Duke Energy
Duke Energy optimizes predictive maintenance of its power plants with AVEVA Predictive Analytics
Duke Energy, a U.S. utility, operates more than 60 power plants across seven states, ranging from coal to simple-cycle combustion turbines, combined-cycle and integrated gasification plants, serving more than 7.2 million customers with 58,000 MW of generating capacity. The group's monitoring and diagnostics center was tasked with optimizing asset performance using the best available digital technology, while remaining independent of existing software and hardware to minimize replacement costs. Deploying AVEVA Predictive Analytics now allows five analysts to oversee more than 87% of the generating fleet through more than 11,000 models and more than 500,000 data points, as part of a predictive maintenance approach. The solution monitors alarms, trends and in-depth analyses in line with EPRI's reliability framework. The results are compelling, with more than $34 million in savings from a single early catch event in 2016.
Integrated AVEVA software
FAQ
How does AVEVA Predictive Analytics monitor 87% of a 58,000 MW generation fleet?
AVEVA Predictive Analytics enables just five Duke Energy analysts to monitor more than 87% of the generation fleet using more than 11,000 models and 500,000 data points. The platform automates the detection of behavioral deviations from assets' normal signatures. Without this industrialization of analysis, dozens of experts would be needed to achieve the same level of coverage. It is an essential skills multiplier for utilities operating dozens of heterogeneous power plants.
Why is a Monitoring & Diagnostics Center strategic for a power generator?
An M&D Center centralizes asset diagnostic expertise for the entire fleet, preventing each plant from building its own analysis capabilities in isolation. At Duke Energy, the center relies on AVEVA Predictive Analytics to monitor more than 60 power plants across seven states, in line with EPRI's reliability framework. This organization encourages the sharing of best practices, speeds up the resolution of recurring issues and ensures a consistent level of expert analysis, whatever the type of asset monitored.
How can the ROI of a predictive analytics project in power generation be quantified?
Duke Energy saved more than $34 million from a single early-catch event in 2016. This pre-failure detection on critical equipment illustrates the operational value of AVEVA Predictive Analytics: avoiding an unplanned outage means saving an entire power output and emergency repair costs. Across a 58,000 MW fleet serving 7.2 million customers, the cumulative value of early detections is considerable and recurs year after year.
How can predictive analytics be integrated without replacing existing systems?
Duke Energy wanted to remain independent of existing software and hardware to minimize replacement costs. AVEVA Predictive Analytics acts as an analytics layer compatible with data sources already in place (DCS, historians, SCADA). This non-intrusive approach delivers predictive capabilities quickly without a costly infrastructure overhaul. It is a key adoption factor for utilities that maintain complex industrial architectures that are slow to evolve.
How can AVEVA help move from reactive to predictive maintenance?
AVEVA Predictive Analytics continuously monitors alarms, trends and in-depth analyses from low-cost sensors and high-fidelity connectivity. The solution alerts teams days, weeks or months before a failure, making maintenance something that can be planned. For Duke Energy, this foresight has made it possible to optimize maintenance windows, reduce unplanned outages and improve the availability of a fleet that is essential to supplying electricity to 7.2 million customers across seven states.