Statistical Design and Analysis of Aircraft Test Programs
Lead Research Organisation:
University of Bath
Department Name: Mechanical Engineering
Abstract
During flight, carbon fibre aircraft components are subject to variety of environmental conditions (eg heat, humidity and impact) that can cause variable reductions in strength. Equally additives such as vertically aligned carbon nanotubes (VACNTs) can lead to significant improvements in strength. The overall probability of failure of a component is dependent on the likelihood of a reduced strength region of the component being subject to a critical loading.
This project will seek to assess the effect of VACNTs on the probability of failure of a composite component. To establish underlying dependency of strength on variable environmental conditions and VACNT additives the Student will undertake their own environmental test program. Multivariate strength distributions derived from these tests will be combined, in collaboration with Martin Gaitonde of Airbus, with time and component dependent in-flight loading distributions and inspectio regimes to establish components failure probabilities. Techniques such as Quantile and robust regression will need to be developed to address the nonlinear and extreme-value aspects of the multivariate time-dependent program. Bayesian experimental design will aid the envisaged test program.
The aim of the project will be fulfilled if the current 'worst case everywhere' philosophy can be replaced by a probabilistic approach to failure that accounts for stochastic variability and the real likelihood of concurrence of severely detrimental events.
This project will seek to assess the effect of VACNTs on the probability of failure of a composite component. To establish underlying dependency of strength on variable environmental conditions and VACNT additives the Student will undertake their own environmental test program. Multivariate strength distributions derived from these tests will be combined, in collaboration with Martin Gaitonde of Airbus, with time and component dependent in-flight loading distributions and inspectio regimes to establish components failure probabilities. Techniques such as Quantile and robust regression will need to be developed to address the nonlinear and extreme-value aspects of the multivariate time-dependent program. Bayesian experimental design will aid the envisaged test program.
The aim of the project will be fulfilled if the current 'worst case everywhere' philosophy can be replaced by a probabilistic approach to failure that accounts for stochastic variability and the real likelihood of concurrence of severely detrimental events.
Organisations
People |
ORCID iD |
Andrew Rhead (Primary Supervisor) | |
James EVANS (Student) |
Studentship Projects
Project Reference | Relationship | Related To | Start | End | Student Name |
---|---|---|---|---|---|
EP/N509589/1 | 30/09/2016 | 29/09/2021 | |||
1939787 | Studentship | EP/N509589/1 | 30/09/2017 | 29/09/2021 | James EVANS |