If Jane Austen were alive today and writing copy for the power grid industry, she might title her masterpieces “Sensors and Sensibility” and “Pride and Predictives.”
The predictive asset management movement is advancing due to the meteoric rise of sensor technology in our lives. Neither your car nor your home cooling unit runs as well without them operating at optimum levels.
The power grid certainly cannot escape the sway of the sensor. Predictive asset management dominates every utility’s thinking and planning to some degree. Outage management, maintenance schedules and transformer performance all are increasingly beholden to the rise of the sensors.
And so we queried some industry leaders on why predictive asset management is such a big thing and so close to grid world domination.
“The benefits of predictive analytics include material savings and the avoidance of lost opportunity costs, which can be in the millions,” Kim Custeau, director of asset management for Schneider Electric Software, said in response to emailed questions from ELP Executive Digest. “Companies such as Duke Energy, EDF Group, Southern Co. and Exelon have all documented these types of savings, initially from power generation equipment and then from transformers, transmission and distribution assets.”
Schneider has a “maintenance maturity pyramid” (shown below) that shows the relationship between certain upkeep strategies and outcomes. The pyramid indicates that the Internet of Things (IoT) offers opportunities to use online monitoring for early warning detection.
“Investment in smart sensors will continue to grow but the gains will only be fully realized if the data is centrally managed, analyzed and provided in the right context to ensure that corrective action plans are properly developed and executed,” Custeau said.
Mike Madrazo, vice president of analytics product management at Silver Spring Networks, noted that meters with edge intelligence built into the communication card can provide many forms of predictive management and real-time monitoring. During this era of distributed generation, such devices can mitigate voltage issues due to excess solar, offer options for demand response programs and deal with on-site outage dangers.
“One major area of major area of focus is vegetation management, where very short voltage deviations and power interruptions can be clustered via local polling to identify vegetation contacts before they cause an outage,” Madrazo replied to ELP.
He sees the greatest need for predictives at the edge of the distribution network. The utilities are going a good job of networking and monitoring centralized assets, Madrazo added.
“However, at the edge either visibility does not exist or external inputs such as solar are impacting the stability of the grid and the performance of local assets,” he said. “These inputs cannot be managed through traditional asset models and therefore require the granular data available from the edge devices such as meters, EVSE (electric vehicle supply equipment) and solar inverters.”
Predictives are not for everyone if they are not in it for the long haul, Madrazo warned. They also may deploy successfully in one situation but not fit for all utilities and regions in the same way.
“Predictive analytics is part art and part science, requiring a great amount of time and resources to achieve the significant rewards that are available,” he said.
From Schneider Electric’s Custeau’s point of view, utilities are “trailing” in taking advantage of the online smart sensor data and incorporating that into reliability programs. They are missing out on thousands to millions of dollars in savings, she said, as well as double-digit percentage benefits in cost reduction and asset utilization.
“Predictive analytics enable engineers to spend less time sifting through raw data and more time improving the reliability assets,” Custeau said.