Targeted Scenario Generation in Low-Voltage Distribution Grids Using a Genetic Algorithm

This transformation introduces new operational behaviour and associated challenges. Among these challenges are limit violations, such as voltage band violations and thermal overloads of lines and transformers, which must be considered in both the planning and operation of distribution grids. To adequately address these challenges within existing planning and operational approaches, as well as to develop new methods, scenarios covering a wide range of operating conditions and adverse situations are required. Such scenarios form the basis for targeted grid studies, for example in the context of grid expansion planning or operational analysis.

To address this need, this work proposes a framework for targeted scenario generation. The framework employs a genetic algorithm to solve an optimisation problem that identifies suitable combinations of installed technologies together with randomly selected corresponding time series. This approach enables the generation of time-series-based scenarios that fulfil predefined violation requirements, ranging from line and transformer overloads to under- and overvoltage violations, as well as combinations of these violation types.

In addition to specifying the type of violation, the framework allows the definition of where and when violations should occur, which grid elements are affected, and the duration of the violation within the considered time window. This work focuses exclusively on low-voltage distribution grids. Future work could extend the proposed framework to applications at higher voltage levels.

The functionality of the framework is demonstrated using a CIGRE low-voltage benchmark grid model, showcasing the ability to generate different scenarios, including a pure uppervoltage violation, a pure line-loading violation, and a combined scenario affecting a selected set of grid elements within a specified violation time window.