**Natasha Bahrani \*, Juliette Blanc, Pierre Hornych and Fabien Menant**

IFSTTAR, Allée des Ponts et Chaussées, 44340 Bouguenais, France; juliette.blanc@ifsttar.fr (J.B.); pierre.hornych@ifsttar.fr (P.H.)mailto:; fabien.menant@ifsttar.fr (F.M.) **\*** Correspondence: Natasha.bahrani@ifsttar.fr

Received: 5 December 2019; Accepted: 26 February 2020; Published: 1 March 2020

**Abstract:** Pavement instrumentation with embeddable in-situ sensors has been a feasible approach to determine pavement deteriorations. Determining pavement deflections during the passage of the load is a promising strategy to determine the overall performance of the pavement. There are different devices that apply loads to the pavements and measure the deflection basin, these include static, vibratory, or impulse loadings. Most commonly used are the static loading like Benkelman beam and impulse loading like the Falling Weight Deflectometer (FWD). However, these techniques are costly and the measurements are recorded infrequently, i.e., once per year or two years. This study focuses on the use of geophones and accelerometers to measure the surface deflections under traffic loading. To develop a method to measure pavement deflections, the sensors were submitted first to laboratory tests, and then tested in situ, in a full scale accelerated pavement test. In the laboratory, the sensors were submitted to different types of loading using a vibrating table. These tests were used to determine the noise and sensitivity of the sensors, and then to evaluate their response to signals simulating pavement deflections under heavy vehicles. The sensor response was compared with measurements of a reference displacement sensor. Different processing techniques were proposed to correct the measurements from geophones and accelerometers, in order to obtain reliable deflection values. Then, the sensors were evaluated in a full scale accelerated test, under real heavy axle loads. Tests were performed at different loads and speeds, and the deflection measurements were compared with a reference anchored deflection sensor. The main advantage of using accelerometers or geophones embedded in the pavement is to enable continuous pavement monitoring, under real traffic. The sensor measurements could also be used to determine the type of vehicles and their corresponding speeds. The study describes in detail the signal analysis needed to measure the pavement deflections accurately. The measurements of pavement deflection can be then used to analyze the pavement behavior in the field, and its evolution with time, and to back-calculate pavement layer properties.

**Keywords:** pavement monitoring; accelerometers; geophones; pavement instrumentation; pavement displacement; condition assessments

## **1. Introduction**

In recent years, there has been a strong interest in long term monitoring of in-service pavements, and in using non-destructive tools to evaluate pavement performance. Several non-destructive testing methods have been developed to measure the surface deflection and to assess the structural capacity of asphalt pavements. The most commonly used are static measurement methods like the Benkelman beam and impulse loading methods like the FWD. These methods present the advantage of reproducing

closely the loading conditions and the stress state in the pavement. However, these techniques can only measure the deflection at discrete locations, and require interrupting the traffic. Hence, these techniques are costly and the measurements are recorded infrequently, i.e., only once per year or two years.

The need to collect the data without disruption of the traffic flow and to measure the response continuously, under normal traffic, has led to the development of pavement instrumentation with different sensors. Strain sensors and temperature probes are the most commonly used devices to instrument and characterize pavement conditions. Some applications have also used piezoelectric sensors to measure stresses and strains in pavement layers, to monitor the fatigue damage of pavements [1–3].

In paper [4] the work focuses on the vibro-acoustic response of transportation infrastructure. It uses a lightweight deflectometer to identify the localization of cracks and variations in the elastic modulus of the pavement by using the road traffic noise as its vibration source and by the help of microphones drilled in the upper layer of the pavement which record the vibration noise. These vibrations are analyzed to identify the severity of the cracks and the level of damage. In [5], the author discusses an approach using the falling weight deflectometer to back-calculate the layer moduli using a probabilistic approach. The static and dynamic deflection bowls are used from the FWD data sets, for the back-calculation of layer moduli and a comparative study is made to accurately back-calculate pavement layer moduli.

The work of [6] focuses on using a laser dynamic deflectometer that has four laser Doppler sensors and a rigid beam to measure the deflection velocity and then compute the deflection basin. It defines the relation between the measured values and the position of the sensors and analyzes the effectiveness of both dynamic and static calibration methods on measurement accuracy. It also studies the impact of the traffic speed on obtained results and concludes that both methods produce effective results.

Geophones and accelerometers are designed to measure the inertial vibrations of the pavement and used in research studies to calculate the deflection basin of the pavement caused by the passage of wheel loading. In the work done by Lie et al. [7], an experimental setup with an array of geophones was used to measure the vibrations during the passage of a known vehicle load. The responses, along with a mobile load simulator and a back-calculation tool, were used to measure the pavement deteriorations.

The work done by Arraigada et al. [8] consisted in measuring the pavement deflections using accelerometers and deflectometers. They also used a visco-elastic pavement model to fit the measurements obtained with the deflectograph. This research points out the difficulties in converting the acceleration into deflections, due to the integration process and noise and drift in the signal. A spline-based correction method was used to convert the acceleration signal into deflections.

The work done by Levenberg [9] describes the use of an accelerometer embedded in asphalt pavement. The loading is applied by a vehicle of known dimensions and weight passing near the accelerometer. These accelerations could then be used for back-calculating the pavement layer properties, without converting the measurements into deflections, which eliminates the integration and amplification errors.

Ngoc Son et al. [10] use geophones to measure pavement deflections and to monitor evolution of pavement layer properties with respect to traffic and environmental conditions. They also focus on the need to improve signal processing to get accurate results using geophones.

This work is part of a research that aims at developing a monitoring system for asphalt pavements focusing on the use of geophones and accelerometers for measuring deflections and accurately back-calculating pavement layer moduli. The sensors are evaluated first in the laboratory, and then on an accelerated pavement testing facility, under various testing conditions. In each case, a reference sensor is used to evaluate measurement accuracy. In the laboratory tests, a laser displacement sensor is used and for the full-scale accelerated pavement tests, an anchored deflectometer is installed. These reference sensors have a good accuracy, but their deployment on a real road site is difficult.

The main advantage of using accelerometers or geophones is the ease of embedment on a real pavement before and even after the construction of the pavement, with a possibility of having continuous pavement monitoring, under real traffic. The determination of the type of vehicles and their corresponding speeds is also possible with the aid of these sensors.
