Summary

On February 21, 2024 the maintainers conducted a webinar hosted by LF Energy to introduce the project to the broader community. The replay page is here.

By the Numbers

  • RSVPs: 209
  • Participants: 144
  • Questions asked: 14

The audience posed some great questions, thank you! We didn’t have time to get to most of them, so we’re answering them here. In some cases we’ve edited the original question for clarity, based on our reading of them.

Q & A

Does the TROLIE interface handle day/night ratings?

The TROLIE Specification currently has no particular means to differentiate ratings based on the ambient conditions that were used to determine the ratings.

When a Ratings Provider proposes ratings to TROLIE, they do so according to their own Facility Ratings Methodology which must, per the Order, take insolation into account. Typically, active insolation is assumed to occur at a facility between sunrise and sunset, but this remains a factor determined solely by the Ratings Provider, not the TROLIE Clearinghouse Provider (Reliability Coordinator), so the Specification has no concept of “day/night”.

There are multiple ways to interpret this question that could imply useful extensions to TROLIE:

  1. Is it important to track whether the rating provided is based on a day or night rating? Is this a common use case for interop?
  2. Currently, the exchange of temperature-to-rating tables is not in TROLIE scope. Should it be? Is this a common enough interop need?

Please submit further questions or propose TROLIE scope changes at https://github.com/trolie/spec/issues.

How does CIM relate to the TROLIE Specification?

The maintainers are familiar with CIM, and it has influenced the development of TROLIE Concepts. However, not all TROLIE Concepts had an acceptable analogue in CIM and the traditional toolchain and serialization conventions of CIM, e.g., RDF XML, do not align well with the typical technology stack of a modern REST API. In short, TROLIE borrows what is most useful from CIM’s semantics while specifying a very conventional web API using JSON media types.

Branches in our system model are nominated by a "From Bus" and a "To Bus"; they are not assigned a (synthetic) unique identifier. How will we be using the `resource-id` field to identify a branch when branches are not assigned a single identifier?

We created a new issue for this: #54. Please join the discussion there.

How are Monitoring Sets, Transmission Facilities and Segments defined and updated by Ratings Provider? What are the validation rules, and how are they coordinated with network model updates?

Exchanging network model data is not in scope for TROLIE. This is a complex problem that likely requires its own set of defined protocols, standards, discussions and vendor adoption, much like the Common Grid Model Exchange Standard (CGMES) used in the ENTSO-E.

TROLIE Specification is intentionally agnostic to how server and client implementations update their network models. The project will, however, develop a Ratings Obligation profile that describes the Monitoring Sets, Ratings Providers, Facilities, Segments, etc. that are assumed by the Conformance testing suite; see trolie/spec#35. Note that this Ratings Obligation profile is not part of the API Specification as such; there’s no “administrative” API surface to exchange this kind of reference data. Therefore, there is no prescribed validation of the same. Moreover, support for the Ratings Obligation profile is only necessary if a vendor implementing the TROLIE Specification wishes to participate in Conformance testing; it’s not strictly required to implement the Specification.

Please consult appropriate vendors, Reliability Coordinators or partners on how their specific software behaves. In addition, we have created issue trolie/spec#57 to document recommended validation rules and model coordination best-practices for TROLIE implementers.

Would TROLIE be a cloud-based solution?

TROLIE is agnostic to where servers are hosted. The specification will work whether clients or servers are hosted either on-premise or in a public cloud. Specific vendor products and reliability coordinator implementations will of course be more opinionated.

Does the specification allow for UTC timestamps?

Yes, TROLIE timestamps use the RFC 3339 format. This format is daylight savings-safe, as shown by example at https://trolie.energy/daylight-savings.

The example referenced above however shows timestamps in local time zones. The common way to specify UTC is to use a “Z” character instead of a UTC offset as a suffix, like in the following example, which evaluates to 7am EDT: 2025-11-01T11:00:00Z

How are the status/state of Real-Time and Forecast snapshots communicated via TROLIE interfaces? How can we know from the TROLIE interface that a new update for the snapshot values is available?

We’ve made an intentional decision to adopt certain design constraints, including the restriction of the TROLIE Specification to classic REST patterns and HTTP/1.1. Thus, we don’t specify WebSockets, Server Sent Events, or recommend HTTP long-polling, instead relying on the Conditional GET pattern mediated by rate limiting.

Can you provide an implementation timeline? Is there an environment for us to test?

The TROLIE Specification is currently a WIP, and we are working on defining milestones for 1.0. Among these milestones will be comprehensive examples. However, since the TROLIE project does not intend to develop a server implementation, we refer you to the appropriate Reliability Coordinator for more information about integration testing.

Our RTO will need to send annual day/night times so each member uses the exact time to define day/night; can TROLIE support that?

This is currently out of scope of TROLIE. As of now, TROLIE does not include anything related to weather data. Whether these timestamps represent weather data that should come from somewhere else, or they are truly key in orchestrating ratings exchange across providers in a common way is an interesting debate.

If this is desired in TROLIE, the need should probably be driven by the RTO by reaching out to one of the project contributors, their vendor, and/or submitting a proposal to https://github.com/trolie/spec/issues.

Does TROLIE handle the exchange of AARs between a Ratings Provider (TOP) using amps and an Clearinghouse Provider (RC) using MVA?

The TROLIE supports the exchange of various kinds of ratings, including apparent power, current, reactive power, etc., so it supports specifying quantities with the appropriate units, including amps and MVA. However, the specification does not require the RC to accept any particular kind of rating, so the Ratings Provider and the Clearinghouse Provider will need to pre-determine the appropriate units. The TROLIE Specification is designed to be flexible enough to accommodate most anticipated situations, such an RC only allowing for apparent power as inputs and sending apparent power as outputs, as well as an RC allowing for current and power factor as inputs. We have a number of tasks identified to properly document this feature.