PhD (M/F) – Structural Health Monitoring using sparse sensor networks and signature-informed modeling

Job Description

Thesis Subject

Profile

- Engineering diploma or Master 2 involving at least one of the following disciplines: computational mechanics, scientific computing, physics-informed artificial intelligence.
- Interest in smart monitoring for mechanical systems, model order reduction, data assimilation, artificial intelligence.
- Motivation for joining an interdisciplinary research project including theoretical, numerical and experimental activities.

Expected skills

- Solid background in scientific computing (numerical methods for PDEs, linear algebra)
- Proficiency in one scientific programming language: MATLAB or Python.

Context

The task of Structural Health Monitoring (SHM) is to assess the integrity of a structure from in-situ sensor measurements [1]. This raises three important questions. The first is to define what information is an indication of structural integrity. It is common practice to focus on physically interpretable parameters which may characterize: a damage, a mechanical property, boundary conditions, an external loading, the onset of an instability, etc. Importantly, complex monitoring problems, involving engineering structures under uncertain operational conditions, require these parameters to be identified within statistical confidence intervals. The second is to determine what prior knowledge on the studied structure is to be used. Relying on a mechanistic model or on a data-base of in-situ measurements leads, in some sense, to orthogonal approaches [2]. On the one hand, mechanistic models are physically explainable, but are limited to relatively simple systems. On the other hand, data-driven models may capture more complex phenomena, but are hardly interpretable. The last is to determine what sensing technology is to be installed on the structure given the constraints. Several criteria may drive the selection process, including: number and location of sensors, overall size and weight, durability and reliability in harsh environment, frugality, compliance with sector-specific regulations, and benefit-cost ratio.
In industrial applications, the benefit-cost ratio often represents a critical barrier to the widespread adoption of SHM. For example, in the context of commercial aviation, the use of SHM remains limited, partly due to challenges in establishing clear business cases [3]. Deploying a dense network of sensors to monitor large areas of an aircraft is not only expensive, but also difficult to reconcile with the stringent regulations governing sensor installation. In addition, achieving high reliability is of paramount importance. Developing minimally intrusive yet reliable smart monitoring systems is therefore a critical and urgent challenge of SHM. Successfully addressing these challenges could unlock economically viable applications for a wide range of applications across various industries (aeronautics, turbomachinery, industrial processes, etc.).
This PhD is part of the funded ANR JCJC project SPARSE-SHM (Sparse structural health monitoring using signature-informed hybrid modeling). The goal will be to develop an innovative SHM framework capable of operating with a very limited number of sensors. The core concept relies on signature-informed modeling. The principle is to develop a reduced-order model capturing just enough spatial and temporal information to estimate parameters of interest with uncertainty quantification. A proof of concept has been demonstrated in [4] for the industrial problem of localizing impacts on composite aircraft fuselages. To further improve the approach and open new smart monitoring applications, data assimilation [5, 6] and artificial intelligence techniques [7, 8] will be used to enhance adaptability to varying operational conditions.

PhD objectives

This PhD opportunity involves theoretical developments, numerical methods, and experimental validation. The candidate will:
- Formalize the concept of signature within a reduced order modeling framework, enabling the development of fast surrogate models that encode critical parameter information.
- Implement data assimilation techniques to enhance model adaptability, particularly in the presence of nuisance factors such as unmodeled operational variability.
- Design an artificial intelligence pipeline to efficiently extract signatures from online measurements.
- Validate the SPARSE-SHM concept on realistic experimental demonstrators.

The proposed methods will be tested across a spectrum of numerical models relevant to smart monitoring, from dissipative vibratory systems (passive SHM) to wave-dominated systems (active SHM).

References
[1] C. R. Farrar and K. Worden. An introduction to structural health monitoring. In: Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences (2007).
[2] R. Liu et al. Dynamic load identification for mechanical systems: A review. In: Archives of Computational Methods in Engineering (2022).
[3] D. M. Steinweg and M. Hornung. Cost and Benefit of Scheduled Structural Health Monitoring for Commercial Aircraft. In: Proceedings of ICAS 2021, Shanghai, China (2021).
[4] D. Goutaudier et al. Long-range impact localization with a frequency domain triangulation technique: Application to a large aircraft composite panel. In: Composite Structures (2020).
[5] Y. Maday and T. Taddei. Adaptive PBDW approach to state estimation: noisy observations; userdefined update spaces, In: SIAM Journal on Scientific Computing (2019).
[6] G. Revach et al. KalmanNet: Neural network aided Kalman filtering for partially known dynamics. In: IEEE Transactions on Signal Processing (2022).
[7] Le-Khac et al. Contrastive representation learning: A framework and review. In: IEEE Access (2020).
[8] E. J. Cross et al. Physics-informed machine learning for structural health monitoring. In: Structural Health Monitoring Based on Data Science Techniques (2022).

Your Work Environment

The PIMM Laboratory is a joint research unit of CNRS, Arts et Métiers and Cnam, dedicated to innovation in the fields of mechanical engineering, materials science and advanced numerical simulation. Located in the heart of the 13th arrondissement of Paris, the laboratory offers a privileged balance between dynamic university life and lively city life.

Compensation and benefits

Compensation

2300 € gross monthly

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions