7.2: Specific Areas of Application of Artificial Intelligence

Community to grid interactions

Over the years, utilities such as PG&E have deployed SmartMeter™ technology to residences and businesses. "SmartMeter™ and Meter-Connector provide communication between PG&E and the grid. This two-way communication allows us to quickly identify outages and resolve other service problems, typically without visiting your home or business."[2] Customers can view the information collected by PG&E providing an online history of energy usage and costs (hourly electric and daily gas and electric consumption) and compare consumption to prior time periods. This information can be useful to both the utility and to customers.

As more customers buy electric vehicles and charge them at home, it is important to do the charging in an economically efficient manner. Managing load shifting is important. It is important economically to do the charging when the rates are most favorable.

As customers deploy solar energy, it is useful to track energy flow.

Energy System Resiliency and Managing Disaster Risks

Various weather events can severely impact a utility's infrastructure. High winds can cause trees or tree branches to fall onto utility power lines. Utilities spend a great deal of effort trying to keep their power lines clear of tree branches by surveying potential exposures proactively using human resources on foot or using drones in the field and dispatching tree trimming equipment.

Utilities have reacted to threats to their infrastructure by embarking on projects to put their transmission lines underground ("undergrounding").

Utilities can use Artificial intelligence to predict equipment failure and schedule replacements before failure.

Utilities can use Artificial intelligence to make predictions on potential high wind events such as Diablo winds in Northern California and Santa Ana winds in Southern California. Models can be trained based on years of historical data. Utilities can selectively de-energize parts of their power infrastructure to mitigate risks.

In the event of a fire, the sooner it is detected and suppressed, the better as far as limiting damage to utility infrastructure and the property of homeowners and businesses.

Autonomous operations in energy plants

Automating work in energy plants can yield operational efficiencies.

Utility equipment maintenance predictions

Mainframe and network equipment vendors have for years used predictive algorithms to schedule equipment maintenance or part replacement. For example, by monitoring the speed of fans that are used to cool equipment, these vendors can anticipate when the fan might fail and schedule replacing it.

Utilities can do something similar for their equipment based on the rated duty cycle of each type of equipment. Utilities could collect operational performance information and use it to predict failures and schedule maintenance in advance. The operational data could be supplemented by machine vision, automatic diagnotic inspections, and the like.

Environmental impacts of AI Usage

It should be noted that even as utilities and other industries use Artificial Intelligence, AI consumes vast computing resources placed in very large data centers. These computing resources consume a great deal of power that can have an impact on the environment.

References

1. Slate, Daniel D., Parisot, Alexandre, Min, Liang, Panciatici, Patrick, Van Hentenryck; Adoption of Artificial Intelligence by Electric Utilities Energy Law Journal, Volume 45, No. 1, 2024., The Energy Bar Association, University of Tulsa College of Law.
2. Pacific Gas and Electric. SmartMeter™. Smartmeter.