Certification for Design Reshaping the Testing Pyramid

New Post

Start Date 01/08/19 - Finish Date 31/07/24

 

 

 

Developing the scientific foundation that enables lighter, safer and more cost-efficient composite aerostructures.

 

This research project is to provide a route for lessening regulatory constraints, moving towards a more cost/performance optimised philosophy by reducing the multiple coupon level materials tests at the bottom of the test pyramid. It will create the scientific basis for a new culture of virtual certification that will enable significant mass savings and target reduction of design costs and development time.

 

The project is developing a new holistic approach to design at a hitherto unattainable level of fidelity. Bayesian techniques such as Monte-Carlo Markov Chains to quantify uncertainty and variability and Bayesian inference process will be used to minimise model uncertainty. Lock-In Digital Image Correlation and Thermoelastic Stress Analysis will be integrated and scaled up to provide a non-contact methodology to capture stress and strain in the vicinity of as-designed internal meso/micro features and in-service damage, and quantitatively identify the mechanisms that initiate damage and lead to failure.

  

 

 

Recent Updates 

 

Recently, we have successfully built a Gaussian Process based Data-Driven Emulator for Quantifying the defects/uncertainty of Composite structure (C-section), The most significant merits of the proposed scheme include: 1) It directly maps from input wrinkle with different parameters to full solution field displacement without time-consuming stiffness matrix assembling and governing equation solving in each single computational prediction. 2) unlike other machine learning algorithms requiring a huge amount of training data, this emulator is capable to provide the predication (expected value) and, from a statistical aspect, the corresponding confidence interval (standard deviation) for the testing input uncertainty at the same time, even the number of training data is limited

 

 

 

Collaborators

 

 

 

 

 Partners

 

  

 

 

 
 
 
Data Centric Engineering

SUBSCRIBE

​Your email:

This field is required.

You are now subscribed!

 

Contacts

Mail  / dce.group@exeter.ac.uk

Phone  /  +44 (01392) 725899

Follow us on Twitter
Website by GFIVEDESIGN

SUBSCRIBE TO OUR NEWSLETTER

& STAY UP TO DATE

image-1 image-5mainMagicWidget74_ey
Follow us on Instagram

©Data Centric Engineering | All Rights Reserved | Privacy & Cookie Policy

Subscribe
Data Centric Engineering