• it/

Search in the site by keyword

Publications - Paper

A Machine Learning Based Tool for Voltage Dip Classification

Related tags

#Electricity Network #Quality of Supply

Publications - Paper

A Machine Learning Based Tool for Voltage Dip Classification

The paper presents a Machine Learning-based tool for making the ex-post analysis of voltage dips (VD) more automatic and effortless. The tool takes as input the full waveforms associated with voltage dips occurring in Italian MV networks and recorded by the QuEEN monitoring system implemented by RSE.

The first tool was developed to classify events based on their HV/MV origin, since utilities will only be responsible for events due to failures occurring in their own networks; it uses Kalman Filter and Support Vector Machine (SVM) self-tuning to extract VD characteristics and classify events, respectively.

On the other hand, the second tool, based on end-to-end Deep Learning, has been developed to distinguish between ‘true’ and ‘false’ voltage dips (VDs); it uses a Convolutional Neural Network (CNN) whose first layers are responsible for feature extraction while the last layers perform event classification.

Download Memory

Rif.RSE​

18006134

Destination​

CIRED 2019, The 25th International Conference and Exhibition on Electricity Distribution, June 3-6, 2019

Publication Date

06 June 2019

Authors​

H. Shadmehr, L. Tenti, R. Chiumeo

Publication Type

Paper

Projects

Development and management of distribution grids (SmartDsys2018)

Related tags

#Electricity Network #Quality of Supply

Comments