2.
Provide your organization’s comments on the Requirement Calculation (including load error correction methodology):
MRP offers comments on two topics: (1) the CAISO scaling the IEPR load forecast and (2) the difference between the net load data available on the CAISO web site and the data the CAISO uses to determine the flexible capacity requirements.
IEPR Forecast Load Scaling
As MRP understands, after the CAISO compared the load ramps from the CEC’s IEPR load forecast with the actual load ramps from 2021 and 2022, the CAISO found that the CEC’s forecasted load ramp was 24.3% higher than actuals. Therefore, the CAISO adjusted the CEC’s 2024 load forecast by multiplying that forecast by an error correction of 75.7%.[1] The CAISO applied this error correction across all intervals in the year, but only for the purpose of determining the net load ramp – the CAISO did not apply the 75.7% correction factor for any other purpose.[2]
MRP has several concerns with this correction.
First, a blanket application of the load ramp error correction across all twelve months does not result in a more precise forecast of demand; it simply brings the load ramps determined from that forecast into alignment with recently observed load ramps. If the IEPR load forecast requires scaling, a more precise method would be to apply an error correction value that is specifically calculated for the specific month. By applying a single error correction value, the resulting forecast monthly load ramps may be lower in some months while being higher in other months when compared to actual 2024 load ramps. Effectively if the forecast is lower than actuals, that means the ISO has over corrected and under forecasted the load ramp which would result in LSEs potentially under procuring Flexible RA capacity.
Second, this error correction methodology was not discussed during the February inputs and methodology meeting. Slide 11 of the CAISO’s February 76, 2023 presentation describes the CAISO scaling the hourly CEC load forecast to produce one-minute load data from which to determine the load ramps but does not describe the CAISO scaling the forecast to achieve load ramp alignment with past year’s actual data. MRP believes the ISO should be required to discuss such significant changes to its methodology during its February meeting, rather than during the draft results study.
The departure from the process laid out in the Tariff and BPM is cause for concern as stakeholders were not made aware of new scaling methodology during the February 7, 2023 meeting. The flexible capacity assessment methodology, which is set forth in Tariff Section 40.10.1.3, does not contemplate the load scaling the CAISO described in the presentation of the draft results. And while Section 40.10.1.4 describes a forecast adjustment, that adjustment relates specifically to changes in the amount of operating reserve. And while the equation for setting the Flexible Capacity Requirements expressly includes an error term (epsilon), the CAISO expressly noted on the April 26 call that this error term would remain at zero (o). From this, it appears the CAISO did not closely follow the process outlined in the CAISO Tarff for conducting its Flexible Capacity Assessment.
Operational Net Load Ramp and ISO Reported Net Load Ramp
The CAISO offered in response to a question MRP asked during the April 26 webinar about why the load ramps determined from the IEPR forecast were so different than the actual ramps that the CAISO thought that the modeling of behind-the-meter solar may be responsible for the significant differences, but that market participants should participate in the Demand Analysis Working Group (“DAWG”) meetings if they wanted to understand the differences. MRP appreciates the suggestion to get involved in the DAWG but is concerned that the CAISO could not articulate the fundamental reason(s) why the IEPR forecast data yielded values that are so different than the actual net load ramps. Moreover, if the CAISO is committed to using the IEPR load forecast for its study purposes, it is difficult to understand why, if the CAISO scaled that forecast for one purpose (determining the FCR), the scaled forecast is not suited for all other purposes.
[1] Draft Assessment at page 10.
[2] Draft Assessment at page 11.
5.
Provide any additional questions or comments on the Apr 26, 2023 Flex RA meeting or discussion:
On the April 26 Call, MRP noted the significant difference between the monthly maximum three-hour net load ramps as indicated by the five-minute data on the CAISO’s web site (at http://www.caiso.com/TodaysOutlook/Pages/default.aspx) and the actual net load ramps as presented by the CAISO (e.g., Figure 3 of the Draft Assessment). The CAISO explained that the difference is due to the inclusion of dynamically transferred resources. It is unclear why such impacts are not included as part of the CAISO’s operational assessments being used for the Flexible Capacity Assessment. If the actual net load ramps are significantly different from those that are shown in the data that is posted to the CAISO website and public materials, then the CAISO should correct the public data.
Finally, MRP notes that the FCR calculations hinge on two sets of inputs, namely, the IEPR load forecast and the CAISO’s forecast of wind and solar production. While the CAISO has compared the ramps from the CEC IEPR load forecast against actuals, the CAISO does not compare its renewables generation forecast against actual renewable production. MRP respectfully encourages CAISO to assess how forecast wind and solar generation values compare to actual values for those inputs and report on that assessment in its future flexible capacity assessments. This will help market participants better understand the variability and uncertainty associated with each of the input variables.
