Hardware implementation of an islanding detection approach based on current and voltage transients

A pattern classification technique for fast detection of power islands in a distribution network is implemented and tested. It utilizes voltage and current transient signals generated during an islanding event to detect the formation of the island. A Decision Tree classifier is trained to categorize the transient generating events as `islanding' or `non-islanding'. It involves two basic stages of signal processing to extract the required feature vectors for the classification. The first stage involves signal filtering and in the second stage signals are processed by rectifying, summing, and low-pass filtering to get the energy content in the three phases during a selected time-frame. Analog filters, rectifiers, adders and micro-controllers complete the implementation. The performance of the design was tested with signals generated using a real-time waveform playback instrument. A simple radial medium voltage distribution system with single distributed generator was simulated in PSCAD/EMTDC to obtain the transient waveforms. The experimental and simulation results give comparable results showing high accuracy in detecting islanding events very fast.