Leveraging machine learning to determine projects most likely to join the power grid
Introduction to ISO Project Success
Most projects that enter the interconnection queues never get built. The queues are growing increasingly crowded, and backlogs continue to persist across multiple ISOs. Renewable project developers and investors are wondering:
- How can I increase my chance of success in the queue?
- What is the probability of projects that I’m competing with?
- Which projects or developers are most likely to join the power grid?
Did you know? In 2024, 23 GW of solar, wind and battery storage projects were brought online by Enverus customers.
How ISO Project Probability Is Calculated
Enverus Project Tracking Analytics provides probability projections for more than 6,500 pre-construction solar, wind and energy storage projects across CAISO, ERCOT, MISO, NE-ISO and PJM. Our first models, released in early 2024, provide a score to projects and developer portfolios based on:
- Location
- Project Type
- Companies Involved
- Capacity
- Days in Queue
- Substation Planned and Operating Capacity
- Study Duration
- Among other inputs
Models are now enhanced for ISO project probability which include additional factors:
Available Transfer Capability (ATC) – the amount of additional power available to be transferred between two points on a grid after accounting for existing demand without inhibiting grid stability. If a project has sufficient ATC, then other factors can weigh more into the project probability. If a project does not have sufficient ATC, then the project probability is likely to be lower.
Power Purchase Agreements (PPA) – an agreement that an entity will purchase the power being produced from a certain energy project. The existence of a PPA for a given project in the queue indicates there are vested interests in the project and increases the likelihood of construction.
Current View of ISO Project Probabilities
Figure 1: Enverus Project Tracking Analytics
ERCOT projects have the highest average probabilities at 26.26%. Of the project types in ERCOT, onshore wind is seeing the most success with an average probability of 37.63%, although it’s worth noting that the number of wind projects in development has been decreasing. Solar and battery storage average probabilities are not far behind with average probabilities of 27.85% and 24.25%, respectively.
PJM projects have the lowest average probabilities at 2.26%. This is no surprise as PJM remains under significant strain with many renewable projects stalled. More than 260 GW of renewable and storage requests are waiting to interconnect in PJM.
Figure 2: Enverus Project Tracking Analytics
Across all ISOs, NextEra Energy and ENGIE stand out as companies with a lot of capacity likely to come online. Currently, NextEra Energy has more than 60 planned projects across MISO, CAISO, ERCOT and PJM with an average probability of 55.56%. ENGIE has more than 40 planned projects across MISO, ERCOT and ISONE with an average probability of 52.64%. High probabilities from these companies can be attributed to secured PPAs, performance in ISO studies and skillful project planning/siting.
Conclusion
The inclusion of ATC and PPA data in the Enverus Interconnection Queue Probability Models has resulted in a 6-8% average increase in accuracy from the previous models, which were already in the 70-80% accuracy range, depending on the ISO. This increased precision enables our customers to make more informed decisions when considering nearby projects in the interconnection queue. Enverus will continue to gather feedback and research other improvements for future iterations of the models to best serve our customers leading the energy transition.