Forecasting the quality

What if you could know what is going to happen in your process hours from now?

What if this information can make you able to adjust the batch before it was too late?


The factory uses organic raw materials in its production. This will cause natural variations and creates major challenges in producing consistently high-quality products. Failing to meet the quality standards will lead to a lower price or potential a wasted batch.


A forecast model was built to predict viscosity 1-2 hours into the future. This allowed the operators and process engineers to understand if the batch would be inside acceptance criteria for a good batch before the batch was produced. This further allowed them to change the recipe so the batch would be accepted. The model was then replicated to predict the viscosity in all silos in the factory.

The first 7 batches are normal, then one "thin" batch and a "thick" batch. These two batches are unwanted. The Intelecy forecast model predicts this before the silo/tank is full, and it would be possible to adjust the viscosity to fit inside acceptance criteria. adjust the viscosity to fit inside acceptance criteria.

bilde fra produkt forecasting.jpg

Result: The forecasting models built using Intelecy is now running live and is now continuously predicting the viscosity levels. If the slurry is going to be too thin or too thick, the operators are being notified, and they can make informed decisions and adjust the mixture ratio and other parameters to ensure they hit the right quality.

With the Forecast tool in Intelecy, operators have optimized the process. They have avoided hours of lost production, stabilized quality, and reduced energy and raw material use.