About

INTENT (Intelligent health monitoring of road infrastructures using bender elements embedded in pavements) is an research project promoted by the Lusófona University, University of Lisbon, University of Minho, and Built COLAB, with the collaboration of BarcelonaTech and University of Illinois. It is funded by the Fundação para a Ciência e a Tecnologia.

Abstract

The expansion of road networks in regions with poor soils has led owners to call for improved and continuous monitoring solutions. Accurate and timely geotechnical information enable better planning of retrofit interventions, with huge savings to the €30B/year EU highway maintenance budget.

Currently, most pavement monitoring solutions are based on periodic inspections of the asphalt course. However, the pavement foundation (subgrade, capping, sub-base) and the granular base course are essential components of the road structure. The foundation acts as a construction platform for the upper layers and must sustain the traffic loads once construction is completed. Likewise, the base course sustains the construction of the asphalt course and spreads the loads to reduce the stresses in the foundation. The failure of the base course will inevitably compromise the pavement and is much harder to detect visually than the damage of the asphalt course.

The objective of this project is to develop a new embedded sensing device, based on the bender element technology, for the continuous monitoring of the dynamic stiffness (shear modulus) of unbound granular layers, and to use it, along with conventional sensors and advanced numerical models, to fuel machine learning algorithms for the continuous monitoring of pavement infrestructure.

The bender elements developed in this project enable the continuous measurement of the stiffness of the granular layers during construction, ensuring that pavement design requirements are met, and during the service life of the structure, enabling retrofit actions in the early stages of damage. They can be embedded in any layer with minimal disturbance, and their signal analysed automatically to extract the shear modulus of the layer. They will be coupled with conventional sensors to get a comprehensive picture of the condition of the layer, and the data analysed in real time by machine learning algorithms for damage detection. Advanced numerical models will help machine learning algorithms distinguish the between gradual loss of stiffness and/or gradual increase of permanent deformation under cyclic loading (concept of fatigue) and the sudden, catastrophic deterioration of the geomaterial (concept of failure). A damage progression prognosis toolbox will be developed to compare the expected and measured rates of degradation of the geomaterial and estimate its future condition.

The project combines complementary competences and equipment from 6 research institutions in Portugal, Spain and USA. It is rooted in the CEN-DynaGeo Project, where numerical models were used for the optimization and interpretation of bender element tests in the lab.