The Dynamics of Interconnection Queues
Assumptions, risks, and costs
Jammed, but not static
Our articles focus a lot on generation interconnection, with some recent exploration of the nuances of interconnection project queues. Previous articles include a general overview of these queues and a close investigation of how interdependent queued projects can be. We’ve explained how, due to interconnection queue backlogs, it may take years for a project developer to receive initial (let alone final) system impact study results.
But, since the interconnection process is a years-long affair, does that mean projects “sit” for years with no changes? The classic engineering answer applies here: it depends. Project developers are allowed to change aspects of their queued project(s) without necessarily losing their queue position. Developers may also withdraw projects from a queue at any point, regardless of whether they have received study results or not. Despite the logjam, interconnection queues exhibit these and other kinds of potentially significant and impactful changes on a daily basis. In this article, we’ll use Southwest Power Pool’s queue as an example to highlight the extent to which changes in the queue might occur in a relevant time window.
You know what they say when you assume…
The foundation of an interconnection study is the set of power flow models on which the study is run. These models comprise snapshots of the grid with currently queued projects, as well as the prior queued projects and their mitigations, connected and energized. Utilities and ISOs make a number of assumptions about their interconnection queues when building these models, often denoted in business practices manuals and/or interconnection-specific user guides (many of these guides are available publicly). On the official study kickoff date, potentially long after the queue submission window has closed, engineers who develop these models take the current state of the queue as gospel. If any changes to an existing project, current or prior, in the queue occur during the study process, i.e., after the models are built, they are not captured in the study.
The assumptions made in the model development process can have a material impact on the results of a study. For example, prior queued generators are generally modeled in a way such that they are shown as injectors of power into the grid. If a prior queued generator withdraws from the queue and the model still has a nonzero injection of power from that generator, then the model is not consistent with the current state of the queue. Additionally, changes in a project’s point of interconnection (POI), injection capacity, or fuel type (e.g., wind, solar, storage, etc.) can also have a substantive impact on the final study results. When assumptions associated with these crucial aspects in modeling go stale, the validity of the study results becomes inherently questionable. If you’re a developer, this means the upgrade costs you are assigned according to the study results might not be what you expected based on your own analysis, or are substantially different than what would be determined if the study were to be redone at that point. This discrepancy is exactly what happens as projects move through the interconnection study process.
Example: SPP DISIS-2018-002/2019-001 cluster: Phase 1 study
Southwest Power Pool (SPP) is in the midst of an interconnection study backlog mitigation plan, which in part involves parallel processing of interconnection clusters (i.e., running multiple studies with different project submission window cutoffs simultaneously). As discussed in earlier Knowledge Center articles, a cluster study is usually split into multiple phases, with time between each phase allowing for developer decisionmaking and modeling refinements, among other processes. Prior to SPP’s backlog mitigation plan, Phase 1 of a given cluster could not begin until Phase 2 of the previous cluster’s study was complete, so decisions about prior queued projects were further along. With the new parallel processing approach, a new cluster’s Phase 1 study can begin upon completion of the previous cluster’s Phase 1 study, when decisions on those prior queued projects are far from final. Further, studies of projects in even earlier clusters may still be ongoing as well.
While intended to increase overall throughput and clear SPP’s backlog of queued projects, a consequence of the parallelized approach is that modeling assumptions can become stale very quickly. This is particularly true for assumptions regarding prior queued generation (i.e., projects submitted to an earlier cluster). The Phase 1 study models for a given cluster contain projects from earlier clusters that may still be under active study in a separate, parallel process. It is possible for these earlier projects to withdraw from the queue before the current cluster's Phase 1 study is even complete. The Phase 1 study results, then, would include the impacts of prior queued generation that would never be built.
Let’s take a closer look at the DISIS-2018-002/2019-001 cluster. The Phase 1 kickoff occurred on August 24, 2022. The cluster’s Phase 1 study results were posted on October 25, 2022. In that window, the following changes to twenty-three prior queued projects from several previous clusters occurred:
GEN-2017-187: POI Updated
GEN-2018-012: POI Updated
GEN-2018-013: Service Type Updated
GEN-2018-014: Service Type Updated
Because all of these changes occurred after the model freeze date, they were not accounted for in the DISIS-2018-002/2019-001 Phase 1 study. This is not a criticism of SPP’s modeling approach, as no matter what, at some point the models have to be built and assumptions must be made. It is also important to note that in the cluser’s Phase 2 study, all of these changes (and more) will of course be incorporated – studies always kick off with the most recent information available. But, developers with projects in the cluster still had to make business decisions based on Phase 1 results that were actually no longer consistent with SPP’s interconnection queue. And when the Phase 2 study is completed, the network upgrade costs may end up being substantially different than those found in Phase 1 due to changes in SPP’s queue. This is always a risk with cluster studies since even current queued projects can withdraw between Phases 1 and 2, but it is exacerbated when assumptions on prior queued projects can change dramatically between phases. All of this underscores the importance of speeding up the study process: the faster studies get done, the less stale the assumptions are in the models.
It is worth noting this situation is not unique to SPP. The Midcontinent Independent System Operator (MISO), for example, has active projects dating back to 2018; PJM, the largest regional transmission organization, has active projects dating back to 2017. Anytime one of these older projects has a material change, later queued projects may be impacted.
A more comprehensive look
The changes listed above may seem like a lot, but it’s important to note that this amount of variability in a queue is fairly normal. The table below shows a count of changes in SPP’s queue in each month from April through December of 2022, categorized by the field changing. This data is taken across all clusters in SPP’s queue. On average, there are roughly 11 material changes in the queue per month.
ISOs publish their most up-to-date queues on their websites, but to “look back in time” for this data, we leveraged Pearl Street’s interconnection intelligence platform. If you’d like an account, feel free to reach out to us at firstname.lastname@example.org.
Despite the logjam, interconnection queues are not static and often change daily. The assumptions used when building interconnection study models, whether it’s an ISO running a Phase 1 study or a developer conducting a prospecting assessment for a future project, can go stale very quickly. Developers in particular must pay careful attention to the assumptions used to build these models and adapt their decision making based on changes in queues that may have occurred since those assumptions were made.
If you enjoyed this article and want to learn more about some of the things we’re working on at Pearl Street, reach out to us at email@example.com. Thanks for reading!