Tue 20.10.Wed 21.10.Thu 22.10.Fri 23.10.


Customer Based Applications 2 - student section

Session CA-2Main

09:40 - 10:45 CEST

Session chair: Kruno Miličević

Play session recording

Customer Based Applications 2 - student papers

09:40 - 10:00 CEST

Saeed Mohammadi, Mohammad Reza Hesamadeh, Ali Vafamehr, Farzad Ferdowsi

Paper Slides

Liberalized electricity markets have been studied for the past few decades with different mathematical techniques. Operating these markets under the growing uncertainties has been challenging in many jurisdictions. Given the recent advances in machine learning techniques and big-data analysis, applications of such techniques have been growing in recent years. In this paper we review state-of-the-art developments of machine learning techniques and their applications in electricity market studies. We briefly provide current market challenges around the world. Then we show how these challenges are addressed using different machine learning approaches. Later, we provide a comparative table where all relevant papers are compared. This table can guide future studies on the machine learning applications in electricity markets by highlighting the promising potential areas. Then, we suggest directions for future researches to pursue machine learning applications in electricity market. Consequently, these approaches can be employed to resolve the uncertainties.

10:00 - 10:20 CEST

Filip Barisic, Kristijan Špoljarić, Hrvoje Hegeduš, Petar Mostarac

Paper Slides

In this paper, a simulation system is presented that measures the stress and contraction of a displacement transducer. The displacement transducer specifications are achieved by the voltage divider described in the paper. Structure of voltage reference, voltage divider, DC voltage generator, instrumentation amplifier and two different data loggers is presented. System was tested in a way that voltage divider produced three voltage levels of -4 mV, 0 mV and +4 mV. Analyzing these three voltage levels conclusions were made about resolution, noise and accuracy. Particular system is tested in laboratory conditions.

10:20 - 10:40 CEST

Marko Kovačić, Marko Jurcevic, Roman Malarić, Antonijo Kunac

Paper Slides Video

In this paper the most commonly used algorithms for estimating phasor parameters are presented. The most algorithms based on recursive and non-recursive DFT algorithm are presented too. For each algorithm is shown: advantages, disadvantages, mode of operation and possible future research. Furthermore, algorithms for estimating phasor parameters based on finite and infinite impulse response filters are presented. For each of them the mode of operation, advantages and disadvantages are shown. Future research is possible for a filter with a infinite impulse response due to better characteristics compared to filters with a finite impulse response as they detect signal more accurately and faster in half cycle period of frequency. Wavelet transform ie. its recursive version is also presented in this paper as they have been proven to be very good for aplication in PMU devices because these algorithms eliminate DC component and work much faster than algorithms based on finite or infinite impulse responses filters. Wavelet transformations have great potential for future research in the field of estimating phasor parameters using new wavelets, multiresolution analysis, connection to filter records and more.

Coffee break


10:45 - 11:00 CEST

Closing ceremony


12:15 - 12:45 CEST

Session chair: Roman Malarić, Marko Jurcevic, Branimir Ivšić

SlidesPlay session recording