To support its clients along these areas, Sopra Steria has developed multiple AI assets and accelerators that are used as catalysts to reduce project development cost and accelerate their time-to-market.
#1: Process adherence
In the Aerospace & Defence industry, manufacturing steps are carefully designed with respect to upstream engineering requirements in order to ensure product reliability. The design of each part or component is indeed created with respect to different
constraints, in particular performance, norms, traceability and certifications. These come also with precise manufacturing processes and steps to guarantee the conformity of the product to its quality requirements.
As a consequence, deviating from standard manufacturing processes that have been defined by engineering for quality purposes generally increases the risk of non-conformity. Tracking these deviations and measuring process adherence with Process Mining
can be done by comparing the logs of activities occurring in the MES with the theoretical process as defined in Manufacturing Engineering.
Beyond process mining and by integrating other relevant features, an AI module can identify the correlation between the contexts of the deviation, in particular the environmental parameters, and the insights of the deviation. In that case, when practitioners
decide to implement a corrective action plan, the AI module can also record the action to mitigate the risk of non-adherence and its final outcome.
This AI system is finally able to recommend, given a particular situation, an effective action plan automatically. Sopra Steria, with its AI and process mining capabilities, has developed assets to accelerate the identification of Supply Chain disruptions
due to quality issues, and support the implementation of corrective action plans targeted towards practitioners.
#2: Intelligent Testing
In the field of Quality Management in Manufacturing or Supply Chain, a key best practice is the inspection (destructive or non-destructive) of the produced or delivered components. This enables the statistical estimation of indicators such as non-conformity
ratio for a production batch and provides concrete insights enabling practitioners to get a better visibility on the risk of Supply Chain disruptions.
The main concern is optimizing when, where and how far inspections or tests must be performed. “Intelligent testing by AI” is a tactical decision support tool designed by Sopra Steria that helps quality practitioners sample the right components
at the right frequency. By combining Machine Learning with a powerful probabilistic model of the manufacturing steps and the production flows, we are able to correlate the testing scenarios with a cartography of the quality risks, allowing decision-makers
avoid any fuzzy context for quality traceability.
Moreover, the link between testing costs –number of tests, duration, etc.- and confidence intervals of the risks is determined, which provides key information to optimize the costs of the inspections or tests for a given confidence level. Products
clustering is also used in order to mutualize when relevant, creating a new dimension –the product- of the solution for the decision support. Finally, the architecture of such AI solution is built to have adaptable parameters that evolve based
on past performance and improving its efficiency through time.
#3: Predictive tracking
In Supply Chain Management, anticipation is key to ensure delivery flows and pre-emptively avoid disruptions. The non-conformity performances impacting supply chain operations is consequently part of the information needed to be tracked and anticipated.
To do so, AI is a key technology that provides valuable predictive information.
In particular, with respect to the procurement process, one of the major challenges is improving one’s assessment of the supplier performance, in particular related to the quality of the delivered components. By integrating a large multi-parametric
environment including supplier characteristics, products, local and global industrial contexts but also exogenous context, AI technologies can track the quality performance of the suppliers and anticipate the risks of non-conformity ratios.
Visibility and anticipation are also critical in the field of quality management within the manufacturing processes. Machine Learning technologies are a powerful tool to predict how non-conformities appear and may arise along quality gates. They also
enable a better understanding of root causes and early prevention of non-conformities at downstream quality gates.
Overall, Sopra Steria uses Machine Learning algorithms to better understand and anticipate as well as Bayesian Networks to propagate the risks of quality disruptions. Such AI technologies enable Sopra Steria’s clients to gain predictive insight
into non-conformity risks, as well as detailed information on risk root causes and localization. Supply Chain or Manufacturing Managers can organize accordingly their resources to mitigate non-conformity risks and ultimately avoid unexpected costly
disruptions.
Benefits
Sopra Steria has helped several aerospace manufacturer in the implementation of AI-based decision support tool for Quality management improving Supply Chain performance. For instance, Sopra Steria has developed Suppliers performances tracking tools with
predictive views. Sopra Steria also worked on a decision support tool to predict the potential non-conformity risks by anticipating the trajectory of different key parameters with Machine Learning.
As an order of magnitude, Sopra Steria estimates, based on its analyses and references, the following benefits related to Quality and Supply Chain thanks to its aforementioned assets.