Acoustic Imaging by distributed sensors

Manager: Emmanuel MOULIN

The "Acoustic Imaging by Distributed Sensors" theme concerns damage detection and imaging in reverberant and complex environments. One of the key ideas is to exploit acoustic signals in all their complexity, in order to extract the maximum information from a limited number of sensors. Ultrasonic codas, in particular, from multiple and intertwined propagation paths in structures offer new perspectives. In particular, the possibility of using natural acoustic sources (also referred to as "ambient" sources, or "opportunity sources") in place of the emission of ultrasonic waves classically used in traditional nondestructive testing techniques, opens the way to passive ultrasonic sensor networks (receivers only), thus low power and potentially autonomous and low-intrusive.

Historically developed in geophysical or underwater acoustics applications, such passive techniques are likely to bring here a number of advantages such as :

  • low power consumption (no ultrasonic emission).
  • easy deployment and installation
  • a limitation of on-board electronics (receiver circuits only, not transceiver).

Thus, the development of an autonomous wireless distributed sensor network for damage detection and imaging based on ambient acoustic field correlation could prove to be a realistic challenge.

One of the main applications considered is naturally the integrated health monitoring of structures that, due to their conditions of use, would be the seat of acoustic wave propagation at frequencies compatible with the application. A notorious example is that of aeronautical structures during flight, in which such an acoustic field is created by turbojet engines and local aeroacoustic phenomena, but also structures and infrastructures related to other modes of transport (notably railways, for which ambient acoustic fields are created for example by the wheel-rail contact).

The ambition here is to cover the entire problem, from upstream research on the physical principles associated with the imaging principle to the implementation of the sensor network based on these principles, with consideration of operational constraints (inter-sensor communication strategy, autonomous power supply by energy recovery, wireless communication...)

Major recent results

A number of important results have been obtained on the use of non-synchronized (thus possibly natural or "ambient") acoustic sources for the characterization and imaging of plate-like structures :

  • Estimation of structural parameters of a reverberant medium from extraction of statistical properties of Green's functions
  • First image of local damage detection and extended defect (corrosion) tomography by passive methods, and quantification of the role of reverberation in passive reconstruction quality.

Figure 1: Example of passive defect localization from frictional noise

  • Development and validation of an original method to characterize a scatterer (local fault), from the processing of reverberation codas measured on arbitrarily distributed sensors.

Figure 2: Comparison between the experimentally estimated effective cross-section (blue dots) and the theoretical value (red curve) as a function of frequency, for a hole of diameter 10.5 mm

  • Development and demonstration of a nonlinear acoustic conversion principle allowing passive damage imaging from ambient low frequency vibrations.

Figure 3: LF-US conversion using rubbing contact resonators

This work has been carried out in particular in the framework of the projects and collaborations below:

  • ANR PASNI white project (2012-2016), led by IEMN/TPIA, partners: Institut Langevin Paris, LaMCoS Lyon.
  • Collaboration with CEA (joint theses) and University of Rome la Sapienza (joint publications).

Work in progress - Prospects

The results obtained to date have been obtained under laboratory conditions. Future work will focus in particular on the problem of applying the developed techniques in real conditions. In particular, the exploitation of the non-linear properties of the damage could allow a greater robustness in the detection. This is the basis of the ANR PANSCAN project, started at the end of 2018 (leader IEMN, partners Institut Langevin, LaMCoS, support AIRBUS).

In addition, the cross-fertilization of skills developed at IEMN suggests the potential of this theme to federate cross-disciplinary activities, in order to lead to realistic and convincing solutions especially in the context of transport-related applications.

Collaborations

Langevin Institute, LaMCoS, University of Rome la Sapienza, CEA, Airbus.
IEMN interns: COMNUM, AIMAN groups.