Classification of voltage dips based on time-frequency parameters

S. Subhani,L. E. Weldemariam,V. Ćuk,J. Cobben
DOI: https://doi.org/10.1109/ENERGYCON.2016.7514012
2016-04-04
Abstract:The objective of this preliminary study was to investigate if voltage dips induced by different root causes have specific time-frequency patterns in their respective voltage dip waveforms by which the underlying root cause can be uniquely identified. Finding specific frequency patterns in the voltage dip waveform provides an innovative basis for the automatic classification of voltage dips based on their probable root cause. Ultimately, progressive insight into voltage dips allows more effective mitigation responses and thereby, conserves economic capital. Time-frequency parameters were obtained using short-time Fourier transforms. Subsequently, cluster analyses was performed to determine if time-frequency parameters from different voltage dip root causes were distinct. The outcome of the cluster analysis indicated that voltage dips which originate from different root causes have distinctive time-frequency patterns within the recorded voltage dip waveform.
What problem does this paper attempt to address?