Integrated Modelling & Testing of Structures for Smarter Decisions
Lead Research Organisation:
University of Liverpool
Department Name: School of Engineering
Abstract
The goal of this project, which is a CASE award with Airbus, is to reduce costs through smarter integration of measurements and predictions based on computational models of asircraft structures and to improve the quality of decision-making associated with design and maintenance.
The research will build on recent work at the University of Liverpool on validation methodologies for computational solid mechanics models involving national labs, industrialists and universities from five countries. Previous work has established a validation procedure for simple structural elements based on strain field data acquired using camera-based systems, such as digital image correlation or thermoelastic stress analysis. Initial work in this new project will focus on applying this work to a simple aerospace component using existing measurement and predictions. Subsequent work will examine more complex elements, such as a bonded or riveted joint, and consider the minimum requirements for data collection in physical tests to provide a reliable validation of computational models. The effective use of statements describing model validity in decision-making will be considered, including updating the information contained in such statements as more data becomes available. In later stages of the research the transfer of validity through a modelling hierarchy from elemental components to complete systems will be investigated.
The main focus of the research will be the effective integration of data from computational mechanics models and tests involving experimental mechanics. An element of philosophy will be required in considering the nature and acquisition of knowledge from computational models.
The research will build on recent work at the University of Liverpool on validation methodologies for computational solid mechanics models involving national labs, industrialists and universities from five countries. Previous work has established a validation procedure for simple structural elements based on strain field data acquired using camera-based systems, such as digital image correlation or thermoelastic stress analysis. Initial work in this new project will focus on applying this work to a simple aerospace component using existing measurement and predictions. Subsequent work will examine more complex elements, such as a bonded or riveted joint, and consider the minimum requirements for data collection in physical tests to provide a reliable validation of computational models. The effective use of statements describing model validity in decision-making will be considered, including updating the information contained in such statements as more data becomes available. In later stages of the research the transfer of validity through a modelling hierarchy from elemental components to complete systems will be investigated.
The main focus of the research will be the effective integration of data from computational mechanics models and tests involving experimental mechanics. An element of philosophy will be required in considering the nature and acquisition of knowledge from computational models.
People |
ORCID iD |
| Eann Patterson (Primary Supervisor) | |
| Antonis Alexiadis (Student) |
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509693/1 | 01/10/2016 | 30/09/2021 | |||
| 1795047 | Studentship | EP/N509693/1 | 30/09/2016 | 30/03/2020 | Antonis Alexiadis |
| Description | The main outcomes are two. The first is the development of two metrics that can be used to assess the validity of a computational simulation when compared to a field of measurements given the uncertainty in the latter. This results in improved evaluations of the capability of a model to represent the real world leading to better assessments of the predictive power of computational solid mechanics models. The second is the development of a method that can accurately represent the measurement uncertainty of the measured field (displacements or deformations) in a low-dimensional or feature vector space. The representation of a spatial field into a lower-dimensionality domain has already been established in engineering through numerous publications. This technique extends the potential of the developed techniques by allowing the measurement uncertainty to be represented as a distribution in the same domain. |
| Exploitation Route | It can be used across disciplines where spatial, gridded measurements are available and employed to characterize the validity of corresponding model predictions. Moreover, it can be used to assess whether significant changes take place in temporally evolving phenomena. These phenomena can range from engineering (e.g. the development of a crack in a structure) to oceanography (characterisation of the El Niño-Southern Oscillation) . These developments are important in situations where decisions associated with the capacity of a model to predict the real world must be made. The benefits of the developed techniques are that they can be easily configured to provide solutions to problems across different disciplines where gridded fields are used. |
| Sectors | Aerospace, Defence and Marine,Environment |
| Description | A presentation at the Winter School on Machine Learning held at Uni. of Liverpool's Risk Instiute. |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | About 40 people attended the Risk Institute Winter School held at the Risk Institute at the University of Liverpool. Most of them were postgraduate students and academics. This resulted in a constructive discussion regarding the uses of the developed techniques across different disciplines. |
| Year(s) Of Engagement Activity | 2020 |
| URL | https://riskinstitute.uk/events/machinelearning/ |
| Description | Presentation at the BSSM's 13th International Conference in Experimental Mechanics (Southampton |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | More than 60 postgraduate students and academics attended the presentation. It resulted in a series of questions regarding the application of the developed method in model validation using spatial measurements. Received feedback and exchanged ideas with members of the audience. |
| Year(s) Of Engagement Activity | 2018 |
| URL | http://www.bssm.org/uploadeddocuments/Conf%202018/2018%20abstracts/55BSSM_ALEXIADIS_FINAL.pdf |