Detection of faulty components on power lines using radio frequency signatures and signal processing techniques

This paper describes a new technique for detecting faulty components such as cracked insulators and faulty surge arresters on power lines. The technique utilises radio frequency (RF) pulses emitted from the faulty components as the monitoring parameter. The high frequency radio pulses are captured using a biconical antenna with a frequency range of 30 MHz to 300 MHz. The radio frequency pulses are stored analysed and characterised using signal processing. Good characterisation of RF pulses in the time domain has been achieved using parameters obtained from statistical analysis such as mean median, peak and auto-correlation. Second analysis method applied for characterising the pulses is the fast Fourier transform (FFT). In order to extract more informative features from the RF pulses, a third analysis, joint time-frequency transformation, is used. These three different methods have successfully revealed unique signatures of different faulty conditions on power line components