| Project Title | Intelligent and Balanced Asphalt Mix Design based on Machine Learning |
| University | Virginia Tech University |
| Principal Investigator(s) | Linbing Wang |
| Funding Source(s) and Amounts Provided (by each agency or organization) | Federal Funds, $50,298 Match, $50,376 |
| Total Project Cost | $100,674 |
| Start and End Dates | 10/01/2018 - 09/31/2020 |
| Brief Description of Research Project | The objective of this study is to develop an intelligent and time-saving asphalt mix design procedure to produce a balanced mixture based on data-driven methods. The objective can be divided into three parts: (1) establish ML/DL models to capture the relationship between properties of the binder, aggregates, gradation, asphalt content, gyration numbers, and asphalt mixture volumetric properties; (2) establish pavement predictive models with the inputs of asphalt concrete properties, other pavement materials information, pavement structure, climate, and traffic based on the data analytics methods; and (3) propose an intelligent and balanced asphalt mix design 2 procedure by combining the optimizing asphalt content module, pavement predictive models, and optimization algorithms. The specific tasks are listed below: § Establish a method to optimize the current asphalt mixture design by predicting effective/total asphalt content, § Propose a methodology by which the mix proportion can be determined when given desired dynamic modulus values based on ML/DL approaches and optimization algorithm, § Establish a rutting depth predictive model with ML/DL methods and data processing techniques, § Propose an IRI predictive model using ML/DL techniques and data processing techniques, § Create fatigue cracking predictive models using ML/DL methodology and data analytics, and § Achieve an intelligent and balanced asphalt mix design by combining the optimizing asphalt content module, pavement predictive models, and optimization algorithms. |