Evaluating different machine learning techniques as surrogate for low voltage grids

Stephan Balduin,Erika Puiutta,Tom Westermann

doi:10.1186/s42162-020-00127-3

Stephan Balduin, Erika Puiutta + Show 1 more

Open Access

https://doi.org/10.1186/s42162-020-00127-3

Copy DOI

Abstract

The transition of the power grid requires new technologies and methodologies, which can only be developed and tested in simulations. Especially larger simulation setups with many levels of detail can become quite slow. Therefore, the number of possible simulation evaluations decreases. One solution to overcome this issue is to use surrogate models, i. e., data-driven approximations of (sub)systems. In a recent work, we built a surrogate model for a low voltage grid using artificial neural networks, which achieved satisfying results. However, there were still open questions regarding the assumptions and simplifications made. In this paper, we present the results of our ongoing research, which answer some of these questions. We compare different machine learning algorithms as surrogate models and exchange the grid topology and size. In a set of experiments, we show that algorithms based on linear regression and artificial neural networks yield the best results independent of the grid topology. Furthermore, adding volatile energy generation and a variable phase angle does not decrease the quality of the surrogate models.

Highlights

The ongoing transformationof the power system requires the involvement of new technologies and methodologies to meet the requirements that arise during this process
Our goal was to enable the creation of larger setups of medium voltage (MV) grid simulations using this kind of surrogate model as a replacement for some of the low voltage (LV) grids
When conducted on the second grid model, the surrogate models generally reached a lower root mean squared error (RMSE) value than they did on the CIGRE LV

Summary

Introduction

The ongoing transformationof the power system requires the involvement of new technologies and methodologies to meet the requirements that arise during this process. Since the power grid is a safety-critical infrastructure, simulation and hardware-in-theloop are used for the development and testing of such new technologies (Steinbrink et al 2017). Smart grid simulations comprise other domains, such as the Information and Communication Technology (ICT) domain, each of which are developed in their own simulation environment. With co-simulation, synchronization and data exchange between these different environments is handled by a co-simulation-framework Balduin et al Energy Informatics 2020, 3(Suppl 1):. Balduin et al Energy Informatics 2020, 3(Suppl 1):24 Building such a large, cross-domain simulation environment is still a complex task and the simulation of the overall system can become very slow (Blank et al 2015). Surrogate models can be used to reduce the simulation time of some of the components in the environment. A surrogate model is a data-driven approximation of a certain function or system, which can be evaluated faster than the original function or system, but which is less accurate (Simpson et al 2001)

Objectives

Methods

Results

Conclusion