Freestyle Drink Dispenser Failure Diagnostics
Nuclear Power Plant Predictive Failure
Manufacturing Equipment Diagnostics & Predictive Failure
Power Plant Gas Turbine Failure Prediction
Telecom Equipment Diagnostics & Predictive Failure
Fleet-Wide Vehicle Battery Failure Prediction
Manufacturing Wood Products
Georgia Pacific streams all plant data to their cloud. We were tasked to support two entire mills and build a predictive failure system to increase availability, to alert maintenance crews with sufficient notice and diagnostics to repair at planned outage points. We provided alerts that enabled to the site to achieve guidance for $4M in annual savings…
- Multi-plant Real Time Monitoring
- +3% Added Availability
- 30 Day Average Prediction Time
- $4M Annual Savings Identified
- 22x ROI
Manufacturing Paper Mill
Georgia Pacific paper mills must maintain 100% uptime for weeks at a time. Unexpected bearing failures on rollers is a common problem preventing this goal. Preventative maintenance results in costly replacement at 60% of life. We used client vibration data from existing IoT sensors to build advanced prognostics models that accurately predict and track failure dates six months in advance. Bearing life can be extended and repairs scheduled at planned machine downtime intervals.
- 9 month accuracy on failure warning time
Power Plant Equipment Health
For Southern Company we were able to establish that its advanced machine learning powered could accurately predict anomalies in equipment operation using legacy historian data. This was a competitive bakeoff against the site builder. GENESIS provided results that were superior to the competition.
Drink Machine Fleet Service
Over 40,000 of Coca-Cola’s Freestyle drink machines stream condition data to the cloud. Field service was experiencing excessive pump replacements. Our analytics platform was able to isolate failure type and then identify the cause as a defective production lot mixed randomly in the fleet. Result...
- Identified bad production lot, over 3,000 defective pumps.
Telecom Equipment Repair
Cox Communications, a top 10 US provider of cable based TV and internet service, determined that 20% of residential "truck rolls" were caused by node box failures (see pic). We built node-failure detection models by correlating customer complaint calls with digital node data. These models were able to reduce the unnecessary rolls by 80%. Results...
- $12 Million annual saving by preventing 60% of wasted residential truck rolls.
Fleet Telematics
Verizon Telematics monitors in excess of 20,000 automobiles. Each has a Verizon cellular device connected to the auto’s OBD-II data port uploading data in real time via their cellular network. The challenge was accurately predicting battery failures while adjusting for winter and summer affects on performance so that it could be offered as a service. This provided a $2M annual revenue opportunity.
