How Predictive Maintenance helped a natural gas company reduce asset downtime and foresee equipment failures
Kalypso was asked to develop a solution that could help predict downtime events in natural gas compressors before they happen. Due to lengthy repair times, our client needed a way to not only gain real-time insights on asset performance but proactively address asset health issues.
A natural gas compressor is the critical beating heart of our clients production, operating 24/7/365 and moving large volumes of gas per day. Downtime can mean thousands of dollars, and because most are located in remote locations the time between dispatching service technicians and repair is significant.
We learned if we could predict events at least 30 mins to 1 hour before the occurrence, our client’s Ops/Facilities team can take appropriate actions to prevent vital asset failures.
Using Predictive Maintenance and Industrial AI, we developed a solution that would identify failure and reliability issues before they occurred. By collecting operational data, real-time field data and historical downtime data, we generated predictive models that tested and validated our dataset again and again, each time learning and improving from previous successes.
The end result was an algorithm that assigned a value between 0 and 1 to indicate the risk of the compressor going down in the next 30 minutes, delivering real-time insights on asset performance.
Our work automated failure detection and early warning notification to help our client react faster, plan and schedule maintenance better and reduce unplanned downtime.
Enabling Predictive Maintenance with Industrial AI Analytics helped our client improve overall machine health, productivity and plant operations.