Optimising power systems by automating large sets of simulations

Abstract

Power system planners need to perform large sets of simulations to assess the performance of the system in its current or future state. The management of all these simulations and the analysis of their outputs turns into a challenging task. Using these outputs, documents must be generated, including a lot of plots for reporting purposes. This kind of study requires a great amount of effort, making it unaffordable to explore all desired power system scenarios to find the optimal configuration for each one. This happens because simulation tools have considerable limitations when it comes to performing exploratory experiments to find optimal configurations. Using optimisation tools is not the solution either, as these are also limited in their ability to express the complexity of the power system being studied. In this paper, a solution for the automation of power system planning studies is presented. PySelf is a software framework that allows handling several simulator definitions, performing simulations and optimisations and generating code for debugging and document elaboration purposes. This framework has been successfully tested with a real use case that involved the execution of more than 1,000 simulations in the context of a task for the European project OSMOSE (agreement number 773406).