Rolls-Royce is renowned for its luxury cars, but it’s also a leading engine supplier in the business aviation sector. The company designs and manufactures cutting-edge engine blades and entire power systems for aircraft worldwide.

Thanks to modern factories outfitted with extensive IoT sensors, Rolls-Royce gathers vast amounts of data on everything from production workflows to engine performance.

Turbine Insight

In Rolls-Royce’s Advanced Blade Casting Facility (ABCF), the Turbine Insight project tackled a core challenge: how to reduce production defects in the creation of turbine blades.

These specialized blades, designed to withstand temperatures higher than their own melting point, demand flawless execution from start to finish.

Initially, the abundant data from factory sensors—covering each station in the multi-step process—was manually processed in Excel. This approach offered little room for detailed analysis or proactive insights.

By examining correlations across thousands of sensor points, I identified potential culprits for defects—anything from a slight temperature deviation to an unexpected fluctuation in cooling times.

Armed with these findings, I orchestrated simulation scenarios that predicted how process changes would impact final product quality.

This gave the factory a data-backed way to optimize blade creation without halting production. With this integrated, iterative approach, the project significantly cut down on faulty blades, boosted overall efficiency, and established a new blueprint for operational excellence at Rolls-Royce.

I led the effort to modernize data handling by setting up an on-premises platform capable of ingesting and storing hundreds of gigabytes of daily sensor information. This involved configuring robust data pipelines that cleaned, validated, and organized the information for further analysis.

Next, I reconstructed the entire production line in a digital format, allowing stakeholders to track every stage of blade creation in real time. Working closely with engineers, I translated the intricate metallurgy and physics involved into quantifiable variables for Machine Learning.

Keywords:

IoT, Big Data, Data Management, Spark, Machine Learning, Simulation

Predictive Maintenance

DRolls-Royce’s factories rely on modern, highly automated equipment to streamline production steps. Any downtime, however, can cause major delays and incur significant costs.

To avoid unexpected breakdowns, devices require routine checkups during scheduled maintenance windows.

Yet predicting which machines might fail next has historically been a guessing game, leading to unnecessary repairs or missed issues.

I introduced a Predictive Maintenance system, consolidating IoT data from equipment sensors and historical performance logs.

I then built and trained a Machine Learning model to forecast potential risks or imminent failures. By implementing semi-automated data collection and integrating these insights into existing workflows, I ensured that maintenance teams could focus on the machines most likely to need attention.

Ultimately, this approach helped Rolls-Royce decrease downtime, save costs, and keep production plans on schedule.

Keywords:

Predictive Maintenance, Big Data, MLOps, Data Analytics, Visualization

Previous

Deutsche Bank
Next

Federal Foreign Office