Global Brewer

16.3% improvement in throughput resulting in $10.2M in savings

Global brewer solves 20-year downtime problem by analyzing thousands of alarms per shift.
modeling industrial data
Challenge

High speed packaging triggers thousands of alarms, many of which in turn trigger other alarms, so that is impossible to identify causes of micro-downtimes as they are happening

Solution

Apply a blend of natural language processing, unsupervised learning, and sequence analysis to analyze alarm data and identify on a continuous basis causes of downtime.

Outcome

Identified insights to reduce downtime events resulting in 16% throughput improvement