Verco’s brand new machine learning based platform, Reduct, is currently being used on a series of trials looking at real world optimisation examples. This is the second in a series of bulletins showing live examples of Reduct’s ability to reduce workload, reduce cost and reduce carbon.

Reduct optimises process water consumption.

For this project we set up an automatic data stream into Reduct, capturing hot water consumption for a single piece of equipment which consumes large amounts of hot water. This process is part of a meat processing facility and runs for 24 hours/day, 5 days/week.

AI and pattern recognition remove the need for complex parameters and human data review.

A conventional platform without integrated Machine Learning would require a series of rules to split out production and non-production days. A 24/5 production is a straight-forward concept for a person but there are complexities:

• Receiving alerts on abnormal behaviour for a profile like this can be difficult without a system that dynamically adapts.
• Occasional overtime shifts create seasonal changes in this regular weekly pattern.

Reduct has been built using AI to ensure that it can pick up and predict these patterns without the need to programme lots of parameters into the system.  With this pattern recognition, abnormalities are identified automatically.

Opportunities were being missed as a result of the reliance on human data review.

This particular site has a piece of critical process equipment which is manually controlled. The equipment requires a tank drain and refill each day and the tank has an internal overflow to prevent overfilling. This overflow isn’t visible to the operators.  Erratic consumption was recorded across a series of initial manual readings, so it was suspected that there were issues with excessive water loss.

With an automatic data feed, Reduct quickly alerts on issues and losses are avoided.

After linking an automatic data feed into Reduct, it became clear that there were multiple instances where the tank was not properly isolated after filling.  The equipment was overflowing hot water which was draining away without any visible sign.  Manual readings taken over the previous 6 months show that this had been happening intermittently for some time, but the data wasn’t being reviewed and the manual readings weren’t frequent enough to draw conclusions on the source of the issue.

In our trials, Reduct spotted and alerted on the first instance 30 minutes after it started.  This would have avoided thousands of pounds in excess water, effluent and gas costs had it been in place from the start of the issue last year.

Watch the Reduct walkthrough

See a demo of Reduct in action, the dashboards and reporting outputs.  Get a feel for what Reduct can do for you!

Reduct transforms your data into insights.

Want to know more?  Our Early Adopters Programme is now full but due to high demand, we’re creating a second wave of Reduct early adopters in Q2 of 2021 and we would love you to join us: