This short dictionary defines some of the terms used by the GO Competition.
The Challenge 1 Final Dataset (C1FD) will be used for the officially scored round of the GO Competition Challenge 1. Conditions will be similar to each of the trial events with the new C1FD used for evaluation and scoring. A deadline for the submission of solution software will be established at least one month prior to the final event. Immediately following the deadline, the software from all competitors will be run and scored against the C1FD. The Challenge 1 Final Dataset score for each competitor submission will be displayed on the competition leader board.
Challenge 1 winners will be determined based on the C1FD dataset score subject to the winning criteria specified in the official Competition Rules document.
The Challenge 1 Original Dataset (C1OD) released at the start of Challenge 1 in order to allow competitors to develop solution methods. The dataset will consist of multiple power system network models each containing multiple scenarios. Competitors will be able to download the dataset in order to test algorithms within their own development environment. Competitors can also submit software to be scored against the C1OD dataset using the official competition platform.
The Challenge 1 Trial Dataset 1 (C1TD1) will be accessible, via submissions to the evaluation platform, approximately 6 months after the initiation of Challenge 1 as the first of two dry-run "trial" rounds for the GO Competition. This previously unreleased dataset is expected to be of similar complexity and scope as the P1OD, but will not be publicly released until after the conclusion of the trial events.
The Challenge 1 Trial Dataset 2 (C1TD2) will be accessible, via submissions to the evaluation platform, approximately 9 months after the initiation of Challenge 1 as the second of two dry-run "trial" rounds for the GO Competition. This previously unreleased dataset is expected to be of similar complexity and scope as the P1OD, but will not be publicly released until after the conclusion of the trial events.
A collection of power system network models and scenario data on those models. Challenge 1 will have four distinct datasets: C1OD, C1TD1, C1TD2, and C1FD.
A dataset score is computed by taking the geometric mean of all power system network models in a given dataset.
An optimal power flow (OPF) expresses the conditions which give the lowest cost per unit of energy delivered. This is achieved by minimizing an objective (cost) function by changing system controls while taking into account various constraints that are used to satisfy power balance requirements and limits. An alternating current (AC) OPF is a nonlinear model; nonlinear because the power flow into load impedances is a function of the square of the applied voltages. A direct current (DC) OPF is a linear, but less accurate, power-flow model.
Unless otherwise specified, OPF is assumed to be (AC) OPF in the context of the GO Competition.
power system network model
A power system network model consists of each hypothetical grid with defined topological structure and characteristics including, but not limited to, locations of generators, loads, transmission lines, transformers, equipment detail, control equipment, and limits.
power system network model score
A (relevant to Divisions 1 and 2 of Challenge 1) power system network model score is calculated by taking the geometric mean across all scenario scores associated with a power system network model.
The preventative security-constrained optimal power flow problem is the form of the SCOPF problem in which none of the contingencies will result in a constraint violation in the absence of corrective action.
The term "scenario" in the context of power systems often means a load snapshot from a coupled sequence of time data. That is not how it is used here. Here a scenario is used to represent an operating instance in time on a power system network model. The scenarios define an instantaneous demand at each bus, renewable resource availability, and other temporary system conditions. This snapshot instance has no correlated relationship to any other instance in the list of scenarios in the dataset. Not only can the instantaneous power demand (load) change between one of the scenarios and another, but so can any of the other parameters, such as line and generator limits and availability. The basic network model, the number of buses, generators, lines, capacitor banks, LTC taps, etc., remains fixed for all the scenarios within a power system network model dataset.
A scenario score (relevant to Divisions 1 and 2 of Challenge 1) is calculated for each scenario of a power system network model. The score consists of the objective function value plus any constraint violation or time penalties. As with the objective function value, a lower score is a better score.
A security-constrained optimal power flow problem not only minimizes the objective (cost) function for the base (no contingency) case, but provides a solution where all constraints are also satisfied for a set of post-contingency states. It can be formulated in either preventative (PSCOPF) or corrective (CSCOPF) mode. In the preventative mode, none of the contingencies will result in a constraint violation in the absence of corrective action. In the corrective mode, however, remedial actions may be necessary to remove constraint violations.