• Renewables
  • Curtailment
  • AI
  • Grid Infrastructure
  • Distributed Energy Resources Management
  • Distributed Energy Resources

Using AI to Fix the Renewables Curtailment Issue

Bill Hughes
Aug 04, 2020

Guidehouse Insights

This blog was coauthored by Michael Kelly.

Many consider large-scale, grid-connected batteries to be the solution to storing excess energy production from renewable generation created during daylight hours for use later in the evening. The common sentiment is that curtailment of renewable power is unfortunate and the grid should use as much green power as possible. A recent Guidehouse Insights article explored this practice and found that solar and wind power generators in California during 2020 were curtailing almost 1% of the total power used.

Exploring the Renewables Curtailment Issue

One might expect that more grid-connected battery capacity should directly result in storing as much green power as possible. In this scenario, the only reason to curtail solar power is if the batteries are already at full capacity. However, data from the California Independent System Operator (CAISO) shows that more than one-third of the power curtailed from solar generation sources occurs simultaneously with grid-connected batteries adding power to the grid. Not only do the batteries have additional storage capacity, but their use also increases the amount of curtailment. The following chart compares the amount and direction of power from grid-connected battery flow compared to the amount of curtailment during a summer day in California.

Grid Battery Power Flow during Curtailment: June 26, 2020

Grid Battery Power Flow during Curtailment: June 26, 2020

(Source: Guidehouse Insights and California Independent System Operator)

This chart shows that the batteries are charging with renewables power in the morning. After 2 p.m., however, the grid-connected batteries are adding power to the grid at the same time renewable power is being curtailed. The regrettable result is that the share of renewables is not as high as it could be.

This situation is not unusual. The chart below shows curtailment occurring at the same time as grid-connected batteries are adding power to the grid over a 2-week period. More than one-third of the curtailment took place when batteries were also adding power to the grid in California.

Percent of Curtailment Offset by Grid-Connected Batteries in California

Percent of Curtailment Offset by Grid-Connected Batteries in California

(Source: Guidehouse Insights and California Independent System Operator)

California is a leader in its use of renewable energy and in the amount of grid-connected storage. Information on the direction of power flow to or from batteries in 15-minute increments is viewable on the CAISO website. One would expect the line to be negative in the day and positive after dark. If this situation can happen in California, it is possible that similar situations take place elsewhere.

The Solution Involves AI

The solution to this situation involves greater use of AI to better predict grid demand and the amount of power that will be generated daily by the renewable generation sources that are in operation. Load forecasting isn’t new for utilities, but the proliferation of sensor technologies (e.g., smart meters, smart inverters) and advancements in AI and machine learning are enabling new capabilities across load and solar forecasting platforms:

  • Use cases have evolved from simple load forecasting to more dynamic applications around revenue, asset management, and locational pricing.
  • Input parameters have expanded to encompass a growing number of datasets, such as cloud cover, wind forecasts, and EV penetration.
  • Vendors are integrating nontraditional technologies into end use patterns.
  • Ongoing innovation is occurring around forecast algorithms and simulating the production and behavior of dynamic profiles for new technologies.

These emerging, dynamic analytics platforms can mitigate some of these curtailment concerns while avoiding the high cost and complexity of larger distributed energy resources management system projects. For more information on how utilities are addressing distributed energy resources management, see the Guidehouse Insights report, DER Management Technologies.