In the November 6 meeting of the Resource Adequacy Subcommittee (RASC), MISO updated stakeholders on potential strategies to model storage dispatch in the Loss of Load Expectation (LOLE) model and the impact of storage generation on the Direct Loss of Load (DLOL) based method. MISO would like stakeholder feedback on the anlaysis and methods shared with the RASC.
Comments are due by December 16.
Battery Energy Storage should have zero winter capacity accreditation. See "Pitfalls of Battery Energy Storage" paper.
The Environmental Sector appreciates MISO’s evaluation of different strategies to represent storage dispatch in the LOLE model and the opportunity to provide feedback on the November 6th RASC presentation comparing possible approaches and their impacts.
The way storage resources are dispatched in the LOLE model, a function of the operational rules in the model, can significantly affect their capacity accreditation under the DLOL methodology recently approved by FERC. Due to this, we encourage MISO to model storage resources using the method that maximizes their potential contribution to maintaining the reliability of the grid, to ensure that storage is being accredited based on its true capacity value. In addition, understanding how to optimally utilize storage during risk periods or emergency events could be useful in informing improvements to market rules (including but not limited to opportunity cost allowances in offer prices), operational strategies, and/or software.
While existing operational decisions may not always maximize the contribution of storage resources to system reliability, we believe the solution to this discrepancy is to improve those operational decisions and reflect such improvements in modeling, and not simply to assume the continuation of existing market inefficiencies. For instance, energy storage resources can change discharge patterns to maintain capacity usefulness once storage supply reaches saturation levels; thus, MISO should assume energy storage dispatch in operations can be modified as storage penetration increases and reflect that in the LOLE model accordingly.
MISO’s current method for modeling storage can result in a phenomenon where storage resources “eat their own lunch”. Essentially, storage resources may be able to effectively meet capacity needs at the beginning of a potential Loss of Load Event (LOLEv), but as their state of charge is depleted, they shift risk hours (and thus the key hours in the accreditation determination) forward to periods when storage resources are no longer available. While MISO’s existing storage modeling methodology may be a reasonable representation of current operator actions, it does not mean that it is an accurate representation of the true potential capacity value of storage resources. In this example, despite storage resources' contribution to reducing the duration or magnitude of a LOLEv, under the current DLOL methodology, storage resources are not given credit for their contribution during hours where they completely mitigate LOL.
MISO presented the “blended” method at the RASC as a way to work around this challenge by forcing 1MW of LOL during all hours of an event, classifying those hours as LOL hours and thus recognizing storage’s contribution to mitigating them. We acknowledge that the “blended” approach is an improvement over the “early” dispatch approach. However, this method does not actually provide a meaningful workaround: it will still reduce the capacity accreditation of storage due to the weighting methodology in MISO’s recently approved DLOL accreditation procedure. Using the “blended” approach in the DLOL methodology, the hours with a 1MW capacity shortfall, when the bulk of storage resources are modeled to be dispatched, will be significantly less valuable than hours with capacity shortfalls of greater magnitude and will therefore be significantly outweighed by later hours in the accreditation calculation. As a result, storage contribution during the forced 1MW LOL hours will continue to be undervalued relative to other resource types. By contrast, the “even loss” modeling methodology presented at the RASC deploys storage more strategically to optimize the reduction of LOLEv depth, effectively using storage at its highest capacity value. We request that MISO provide additional information about the relative weighting of the hours included in calculating the accreditation values for storage under both the “blended” and “even loss” approaches.
Overall, the Environmental Sector believes that MISO should pursue an approach that maximizes the contribution of storage resources to maintaining reliability during the entire duration of an event, which in this case is the “even loss” modeling method. Under the “even loss” method, storage resources are able to minimize the depth of loss of load events rather than the length of such events. Optimizing storage to reduce the depth of events may allow MISO and its utility members to be more targeted in load shed to limit harmful impacts on essential services. It may also improve the effectiveness of combining targeted DR curtailment and storage energy together to mitigate LOL risk.
Additional Feedback/Future Analysis:
Evaluating storage capacity additions without considering renewable energy growth does not yield meaningful insights. MISO conducted simulations of two scenarios when evaluating energy storage: one with only storage additions and the other with both storage and significant wind and solar resource additions (described on slide 9). However, substantial evidence indicates that the value of storage, particularly short-duration storage, is closely tied to renewable intermittency; this mirrors what is shown in MISO's results on slides 12 and 13. In MISO, we expect future energy storage resources will be developed in parallel with further development of wind and solar generation. To accurately reflect expected future system conditions, MISO should always include a reasonable level of additional renewables when evaluating the accreditation of storage resources.
MISO should compare the impact of storage dispatch on the depth of LOLEv. We request that MISO also compare the depth of the loss of load events in the model using these different approaches to better understand the value of storage under each approach.
Additional analysis should include a fair evaluation of long-duration energy storage resources. We also encourage MISO to evaluate the capacity contribution of longer duration storage, including both 8 hour storage and long- duration storage (LDES). While 4 hour storage may be reasonable for the prompt year and the near future, we expect longer duration storage to become an increasingly competitive resource on MISO’s system. The Department of Energy projects that several kinds of LDES technologies will be deployed commercially starting in 2028. Demonstration projects have already broken ground across the United States; for instance, in the MISO footprint, Xcel Minnesota has received regulatory approval and broken ground on a multi-day (100 hour) energy storage system, using FORM Energy’s iron air battery technology, which is expected to come online as early as 2025.
Protocols should be updated to allow operators to strategically use the various modes of operation energy storage resources offer. Storage resources, including batteries, are capable of multiple modes of operation designed to optimize different functions and meet different grid needs. MISO operators should be trained to use these different modes according to protocols defined in MISO’s Reliability Operating Procedures. MISO should consider developing different procedures for the strategic use of storage during both “Normal” and “Emergency” Operating Conditions.
Normal Operating Conditions: In normal operating conditions, storage resources should be used in conjunction with renewable energy generation, as configured in the early dispatch scenario. Under these conditions, storage resources add value by reducing renewable variability in sudden low-wind and low-solar conditions. Additionally, storage resources operated in this way can effectively reduce the need to ramp-up thermal resources in the afternoon (as shown by the IMM’s duck curve).
Emergency Operating Conditions: In emergency operating conditions, such as when operators are faced with conditions of grid instability (winter storms, hurricanes, etc.), storage resources should be dispatched under the even loss strategy to best serve reliability and reduce the need for load shedding. Without protocols that designate how storage resources can be used during emergency conditions, operators cannot deploy the resource to its maximal potential.
RASC: LOLE Modeling Enhancements - Storage Modeling (20241106)
Invenergy appreciates the opportunity to provide our feedback to MISO’s strategies to model storage dispatch in the Loss of Load Expectation (LOLE) model and the impact of storage generation on the Direct Loss of Load (DLOL) based method.
Invenergy supports MISO’s “blended” option which includes both the preservation of the current dispatch method (which recognizes and deploys the ‘as early as needed’ approach which mimics real time market operation) but also recognizes the importance of the reliability contribution of storage prior to a full loss of load event by spreading the event to all hours and/or hours before full loss of load. This results in all the hours in the event to be part of the Critical Hours pool, thus counting storage’s availability and reliability towards accreditation.
Invenergy recommends MISO continue to collaborate with Astrapé to add storage availability to the reporting of margin and communicate with stakeholders for them to better understand this nuance (i.e., is this a fundamental gap or simply a reporting feature that needs to be resolved).
Invenergy would also like to recognize and thank MISO for fulfilling their commitment as noted in their DLOL FERC filing to pursue this process enhancement.
The Entergy Operating Companies ("EOCs")[1] appreciate the opportunity to provide feedback on MISO’s storage modeling proposal.
The EOCs support MISO’s efforts to improve how storage resources are modeled and accredited under DLOL. Further, the EOCs are generally supportive of the “blended” concept proposed by MISO. However, the EOCs are concerned that MISO’s proposal is assigning a minimal weighting to the critical hours that occur during the storage dispatch period which results in lower storage accreditation ratings. This is caused by MISO’s design decision to specify only 1 MW of load shed to the storage dispatch hours. The EOCs believe that MISO should explore other design alternatives that would assign a higher and more reasonable weighting for the critical hours that occur during the storage dispatch period.
Further the EOCs request that MISO clarify whether the 1 MW of load shed assigned to the storage dispatch hours will influence the accreditation for all resource types or just storage resources.
[1] The Entergy Operating Companies are Entergy Arkansas, LLC, Entergy Louisiana, LLC, Entergy Mississippi, LLC, Entergy New Orleans, LLC, and Entergy Texas, Inc.
Clean Grid Alliance (CGA) appreciates this opportunity to provide feedback.
CGA supports the Environmental Sector's feedback.
MISO should continue to develop the “even loss” approach for energy storage class accreditation. The “blended” method would unduly harm the storage resource class (sometimes to the undue advantage of other resource classes). In particular, that's because the “first come, first served” element of the method forces the modeled depletion of storage capacity before other weighted risky hours that might count much more for accreditation. Storage resources are able to make themselves more useful than the blended method would allow, thus MISO should pursue the "even-loss" method.
Besides, with appropriate Energy and Operating Reserve markets rules allowing storage to incorporate reasonable opportunity costs into offer prices, the even-loss method would be the most reasonable choice. If the Independent Market Monitor (IMM) allows storage resources that prioritize being available for discharge during the most extremely risky hours (e.g., “Max Gen” emergency conditions) to offer energy at very high prices in other hours, then that storage can more easily be available across more risky hours. Further, energy storage can offer energy on an emergency-only basis. So, contrary to MISO’s claim on slide 7 of its presentation, operators would be more likely “to defer [some] storage in the presence of an event,” because storage could offer in ways (i.e., at prices or as emergency-only) that would prevent them from discharging in lower risk hours and allow them to discharge only during higher risk hours. With such market rules in place or clarified, MISO then could assume market behavior according to those rules in DLOL modeling, which would allow DLOL modeling of a more realistic storage resource class that would be available to serve more of the modeled risky hours.
Regardless of the accreditation method chosen for storage, MISO and the IMM should clarify how the IMM will allow storage to define opportunity costs for Energy and Operating Reserves offer prices. MISO also should explain the interaction of those market rules and DLOL modeling assumptions.
Lastly, MISO should explain how Long Duration Energy Storage (LDES) with output duration (materially) over 4 hours would fare under these rules (making proper assumptions about other resource classes). Adding more LDES to the storage resource class will make more reasonable the recommended DLOL modeling that allows storage to be available to serve more of the modeled hours with highest risk.
I'm happy to discuss.
David Sapper
dsapper@cleangridalliance.org
DTE appreciates the opportunity to provide feedback on MISO’s analysis of various ESR modeling methodologies within the LOLE and their impacts on DLOL class-level accreditation results. DTE believes that ESR modeling should closely reflect real-world operations, as deviations could lead to improper accreditation.
With this guiding principle, DTE supports the ‘Early Dispatch’ method as the best option among the three proposed for modeling ESRs. If MISO opts for the ‘Blended’ or ‘Even Loss’ approach, it will diverge from actual ESR dispatch practices in the MISO market and could be perceived as preferential treatment for battery resources. Using the ‘Blended’ or ‘Even Loss’ approach could also result in high class-level accreditation for 4-hour ESRs, despite their inability to provide the accredited capacity value in real operations. However, if MISO modeled ESRs how they are dispatched in actual operation, the potentially lower class-level accreditation results could highlight the need for longer-duration storage otherwise missed if other approaches are used.
Instead of altering the dispatch methodology in the LOLE to accredit ESRs at a higher percentage, MISO should explore better ways to incorporate actual operation of all resource types in the LOLE, including adding intra-day charging and discharging for ESRs. This would result in the LOLE model better reflecting real-world operations while ensuring ESRs are more accurately accredited and are not negatively impacted by operational limits imposed in the LOLE model, such as the inability to properly dispatch ESRs during intra-day dual peak Loss of Load events.
MN8 Energy appreciates the opportunity to comment on MISO's storage modeling enhancements presented during the November 6th Resource Adequacy Subcommittee.
MISO's current proposal would weight hours based on a measure of supply margin that would assign greater weight to hours with higher unserved energy than those with lower unserved energy, after already weighting for probability of occurrence. Any MW delivered during a loss of load event, or to mitigate a loss of load event, has an equal impact on loss of load risk and unserved energy, all else being equal, and hence, should be valued equally. MISO should rework its approach to DLOL to correct this discrepancy between actual reliability impact and accreditation.
The OMS Resources Work Group (OMS RWG) provides these comments to MISO in response to its potential strategies to model storage dispatch in the LOLE model and its proposal to evaluate the impact of storage on its DLOL accreditation methodology. This feedback is from an OMS work group and does not represent a position of the OMS Board of Directors.
The OMS RWG appreciates MISO’s efforts to review how storage resources are dispatched within the LOLE model and consider alternative approaches. This topic received significant attention during the MISO stakeholder discussions around the Direct Loss of Load accreditation methodology, as the approach used to dispatch storage resources in the LOLE model have a significant impact on which hours end up being designated as loss of load hours, and thus, which hours have the greatest impact on resource accreditation.
The OMS RWG agrees with MISO’s decision to evaluate three different approaches on how to model storage dispatch. The “Early” dispatch – which is how storage is currently modeled – and “Even Loss” dispatch both warrant further analysis and discussion, and the “Blended” dispatch MISO developed could capture the advantages of both approaches. The OMS RWG also agrees with the multiple stakeholders that made comments during the November 6th RASC stating that the chosen storage dispatch methodology used in the LOLE modeling need not align with how operators deploy storage. Given the weighting of hours in the Direct Loss of Load accreditation methodology, the storage dispatch in the LOLE model needs to be done in a way that most accurately reflects the reliability value that storage resources provide for the system.
MISO RASC: LOLE Modeling Enhancements - Storage Modeling (20241106)
NextEra Energy Resources Feedback
Date: December 16, 2024
Please see the feedback provided via email by NextEra Energy Resources due to the inclusion of graphics that are not supported by the Feedback Tool. Thank you.