The Next Best-Practice in Reliability

Trekker Edge Computing Sensors on Garbage Trucks Are Monitoring the Grid & Providing Machine Learning Analytics for Improved Reliability

There’s something very interesting happening in the utility industry.  Garbage trucks are being used to improve the reliability and the resilience of the overhead electrical grid. Patented, multifunction edge computing sensors mounted to garbage trucks  provide weekly monitoring of overhead grid health. The sensors identify deteriorating equipment emitting partial discharge pre-fail signatures which allows utilities to replace equipment before an outage occurs. Findings are mapped, time-coded, and become powerful conditions-based data points for driving machine learning algorithms that reveal how the grid responds to storms, humidity, heat, wind, snow, ice, load, and other factors. The AI and machine learning analytics can provide powerful insights for pro-active predictive maintenance, storm response resourcing, financial planning, and drive other operational efficiencies. For more information, contact John Lauletta, CEO, Exacter, Inc. [email protected]

As the trucks drive their weekly routes, Trekker sensors identify and map GPS coordinates of deteriorating equipment on the grid and associated environmental conditions. Knowing where incipient problems are located allows the utility to make repairs before a power outage occurs.

What’s even more interesting is the data gathered provides a powerful source for machine learning models that will evaluate how the grid responds to load, weather, air contamination, and other factors that lead to outages and pole fires.  The analytics will provide utilities unique insights for financial forecasting and planning of intelligence-based predictive maintenance.

For example, by reviewing the number of deteriorating components before and then after a storm, the utility has an accurate assessment of actual impact of the storm. As more events occur, the quality of the repetitive route data to evaluate grid conditions drives powerful machine learning algorithms to analyze the delta.

This enables the utility to predict damage and prepare manpower resources to address the storm BEFORE it happens.  Forecast capability like this is a huge benefit to utility operations striving for improved grid reliability and resilience. Because the fleet of garbage trucks traverses the same routes every week across each city, the body of data gathered becomes a valuable resource for proactive management and maintenance of the grid.

What differentiates the Exacter Garbage Turck project from other predictive analytic efforts is the quality and criticality of the data being gathered by the Trekker sensors? Trekker identifies partial discharge in the form of arcing, current leakage, or tracking taking place on the overhead equipment.  According to industry experts, partial discharge is the most indicative sign of equipment deterioration that leads failure.

In other words, the partial discharge locations discovered by Trekker represent the most eminent points of risk, the points most likely for an upcoming outage on the grid. Remove these points of risk . . . And you remove the opportunities for failure, which makes the grid inherently more reliable.

Advanced Analytics from Strategic Partner SAS
Exacter has engaged in a strategic partnership with SAS to develop powerful AI and algorithmic analytics to provide utilities with unparalleled insights into grid behavior that will allow utilities to reduce outages, minimize truck rolls, improve worker safety and gain deep understaning about the health and resilience of the grid.  Their product offering, Grid Guardian AI is already being used by a number of utilities around the country.  Click Here for More information.

Installation of Trekker sensors on the trucks is easy.  The sensors require no external wiring or operator intervention.  They are simply mounted on any fleet vehicle, and as the truck moves, they automatically collect data.  There is no operator intervention.

Imagine the potential of 52 weeks of data and the ability to correlate deteriorating equipment with a host of environmental conditions. This is truly a first-of-its-kind conditions-based health assessment of an entire US cities   It’s what we believe will be the first of many.