A report published on Energy Central in January 2019 described how two ISO’s (ISO New England and the New York ISO) were factoring in Behind-the-Meter PV (BTM PV) supply into their Load Forecasts. Under normal conditions the loss of a generating unit only effects the amount of energy that is available to meet demand. But BTM PV is different from other generating resources in that the loss of this supply resource has a direct impact on demand. Likewise, the availability of this resource also has a direct impact on demand. This yin-yang effect on supply/demand is unique to BTM PV supply resources and is explained further in the report cited previously.
A recent article published by Jon Black, Load Forecasting Manager at ISO New England (ISO-NE), describes the changes that ISO New England is making to it’s load forecasting methodology to address this BTM yin-yang effect. As Jon points out in his article “legacy modeling approaches that do not explicitly factor for BTM PV will generally exhibit deteriorating forecast performance. To help mitigate this tendency, ISO-NE has now built data processes and related tools to enable the explicit accounting of BTM PV in its load forecasts.”
BTM PV presents another challenge to load forecasters in ISO’s; the majority of these small resources are installed “under the ISO’s radar”, so it’s very difficult for an ISO to accurately estimate the real impact on supply/demand from these resources. The forecasters simply do not know where these BTM PV are operating on the grid or how much supply capacity they provide, nor do they know the amount of demand they may be masking. As Jon stated, the ISO needed “ground truth” about these resources. To overcome this problem ISO New England engaged an outside vendor to provide “data from approximately 10,000 reference systems, binned temporally in 5-minute increments and geographically at the town level.”
This BTM PV data was then incorporated into ISO-NE’s load forecasting models. Jon wraps up the article with a nod to industry: “The newly developed data intelligence and tools will better position ISO-NE to forecast the region’s rapidly evolving load profiles. In addition to improved net load forecasting performance, reconstituted load profiles also provide critical information for forecasters, since they reveal the amount of load that can suddenly become “unmasked” during periods of heavy rain or snow cover. As the industry accumulates greater experience load forecasting with larger BTM PV penetrations and other emerging technologies at the customer end of the power system, we all stand to benefit from learning from one another to improve overall methodologies and best practices.”
Well said Jon. Perhaps other ISO’s will consider ISO-NE’s approach to incorporate BTM PV into their load forecast processes.