eduzhai > Applied Sciences > Engineering >

Evaluating Different Machine Learning Techniques as Surrogate for Low Voltage Grids

  • Save

... pages left unread,continue reading

Document pages: 16 pages

Abstract: The transition of the power grid requires new technologies and methodologies,which can only be developed and tested in simulations. Especially largersimulation setups with many levels of detail can become quite slow. Therefore,the number of possible simulation evaluations decreases. One solution toovercome this issue is to use surrogate models, i.e., data-drivenapproximations of (sub)systems. In a recent work, a surrogate model for a lowvoltage grid was built using artificial neural networks, which achievedsatisfying results. However, there were still open questions regarding theassumptions and simplifications made. In this paper, we present the results ofour ongoing research, which answer some of these question. We compare differentmachine learning algorithms as surrogate models and exchange the grid topologyand size. In a set of experiments, we show that algorithms based on linearregression and artificial neural networks yield the best results independent ofthe grid topology. Furthermore, adding volatile energy generation and avariable phase angle does not decrease the quality of the surrogate models.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×