Crack initiation and propagation affect the mechanical properties of asphalt pavement and significantly limit its design life. To assess the structural health of asphalt, knowledge of the strain condition has proven to be useful and optical fibers, conventional strain gauge, metal-foil-type gauges, can be used for this purpose. The high cost of optical fibers has limited their use in this application and although conventional strain gauges have an excellent capability of assessing even large strains, they are rarely used because asphalt paving has high temperature (up to 150 ℃) and pressure (around 290ksi) during installation. A new multi-layered strain sensor has been proposed to overcome the installation challenges in asphalt pavement and provide a reliable structural health monitoring of capability. In this proof-of-concept study, a finite element model was created to simulate the pavement–sensor interaction for the proposed strain sensor in order to optimize its design and potential.

The core of the strain sensor is an H-shaped Araldite GY-6010 epoxy structure with a set of PVDF piezoelectric transducer in its center beam and a polyurethane foam layer at its external surface. An extra thin layer of urethane casting resin is then coated at the external surface of the polyurethane foam to prevent the sensor from being damaged under harsh installation condition. During the life of the pavement, traffic loads deform the embedded sensor. The sensor deflection variations register as voltage amplitude changes by the piezoelectric transducers that are later recorded as strain changes. When a crack initiates and propagates in the pavement, sensor deflection increases dramatically as the pavement fails to maintain its designed tensile and flexural strength.

The simulation of the pavement–sensor interaction resulted in a modification of several key elements of the original design. The internal Araldite epoxy structure, and outer foam layer were optimized with thicknesses of 11mm, and 2.5mm, respectively. Two design parameters for the sensor, side wing length/middle beam length ratio, and pavement thickness, were also adjusted for optimal design and location of the sensor. The numerical model was created and the simulation performed using COMSOL. Using a fixed traffic load, and varying the side wing length between 20mm-50mm, and the middle beam length between 80mm-200mm, with the best ratio of side wing to middle beam length of 0.3125 was determined as a result of this study.