Recent Advances in the Use of Artificial Intelligence for Rigid Airfield Pavement Analysis and Design

Historically, analysis and design methodologies for rigid airfield pavement systems have primarily relied on empirical equations derived mostly based the field performance. However, the FAA has implemented mechanistic-based approaches, including three-dimensional finite-element (3D-FE) procedures, for the design and analysis of rigid airfield pavements. This decision was made in response to the emergence of new wide-body aircraft and other design complexities, such as heavier aircraft and intricate gear configurations. On the other hand, the time-consuming nature of 3D-FE computation for analyzing multiple slabs subjected to aircraft and environmental loads has rendered routine design and analysis impractical. To address this issue, AI-based alternatives offer significant potential for producing accurate and rapid rigid pavement-response estimations compared to traditional FE-based design programs. Using AI-based modeling is a convenient option instead of conducting lengthy 3D-FE computations. In this paper, we introduce a recent FAA-sponsored research study conducted at Iowa State University (ISU) that utilizes AI-based alternatives to performing full 3D-FE computation for analyzing and designing rigid airfield pavement systems. The paper contains two case studies: (1) creating strong AI models to predict critical rigid airfield pavement responses and analyze top-down cracking behavior and (2) developing fast AI-based models to predict pavement foundation response and moduli for designing new and rehabilitated rigid airfield pavement structures. The capability of AI-based surrogate-response models for analyzing and designing rigid airfield pavement systems is successfully demonstrated and discussed in this paper.