CIFRE Thesis: Statistical Tolerancing & Big Data Integration (m/f)

Description de l'offre

CIFRE Thesis: Statistical Tolerancing & Big Data Integration (m/f)

Airbus Toulouse

Airbus is a global leader in aeronautics, space and related services. In 2016, it generated revenues of € 67 billion and employed a workforce of around 134,000. Airbus offers the most comprehensive range of passenger airliners from 100 to more than 600 seats. Airbus is also a European leader providing tanker, combat, transport and mission aircraft, as well as Europe's number one space enterprise and the world's second largest space business. In helicopters, Airbus provides the most efficient civil and military rotorcraft solutions worldwide.

Our people work with passion and determination to make the world a more connected, safer and smarter place. Taking pride in our work, we draw on each other's expertise and experience to achieve excellence. Our diversity and teamwork culture propel us to accomplish the extraordinary - on the ground, in the sky and in space.

Description of the job

A vacancy for a placement in Statistical Tolerancing & Big Data Integration (m/f) has arisen within Airbus Toulouse Saint-Martin. You will join the department in charge of tolerancing.

Profil recherché

Tasks & accountabilities

In the context of the CIFRE contract on the following subject: "Improvement of statistical tolerancing in an environment with diverse industrial issues and the integration of feedback in production under the form of Big Data", you will be required to handle the following topics on the scale of a complete aircraft.

Your main tasks and activities will include:

·  Assessing the currently used statistical methods,
·  Improving statistical methods taking into account the context of the targeted plant,
·  Providing recommendations related to tolerance control,
·  Providing methods, recommendations and algorithms for validation of models, tolerancing,
·  Incorporating feedback on production (capability) in models and optimising the sharing of tolérances,
·  Incorporating partial measurements in models and conformity prediction,
·  Summarising the results and proposing a clear and comprehensive graphic interface.

The objectives of this thesis are to rethink and refine the statistical approach of the processing of chain of dimensions.
The advantages of the current method (ASCR) must be maintained, i.e. that it is a quick analytical method based on an RSS approach and implementing a safety coefficient that varies according to the topology of the chain of dimensions.

After the assessment of the relevance of the current method in various environments, the topological coefficient must be completed by an industrial environment coefficient.
Indeed, each Airbus plant is a centre of excellence for a specific type of product, they are therefore subjected to different constraints and sources of variations.
For example, the production of an engine pylon suffers from less extensive deformations than that of an entire section of fuselage.

The approach must not limit itself to the assessment of each Airbus plant at a given time (t) but should provide an assessment grid that allows for a regular review of the assessment of each plant and that can also assess our aerostructure suppliers. Indeed, the failure to obtain the tolerances negotiated with a supplier can jeopardise the industrial project at an aircraft level.

Going from one method to another will allow us to re-evaluate the risk levels of each calculation and identify a potential margin enabling us to perform an upward re-assessment of certain tolerances.

In this context, one of the major unknowns with these purely numerical re-assessment approaches remains the relevance of each 3D simulation model of the assemblies performed, in particular regarding the direction of the measured variations. These 3D models can be seen as black boxes linking inputs to outputs in a linear way. A numerical method allowing to validate 3D tolerancing models afterwards thanks to measurement data, and the proposed direction can be corrected, if necessary (self-learning).

Finally, there is the question of the propagation of results through the tolerance network. This network is currently composed of more than 60,000 interconnected inputs, a Big Data approach will need to be implemented.

The final objective will be to perform a real-time mapping of industrial risks at the level of the entire aircraft, which will be individually updated for each MSN (aircraft identifier) according to the measured unit value, but also the usual measured variability of the variables that are not yet captured.

Required skills

As the successful candidate, you should be able to demonstrate some of the following skills and experience:

·  Educated to a 5-year degree in Statistics/Applied Mathematics, Big Data or a related discipline,
·  A degree in Engineering + a 5-year degree in research would be a plus,
·  Knowledge of machine Learning,
·  Knowledge of engineering and mechanics would be a plus,
·  IT languages: R, Python,
·  Independent, pragmatic, ability to innovate and be a driving force for suggestions,
·  An advanced level of English and French.

This thesis is in partnership with the Toulouse III Paul Sabatier university.