MSCA-COFUND-CLEAR-Doc-PhD Position#CD22-38: Study and treatment of the environmental perturbations impact on distributed fiber optic sensors using AI solutions

State of the art

To date, the continuous development in distributed fiber optic sensors (DFOS) technologies for physical parameters measurements (e.g. temperature, mechanical strain, mechanical and acoustic vibrations) has enabled the deployment of these sensors in real applications in numerous domains (e.g. civil engineering, geotechnics, intelligent structures) [1-6]. Therefore, DFOSs can be particularly suitable for instrumentation in areas that are difficult to access or in restrictive environments where conventional sensors are not functional. In the case of complex and large-scale civil engineering structures, the DFOS instrumentation network and the installation geometry require the compatibility between the following three components:

1.The optoelectronic interrogation unit:

This unit is the "brain" of the network that carries out an interrogation on the optical fiber core and translates the intrinsic measured characteristics into strain or thermal values. The interrogation techniques are based on optical time (or frequency) domain reflectometry; making it possible to obtain the profile of strain and/or temperature along the optical fiber with tens of thousands of measurement points. This unit is mainly characterized by:

-the investigated backscattering phenomena (Rayleigh, Brillouin or Raman)

-the optical power unit and the detection sensitivity

-the electronic processing unit capacity

These characteristics would allow, depending on the type of the sensing optical fiber, to specify the range, the spatial resolution, the mechanical/thermal sensitivity and finally the acquisition frequency.

2.The fiber optic sensing cable (i.e. neuron) installed in a host medium:

Consists of one optical fiber or more assembled in a single cable. The different layers of sensing cable should protect and fully transfer the structural strain/thermal field into the optical fiber core. The sensing cable design should satisfy all the constraints related to host medium the installation, conditions of use and operation [7-9]. DFOSs that are based on optical reflectometry techniques explore the optical fiber as uniaxial sensor along its axis of symmetry [10-11]. "Neuron" is characterized mainly by its geometry, materials, mechanical/thermal characteristics, mechanical adhesion/thermal conductivity between the various layers of the cable and their long-term chemical stabilities.

3.The interconnection cables:

The connection fibers must guarantee an optical signal transfer between the optoelectronic unit (i.e. brain) and the sensor cable (i.e. neuron), without degradation. It may contain active and passive optical components (e.g. optical switches, optical amplifiers) to distribute the optical signal to the different “neurons” depending on the complexity of the instrumented structure. However, the implementation of an efficient and sustainable DFOSs network into infrastructures depends on the following instrumentation objectives:

-Providing structural health alerts

-Detection, quantification, classification of anomalies (e.g. cracks in concrete) and monitor their evolution

-Measurement of strain, mechanical vibrations and temperature with high precision [6]

This implantation works perfectly for all structures in laboratory conditions and in controlled environment. However, as soon as the structure is exposed to external or environmental disturbances, the different measurements become unusable and the instrumentation network becomes non-functional. Up today, the majority of DFOSs applications in large structures instrumentation have been carried out with low external disturbances and limited durations.

In general, the acquired information delivered by the optoelectronic interrogation unit needs post-processing operations with a significant investigation time. This procedure becomes very complex for massive data in the presence of various disturbances; particularly, the instrumentation of large structures and various parameters.

Therefore, the use of DFOS has been limited to specific cases such as linear installation of optical fibers, low environmental disturbances within a short term. IMSE Lab. has developed its expertise to deliver an accurate measurement of the strain and thermal variations profile that allows early identification of anomalies in different types of structures and applications [12-17].

Scientific objectives:

The objective of this thesis will be to study and validate the concept of a “smart” DFOS: it is required to develop a method with massive processing of measurements obtained by optical fibers in real structures with a complex geometry, that may undergo various environmental disturbances. The main expected results are: an accurate profile of mechanical strain and thermal variations and an automatic diagnostic of the opening of cracks in contact with the sensing optical fiber and the identification of microcracks.

Therefore, we are considering the following three axes:

a-Optical computation:

To work on optical signal processing in order to improve signal quality and translate the acquired data. Analyzing the instrumentation network considering the parameters of the optoelectronic interrogation unit, connection cables and sensors. We will mainly use the optical frequency domain reflectometry technique based on Rayleigh backscattering. This technique provides access to the intrinsic data of the optical fiber with millimetric spatial resolution; particularly the evolution of the polarization of light in the fiber during its propagation. Additional methods like Brillouin based techniques will be investigated.

b-Experimental study:

Fabrication of different samples of reinforced concrete (i.e. host medium). DFOS sensing cables will be installed inside the samples during manufacturing, supplementary cables would be installed on the surface afterword. In the meanwhile, conventional instrumentation will also be put in place. External mechanical disturbances will be introduced via piezoelectric components, cyclic thermal variations and then a controlled mechanical degradation of the samples. We will analyze the sensor performance with different optoelectronic transducers and specially the photoreceiver unit. Integrated semiconductor optical amplifier and photodiode (SOA-PD) will be used and compared to classical solution EDFA-PD (Erbium Doped Fiber Amplifier – PD).

c- AI implementation:

Implementation of algorithms to obtain strain profiles as a function of the temperature of the host medium to identify cracks and microcracks. In this phase, the development of a mechanical and thermal transfer function of optical fiber sensing cables should be studied. These algorithms will permit to optimize the post-processing of the acquired signals.

References

[1] https://doi.org/10.1088/0957-0233/10/8/201

[2] https://doi.org/10.1007/1-4020-3661-2

[3] https://doi.org/10.1016/j.protcy.2016.08.065

[4] https://doi.org/10.1016/j.jrmge.2015.01.008

[5] https://doi.org/10.1016/j.proeng.2016.08.833

[6] https://doi.org/10.1201/9781315119014

[7] https://doi.org/10.1061/(ASCE)0733-9399(1998)124:4(385)

[8] https://doi.org/10.3390/s19030742

[9] https://doi.org/10.1061/(ASCE)ST.1943-541X.0002209

[10] https://doi.org/10.1016/B978-0-08-102181-1.00005-8

[11] https://doi.org/10.3390/s17040667

[12] A. Tixier, Université de Grenoble 2013

[13] H. MOUZANNAR, Université de Lyon 2016

[14] D. A. Ho, Université de Lyon 2017

[15] V. RAMAN, École centrale de Nantes 2017

[15] J. G. De Sauvage, Université de Lyon 2018

[16] I. ALJ, Université Paris-Est 2020

[17] https://doi.org/10.1117/12.2229057