vs. ground and grooved) as input parameters. It predicts the surface friction (measured via the dynamic friction test, DFT, ASTM E 1911; Komaragiri et al., 2020) and surface texture (measured via the circular track meter, CTM, ASTM E 2157; Komaragiri et al., 2020) of concrete pavements undergoing accelerated polishing tests as a function of number of polishing cycles as well as the terminal friction and texture values associated with long-term polishing. The model further uses these outputs to predict the international friction index (IFI) parameters, including the friction number (F60) and slip speed (SP), as well as the equivalent skid number (SN) measured by locked-wheel trailer test (ASTM E524) at a given speed and for a given concrete pavement. |