Power generation, transmission, and distribution plants and facilities—like many industries—are awash in data, but often lack insights due to the use of insufficient or incorrect software analysis tools. Traditional tools, such as spreadsheets, were originally designed for financial use, but have been pressed into service to analyze time series data. However, these and other general-purpose tools are cumbersome for performing the analytics required to create insights.

A better alternative is to use a purpose-built solution, one designed from the ground up for analyzing time series data. This type of software is commonly referred to as advanced analytics and is available from a number of suppliers.

Figure: Power utilities can use advanced analytics software to create insights from time series data.

Following are a few examples of how advanced analytics software is being used to improve power industry operations across multiple facilities and entire fleets.

A large utility in the southeastern U.S. was experiencing an excessive amount of failures of the bypass main contactors on their wind turbine assets, causing unplanned downtime and necessitating expensive reactive maintenance.

The utility’s maintenance team used advanced analytics software to predict failures before they occurred. This was done by engaging the utility’s turbine experts to identify situations when prior failures occurred, and then comparing these to conditions when the system was operating normally. A machine learning algorithm and automated profiling tool was developed using the software and then back tested on two and a half years of historical data. This resulted in identifying 12 errors, including 11 true positives and 1 case where the asset failure was imminent.

By putting powerful machine learning tools in the hands of subject matter experts, each of whom possessed extensive knowledge of the assets and the signatures of particular failure modes, the utility was able to quickly build a robust model for detection of future problems. This solution was then rolled out across all of the utilities wind farms, providing a fleet-wide solution.

In another case, a major utility in the midwestern U.S. needed to anticipate electrical distribution system substation loading during summer peak demand periods. Advanced analytics software was used to develop a dashboard for all of the utility’s substations across their entire service area, allowing quantification of how much time each of the substations spent at high loads close to or above manufacturers’ rating for key assets such as transformers and switchgear. This dashboard is now used for compliance reporting and maintenance prioritization, as well as for determining where new battery storage assets should be installed to provide maximum benefit. A further use is shifting load from one substation to another by comparing assets across the organization.

Another utility was experiencing a phase voltage anomaly on one of their generators. As with many difficult problems, this one was intermittent and therefore hard to spot due to low rates of data collection and other issues. Advanced analytics software was used to plot the variation of data away from the mean, revealing the problem to be an improperly installed fuse holder. Due to success with resolving this longstanding issue, the utility quickly scaled the solution to cover multiple assets fleetwide.

Any generation asset trips are a serious matter, and one facility was experiencing multiple incidents with no success identifying the root cause. The plant looked at extracting data from a historian and analyzing it with spreadsheets, but the amount of data and complexity of the required analysis made this an untenable solution.

Instead, advanced analytics software was applied to overlay transient events, add or remove signals as needed for troubleshooting, highlight events, and add markers at particular times. This investigation identified the cause of the trip as an instrument air problem, which precipitated an issue with a pneumatic steam injection control valve, the root cause of the trips.

The power industry has changed dramatically over the past few decades with the emergence of distributed energy resources, many of them small in scale with intermittent output. This is prompting a search for better solutions to increase reliability, effectively manage the grid, and improve maintenance by moving from reactive to proactive. Advanced analytics software is being applied to the time series data collected and stored by utilities to address these and other issues by quickly creating insights to improve operations across entire service areas and asset fleets.