The dynamic modulus of hot mix asphalt (HMA) is a fundamental material property that defines the stress-strain relationship based on viscoelastic principles and is a function of HMA properties, loading rate, and temperature. Because of the large number of efficacious predictors (factors) and their nonlinear interrelationships, developing predictive models for dynamic modulus can be a challenging task. In this research, results obtained from a series of laboratory tests including mixture dynamic modulus, aggregate gradation, dynamic shear rheometer (on asphalt binder), and mixture volumetric are used to create a database. The created database is used to develop a model for estimating the dynamic modulus. First, the highly correlated predictor variables are detected, then Principal Component Analysis (PCA) is used to first reduce the problem dimensionality, then to produce a set of orthogonal pseudo-inputs from which two separate predictive models were developed using linear regression analysis and Artificial Neural Networks (ANN). These models are compared to existing predictive models using both statistical analysis and Receiver Operating Characteristic (ROC) Analysis. Empirically-based predictive models can behave differently outside of the convex hull of their input variables space, and it is very risky to use them outside of their input space, so this is not common practice of design engineers. To prevent extrapolation, an input hyper-space is added as a constraint to the model. To demonstrate an application of the proposed framework, it was used to solve design-based optimization problems, in two of which optimal and inverse design are presented and solved using a mean-variance mapping optimization algorithm. The design parameters satisfy the current design specifications of asphalt pavement and can be used as a first step in solving real-life design problems.

Permanent deformation is a major load-associated distress occurring in flexible pavement systems and increases with load repetitions affecting road roughness, serviceability, and the international roughness index (IRI). Early detection of rutting is necessary for maintenance and rehabilitation activities, but due to the complex behavior of asphalt mixtures, accurately predicting the permanent deformation of asphalt pavement is difficult. Historically, multivariate regression modeling and recently, artificial neural networks (ANNs) are used widely for material properties prediction. The ability to model accurately the response variable is adversely affected when inputs have pairwise correlations. To overcome this barrier, principal component analysis (PCA), as a dimensionality reduction technique, can be used to produce uncorrelated linear combinations of the original inputs as illustrated in this work using 83 (i.e., samples) laboratory compacted specimens from the State of Wisconsin. Asphalt binder, aggregate, and mix properties are obtained and used as the model inputs. The response parameter is the accumulated strain at the corresponding flow number. Using the developed pseudo inputs from PCA, a multivariate regression and an ANN model are generated and were able to fit the test cases with r_fit $r_{\mathit{fit}}$ <!--mathContainer--> <!--Loading Mathjax--> of 0.8 and 0.97 respectively. The developed machine learning-based framework is shown to be a capable tool in estimating the rutting behavior of asphalt mixture.

Historically, asphalt mixtures in Minnesota have been produced with fine gradations. However, recently more coarse-graded mixtures are being produced as they require less asphalt binder. Thus, it is important that pavement performance for coarse gradations be evaluated. Within this research work, performance evaluation took place with the use of the dynamic modulus test in indirect tension mode on coarse-graded mixtures consisting of field cores from 9 different pavements located in five districts of Minnesota. From each pavement’s surface layer, 3 specimens were tested at three temperatures; 0.4°C, 17.1°C, and 33.8°C each at nine frequencies ranging between 0.1 Hz and 25 Hz. Additional volumetric characterization of the field mixtures was done to determine asphalt content, air voids, and blended aggregate gradations. Asphalt binders were extracted and recovered for use in determining binder shear complex master curves. Through this information the modified Witczak model was used to create |E*| master curves which were then compared against the indirect tension (IDT) test |E*| experimentally created master curves. From the results the modified Witczak model needs to be modified for IDT collected dynamic modulus data.