Proactive Power Plant Maintenance
Shifting the focus of data collection and analytics to the edge of the network improves equipment reliability while preventing costly, and potentially disastrous, downtime
The unexpected failure of a steam turbine can create substantial disruption, damage, and economic loss, not only to its power plant but also the downstream power grid and the business and customers who depend on a reliable and continuous source of electricity. To prevent this, predictive maintenance has become a key part of the modern maintenance department. More power utilities are deploying such processes to maximize the reliability of their equipment by detecting potential failures before significant problems arise.
With predictive maintenance, any variance from the “expected behavior” can trigger an alert. These triggers could be based on predetermined performance rules, such as temperature or vibration outside of an acceptable range or be based on deployment of predictive models at the edge that account for historical data. The use of predictive analytics enables users to explore potential root causes and appropriate remedies.
Power plant maintenance is a major undertaking so deployment of predictive analytics at the edge allows power plant operators to proactively schedule maintenance in a non-disruptive fashion to avoid outages altogether and to do so in the most cost effective manner. One of the advantages of an intelligent edge solution is that it can respond immediately to any anomaly or malfunction before it becomes a much more serious and costly issue. It also eliminates the need to upload massive amounts of data to remote data centers because many maintenance issues can now be resolved on site using analytics and processing applications deployed at the edge.
The data center or cloud will continue to play a key role by providing overall system management as well as analytics and machine learning functions that require the computing power of a cloud/datacenter. However by relying more on edge intelligence for critical latency-sensitive functions, utility companies can reduce overall bandwidth and data processing costs and operational delays by sending only non-critical information to the data center or cloud for analysis.