Fatigue cracking is a major issue in asphalt pavements, reducing their lifespan and increasing maintenance costs. This study develops an artificial neural network (ANN) model to predict the onset and progression of fatigue cracking. The model is calibrated utilizing Falling Weight Deflectometer (FWD) testing data, alongside essential pavement characteristics such as layer thickness, air void percentage, asphalt binder proportion, traffic loads (Equivalent Single Axle Loads or ESALs), and mean annual temperature. By analyzing these factors, the ANN captures complex relationships influencing fatigue cracking more effectively than traditional methods. A comprehensive dataset from the Long-Term Pavement Performance (LTPP) program is used for model training and validation. The ANN’s ability to adapt and recognize patterns enhances its predictive accuracy, allowing for more reliable pavement condition assessments. Model performance is evaluated against real-world data, confirming its effectiveness in predicting fatigue cracking with an overall R2 of 0.9. This study’s findings provide valuable insights for pavement maintenance and rehabilitation planning, helping transportation agencies optimize repair schedules and reduce costs. This research highlights the growing role of AI in pavement engineering, demonstrating how machine learning can improve infrastructure management. By integrating ANN-based predictive analytics, road agencies can enhance decision-making, leading to more durable and cost-effective pavement systems for the future.
1 October 2025 in Canadian Journal of Civil Engineering
This study presents a physics-informed neural network (PINN) framework for predicting the international roughness index (IRI) of flexible pavements, utilizing the long-term pavement performance database. A total of 390 observations from 74 pavement sections across the United States were utilized. The architecture of the PINN combines a mean square error (MSE) loss function with a custom physics-informed MSE (PMSE) loss function. It integrates mechanistic IRI equations sourced from the mechanistic-empirical pavement design guide (MEPDG) into the training process. Comparative analysis was performed against standard regression models and the MEPDG IRI prediction equation. Results indicate that the PINN model significantly improves prediction accuracy, achieving a coefficient of determination ( R2) of 0.743 and a root mean squared error of 26.57, outperforming traditional methods. Furthermore, a closed-form equation from the model was derived. By utilizing this new simplified equation, transportation agencies can allocate resources and prioritize maintenance activities efficiently through the IRI forecast PINN Model for asphalt pavements.
1 August 2019 in Journal of Materials in Civil Engineering
Fatigue cracking and permanent deformation (rutting) are major distresses that occur in flexible pavements. One of the most significant solutions to prolong the fatigue life as well as rutting resistance for an asphaltic mixture is to utilize flexible materials such as fibers. The impact of fiber addition to hot-mixed asphalt (HMA) was investigated utilizing a developed testing program in which the mechanistic and mechanical performance of fiber-reinforced mixtures as well as the economic attributes were evaluated. All beam fatigue tests were strain-controlled tests performed according to current guideline procedures. The testing results indicated that the fiber-reinforced mixtures had a better performance compared with conventional mixtures in terms of rutting and fatigue resistance. Additionally, a mechanistic analysis coupled with a cost-effectiveness analysis study based on the fatigue and rutting resistance performance on the two mixtures was performed. Overall, analysis showed that fiber-reinforced asphalt mixtures exhibited significantly lower cost of pavement per 1,000 cycles of fatigue and rutting life 1.6 km (1 mi) compared with conventional HMA mixture.
There is currently uncertainty in how to properly account for partial recycled binder availability within asphalt mixture design procedures so that content of recycled asphalt materials (RAM) can be maximized while still achieving good performance. This study evaluates and compares two alternative approaches to consider partial availability in mixture design: availability adjusted mix design (AAMD) and corrected optimum asphalt content (COAC). The AAMD method revises the calculation of volumetric properties by considering unavailable binder as part of the bulk aggregate volume and uses the RAM gradation to design the aggregate structure. In the COAC method, a mixture is first designed following the conventional Superpave procedures and then a specified increase to the virgin asphalt content is made. Two “control” mixtures containing RAM are evaluated that were originally designed according to the standard Superpave method. Alternative designs were prepared according to the COAC and AAMD methods while maintaining the RAM content in the respective control mixture. Baseline virgin mixtures and one mixture designed following the AAMD method with higher reclaimed asphalt pavement (RAP) content were also prepared. The cracking and rutting performance of the resultant mixtures was evaluated. The results show that the AAMD and COAC approaches lead to an improved cracking performance compared with the control mixtures even for the mixture designed with AAMD at a higher RAP content. However, the mixture designs prepared according to the COAC method presented increased rutting whereas the AAMD mixtures remained at the same level as their respective control mixture.