In a complex industrial environment such as aeronautics, meeting delivery schedules is a constant challenge, and OTD measurements quickly become the yardstick for measuring progress. To better understand its potential delivery delays and make its OTD a consistent lever for improvement, an aerostructures subsidiary of a major European aircraft manufacturer called on Expleo’s know-how to develop a predictive data analysis tool better able to capture the complex balance of a supply chain.
The Challenge
As the guarantor of the smooth running of a supply chain, OTD is probably the most important key performance indicator for companies, providing them with an assessment of whether they are meeting the delivery deadlines of their orders. But this is true of all performance indicators; OTD can only be of real value if the data collected and analysed to determine it is of good quality.
The aeronautics industry combines many components necessary for its proper execution –structuring the information associated with it, making it flow smoothly between all parties, and ensuring it is easily readable in its correlations is not easy.
To increase its production rate in the face of growing market demand and better understand and anticipate its potential delivery delays, an aerostructures subsidiary of a major European aircraft manufacturer decided to call on Expleo’s know-how to develop an application capable of better predicting its OTD. Its underlying objective is, of course, to benefit from decision support in its organisational processes.
Solutions
The four-month project led by three Expleo specialists – a data scientist to design the mathematical model, a data engineer to collect the data and an expert already aware of the customer’s organisational processes – is now at the end of its proof of concept with some notable initial results. To do this, the Expleo team, in collaboration with experts from the customer’s supply chain, first had to analyse the many factors that can influence delivery delays.
Thus, the data from HR departments, quality monitoring, supplier inventories, production flow status, etc., were as numerous as they were disparate, and it was necessary to proceed with vast data recovery, cleaning and merging plan using the Python programming language and the TIBCO Spotfire data visualisation software.
By cross-checking and progressively narrowing the funnel according to the information collected relevance, the Expleo team established a machine learning model capable of explaining the delays in previous operational flows and thus better predicting OTD.
Outcome
With this project’s success, our customer is fully convinced of the relevance of using data-driven tools to respond to issues arising within its production line. Thanks to the tool developed by Expleo and the precision of its machine learning algorithms, our client can now easily cross-reference all the essential data sources potentially concerned and derive valuable decision keys for more optimal performance management. With the dashboard offered by the TIBCO Spotfire data visualisation platform, they also benefit from a clear and rapid display of the operational flows ins and outs.
In this project, sorting and analysing the data requires a high level of manufacturing engineering expertise, and this is where Expleo comes into its own. For example, for the study of the one-year follow-up of an aerostructure section on 344 examples of a given aircraft model, only 15% of the mass of collected data is, in the end, usable.
