To date, there has been limited investigation of how the differences between assumed and foreseeable conditions translate into the potential for extended operating lifetime of structures. Consequently, it is difficult to assess the probability of the assumed life extension being achieved, particularly as life-limiting factors are likely to differ between projects and designs. This is a very important aspect in different engineering applications for the adoption of a new circular economy action plan that aims to ensure that the resources and structures are kept operational as long as possible.
The rationale is that integrated continuous monitoring systems allow learning from the past to decide in the present about the structural integrity (short- and long-term diagnosis), and to predict in the future the remaining useful life (long-term prognosis). In recent years, research on Structural Health Monitoring (SHM) has sought for solutions to close the loop between designing, manufacturing, building, and maintaining structures driven by continuous measurements of structural data. However, reliable higher-levels in the SHM hierarchy [1] (i.e. quantification and estimation of the useful remaining life) need further developments. In particular, they are only possible by merging physics-based models that underline the mechanics of the time-variant evolution of the structure with data measurements from the operational structure [2].
In practice, there are two main Challenges: (i) Robust extraction of dynamical features (DF) for continuous monitoring which are insensitive to Environmental and Operational Variabilities (EOVs) [3]. These DF should be interpretable during the entire evolution of the structural performance and they should be able to accommodate dimensionality and complexity reduction of their associated non-linear time-variant nature. (ii) And there is a need of developing measures to quantify and propagate uncertainty towards the estimation of the remaining useful life on structures.
The main aim of the project is to reformulate the current techniques to make robust and reliable long-term monitoring by accommodating hybrid models to describe better the time-variant evolution of structural engineering systems. Moreover, the physics-based model will guarantee that predictions made, at future stages of the structure, will adhere to known underlying physical laws of the system model evolution. In particular, the project will evaluate the level of complexity needed of the physics-based models by means of different model order against their capabilities for long-term monitoring. In order to address the many challenges of the project, we shall adopt the following objectives:
O1 - Development of a methodology that can facilitate the extraction of adaptive features for robust continuous and automated monitoring adequate for a time-variant structure evolution.
O2 - Incorporate principled means of inference (physical principles) that can facilitate the interpretability of dynamic features on structures with unknown excitation and non-stationary dynamics.
O3 - Evaluate the contribution of including prior knowledge of the physics-based model coupled with continuous data measurements for predicting future states of the structure and estimation of remaining useful life.
This computational project is supervised by Dr David Garcia Cava (School of Engineering, University of Edinburgh). It will involve regular interaction with collaborators from academia and industry. Interested candidates may contact the supervisor for further information (david.garcia@ed.ac.uk).
References
[1] Farrar, Charles R., and Nick AJ Lieven. Damage prognosis: the future of structural health monitoring. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 365.1851 (2007): 623-632.
[2] Cross, E.J., Gibson, S.J., Jones, M.R., Pitchforth, D.J., Zhang, S., Rogers, T.J. (2022). Physics-Informed Machine Learning for Structural Health Monitoring. In: Cury, A., Ribeiro, D., Ubertini, F., Todd, M.D. (eds) Structural Health Monitoring Based on Data Science Techniques. Structural Integrity, vol 21. Springer, Cham.
[3] García Cava, D., Avendaño-Valencia, L.D., Movsessian, A., Roberts, C., Tcherniak, D. (2022). On Explicit and Implicit Procedures to Mitigate Environmental and Operational Variabilities in Data-Driven Structural Health Monitoring. In: Cury, A., Ribeiro, D., Ubertini, F., Todd, M.D. (eds) Structural Health Monitoring Based on Data Science Techniques. Structural Integrity, vol 21. Springer, Cham
Further Information:
The University of Edinburgh is committed to equality of opportunity for all its staff and students, and promotes a culture of inclusivity. Please see details here: https://www.ed.ac.uk/equality-diversity
Closing Date:
Principal Supervisor:
Eligibility:
Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree. Further information on English language requirements for EU/Overseas applicants.
Applications are particularly welcome from candidates expecting to receive a first-class degree in mechanical engineering, physics, applied mathematics or a closely related subject.
Interests on: Structural mechanics and dynamics, Stochastic modelling and uncertainty quantification, understanding environmental and operational variabilities and their impact in structures and structures for renewable energy is particularly welcome.
Funding:
Applications are welcomed from self-funded students, or students who are applying for scholarships from the University of Edinburgh or elsewhere.*
Further information and other funding options.
*Competition (EPSRC) funding may be available for an exceptional candidate but please note you must be a UK student or an EU student who has lived in the UK for at least 3 years.
