Predict Deviations in Operational Performance

Ever wished you knew the moment your equipment started deviating from normal operation?  Perhaps if you had the insight early, you could have adjusted a setting, topped up a level and realigned the course trajectory to avoid equipment downtime.

Large manufacturers have thousands of pieces of equipment and tens of thousands of sensors, each sending data through continuously, making it hard for operations teams to monitor each component.  Up until now, manufacturers have relied on condition monitoring, which alerts an operator when a threshold has been met. However often this alert is too late and failure is imminent. 

With artificial intelligence and machine learning technology, operators can predict the future performance of each system in their plant, detecting when a component of a piece of equipment deviates from normal operation. Deviations typically start out inconsequential, and the AI models study the trends, predicting if they are likely to cause issues or faults.  By altering operations teams early, additional time is gained to plan necessary interventions.  

VROC's AI platform OPUS includes a simple AI model wizard, which operators can use to build deviation models of their systems and processes, these models work tirelessly in the background continually forecasting equipment performance and predicting deviations.  

Watch the video to see a deviation AI model action

Benefits of Deviation Models

Deviation Models learn from historical and real-time data to predict when a critical component of a process deviates from normal operation. This advance prediction provides time to plan interventions which leads to increased operational efficiency, reduced asset downtime, .

The following case studies have all used VROC's AI deviation modelling technique:

- Leaking non-driven end mechanical seal on a produced water pump

- Increased bearing vibrations of a Gearbox

- NDE bearing degradation on a water supply pump

- Gas air heater seal issue on coal fired power plant

- Boiler feed pump shaft bearing misalignment 

- Ventilation fan rusted linkage

- Haul truck tyre degradation from sidewall cut 

No code Deviation Modelling with OPUS

With VROC’s OPUS platform, building and deploying Deviation Models is a seamless process that doesn’t require any coding skills. Engineers and operators can utilize OPUS's model wizard to create sophisticated predictive models.

• No-Code Platform: Intuitive model wizard allows users to build machine learning models without coding.
• For Engineers and Operators: Designed to be used by non-data professionals.
• Rapid Deployment: Quickly build, test, and implement models without lengthy development cycles.
• Model Management: OPUS's inbuilt MLOps ensures models remain accurate, continuously monitoring and analysing equipment performance
• Scalable: Quickly build models for all critical machinery and processes enterprise wide
• Interoperable: Pull data from your existing historian, PLC, SCADA or direct from sensors on your plant, so you can build accurate models