Alpha-i wins the Rolls-Royce challenge of AI automated decision-making for aero-engine inspection

At Rolls-Royce’s manufacturing facilities, inspection is carried out to measure the moment weight of rotating components, to determine optimal alignment to the engine.

High conformance standards currently entail high failure rates for the components, requiring rework that causes disruption and increased manufacturing costs.

The challenge was to produce an AI algorithmic model that allowed automatic adjustments throughout the manufacturing process to increase the conformance rate.

Challenge Highlights

  • 5 years of data:
    • 25,000 components
    • 7-week manufacturing line
    • 4 monitored processes
  • Evaluation based on:
    • Expected savings
    • Innovation
    • Feasibility
    • User Experience

Alpha-i was declared the winner of the challenge as the solution with the highest potential to improve and automate quality control for critical components in the aerospace industry.

We achieved this by combining AI algorithmic insights with a user friendly platform that could support engineers in identifying causes of non-conformance and in continuously improving the manufacturing process.


The event was organised and sponsored by:

The Rolls Royce Manufacturing Hack gave us insight into the challenges faced by the industry and allowed us to address these problems using data science. Producing a solution in a short time was exciting and it was great to see creative solutions from other teams as well. First, we spent time understanding the problem by analysing the data and also validating our solution with the domain expert. Next, we collated the data together in a practical format and finally developed machine learning models to perform anomaly detection. This proved challenging due to the lack of abnormal examples, but we overcame this using various statistical techniques. When presenting the front end of our solution, we focused on how it would help Rolls Royce’s engineers on a day to day basis. Throughout the process, we constantly liaised with the domain expert to ensure our solution would add tangible value. Through this hackathon, I have learned that good communication with the customer is vital in order to develop a viable solution.

William Tai – Machine Learning Engineer