Power Quality Issues Dataset Using Short Time Fourier Transform (STFT)

Power Quality (PQ) has become one of the most crucial issues in contemporary power systems because of the large-scale growth of renewable energy sources, the existence of complex industrial loads, and sophisticated electrical and electronic apparatus. A properly designed dataset is necessary to detect and categorize various PQ disturbances with the help of artificial intelligence (AI) techniques. In this work, the PQ dataset is created through a method of Short-Time Fourier Transform (STFT). There are five key disturbances of PQ that are considered, which are voltage sag, swell, harmonics, interruption and transients. To obtain the original dataset, 100 signal samples are produced for each disturbance, and 2091 features are extracted using STFT. To reduce the burden of computation, the standard deviation method is conducted, which leads to a reduced dataset with 794 features per sample. A total of 500 samples and 794 features are obtained, and all PQ disturbances are numbered as follows: 0 Sag, 1 Swell, 2 Harmonics, 3 Interruption, and 4 Transients.