Case study card:

Digital twin for prescriptive quality

Digital twin for prescriptive quality

Vertical

Primary Industries and Infrastructure

Business Challenge
Soybean Flour was facing quality control problems. Quality is measured mainly through the degree of Protein, Moisture, Fiber and Oil.
The challenge was to reduce the Moisture standard deviation of 0.3% equivalent to obtaining 11.8 - 12.1% moisture)
Prediction must be made before the whole batch is wasted.

Supplier Solution
The supplier simulated the production line to train hundreds of machine learning models simultaneously.
The model analyses millions of data relations and generates predictions 7 hours before a quality deviation could occur
The models prescribes the adjustments needed to avoid the deviation
It runs on the cloud (AWS) and receives the streaming in real-time from the process variables
A dashboard with customized UX/UI was built so operators can follow recommendations suggested by the model

Industry

Manufacturing

Client

Confidential

MCCai_short_logo_black.png
MCCai_long_logo_black.png