MSC: Improve Accuracy of Resource- Submitted Forecasts for DIRs (MSC-2021-11) (20210805)

Xcel Energy appreciates the opportunity to provide feedback regarding the accuracy requirements of resource-submitted forecasts for DIRs. 

Forecasting Technology and Support of a 50/50 Forecast

Our forecasting methodology averages several weather models together and weights those with the best recent performance the highest.  That result is then converted from a wind speed forecast to a power generation forecast based on power curves.  The 50/50 forecast is the forecast we use, but other percentile forecasts are available. 

Forecast Performance Expectations 

There are some considerations to keep in mind when considering the accuracy of a RT wind forecast: 

• Icing – Some vendors do not incorporate an icing forecast so adjustments have to be made manually based on conditions and turbine specifications.  It is not difficult to make a one-time adjustment to the day-ahead forecast each morning, but  continuous near-term submittals for RT would need to incorporate an automated process that may not be available.   Icing adjustments are further complicated by the technology required to determine when icing is occurring/is sustained at the wind farm.   
• Outages can create a major difference in forecast accuracy.  MISO should consider the outage rules as they apply to DIR resources that may not be staffed 24/7 and the impacts of real-time outages to the forecast. 

WPPI provides the following feedback on improving the accuracy of resource-submitted forecasts for DIRs:

(1)    Please comment on the forecast technology you use and whether your technology supports a 50/50 forecast.

Not applicable. WPPI uses MISO’s forecast for its DIRs (for real-time operations).

(2)    How can MISO best characterize forecasting performance expectations to support reliable operations?

(3)    Please comment on the proposed Tariff language that is posted along with the presentation.

In response to questions (2) and (3), the proposed Tariff language seems to provide sufficient criteria to facilitate accurate resource-submitted forecasts for DIRs (Forecast Maximum Limit), “directly derived from”:

•  “a timely, accurate, and unbiased forecast”
• “ actual known physical capabilities”
•  “actual generation data”
• applicable weather information/models

In addition, the proposed Tariff language allows MISO to determine that a Market Participant’s Forecast Maximum Limits are persistently inaccurate based on a comparison with MISO’s calculated Forecast Maximum Limits and/or actual generation data.

It seems to WPPI that the proposed tariff language above suffices to facilitate accurate resource-submitted forecasts for DIRs and requiring the Forecast Maximum Limit reflect the “most likely forecast outcome” is not necessary, may not be helpful, and does not unambiguously identify a specific statistical measure.

Consumers Energy appreciates the opportunity to provide feedback regarding the Improve Accuracy of Resource- Submitted Forecasts for DIRs (MSC-2021-11) issue discussed during the August 5, 2021 MSC meeting.

Consumers supports MISO's proposed Tariff language posted along with the presentation to define standards intended to improve accuracy of resource-submitted forecasts for DIRs.

Clean Grid Alliance (CGA) Comments on MISO’s Proposed Intermittent Resource Forecasting Tariff Changes

At the August 5, 2021, Market Subcommittee Meeting, MISO proposed tariff revisions intended to help ensure MISO has accurate forecasts for Dispatchable Intermittent Resources (DIRs) as it manages the real-time market.  CGA appreciates and supports MISO’s goal of having accurate forecast data for DIRs, especially as the region moves towards higher penetrations of these resources.  We do, however, have some concerns about the lack of clarity in the proposed tariff language as detailed below. 

We encourage MISO to remove some vague language from its proposed changes and to provide language that clarifies how MISO will determine whether a Market Participant’s (MP’s) own forecast is not as accurate as required.  The language below from the Forecast Maximum Limit section is too subjective to be useful in evaluating the accuracy of a MP’s forecast and we urge MISO to remove it. Some potential alternative language is included.

“(D) the submitted Forecast Maximum Limit must reflect the most likely forecast outcome, and be directly derived from a timely, accurate, and unbiased forecast,be consistent with the requirements of this Tariff, the Business Practices Manuals, and in accordance with Good Utility Practice, with said forecast from which the limit is derived being directly derived from the actual known physical capabilities and characteristics of the Resource, actual generation data for the Resource, weather information specific to the physical location of the Resource, and weather prediction models for the physical location of the Resource, all relevant as of the time of submission.”

We also suggest that MISO develop specific tariff provisions or BPM language that is focused on defining what “persistently inaccurate” means and how MISO will evaluate the accuracy of a MP’s forecasts relative to MISO’s own forecast.  It is only when a MP’s forecast is less accurate than MISO’s forecast that MISO should take action to override the MP’s forecast.  We suggest the language clarify that:

• If the MP’s provided forecast is more accurate than the MISO forecast or within a certain small deviation band[1] from MISO’s forecast, the MP’s forecast will be used.
• If the MP’s provided forecast is consistently outside this acceptable deviation band, MISO will substitute its own forecast.
• The evaluation method that will be used to determine if the MP’s forecast is outside the deviation band.  This method should be described in either a Business Practice Manual or Tariff Section and referenced in the tariff section that describes the required accuracy.

CGA appreciates the opportunity to provide this feedback and welcomes any questions.

Respectfully submitted,

Natalie McIntire

Consultant for Clean Grid Alliance