Before
Glue manufacturing is a complicated process, where tiny changes in the control parameters can cause serious quality issues. The process to ensure quality is heavily people dependent which is very costly.
After
An automated product quality inspection system which has increased quality levels and saves £20m per plant annually.
The How
- Collect control data from the PLCs along the production line
- Build a multiple input and output regression model with the Gaussian process
- Predict the product quality in terms of stickiness
- Use this information to gauge the current quality levels and also highlight any potential issues that may be on the horizon
