Problem Statement

At a large aircraft manufacturing company, during Wing assembly stages, small gaps were seen between the interfacing components. These gaps had to be filled by a manual ‘overhead operation’ called Shimming. This operation was a significant disruption to enable the desired ramp rate and increased the manufacturing costs. Hence the objective of this activity was to understand the root cause leading to gapping and extract non-intuitive insights to reduce shimming process.

Solution Main Activities

  • Collected key product characteristics data for all the interfacing components (i.e. Wing Covers, Ribs etc.), Jig calibration data and Wing deflection
  • Combined previously isolated ‘historical data set’ with the corresponding ‘events in real life’ i.e. gaping during assembly
  • Developed supervised machine learning algorithm to identify the root cause characteristics leading to gaping

Key Achievements

  • Predictive data analytics that enabled the aircraft manufacturer to reduce and eliminate cost of non-quality i.e. reduce amount of gap-filling per wing assembly and enable production ramp up
  • Prediction algorithm predicting gap filling cases with 95 % accuracy to enable accurate workforce deployment planning
  • Spread awareness about benefits of data analytics techniques applicable to shop floor environment
  • Promote a culture for data curation and data harvesting within the business