Recognition of Residual Current A+6 Waveform Based on Random Forest Algorithm

In the field of power system safety protection, residual current detection technology plays a vital role, which is essential to ensure the safe and stable operation of the system. Accurate identification of the waveform of the residual current is necessary for the residual current protection device to perform the disconnect operation correctly. At present, the recognition of specific waveforms such as F, A0, A90, A135, and SDC in the A+6 waveform is not very sufficient, especially in terms of compatibility and recognition efficiency at 50Hz and 60Hz frequencies. In this study, we propose a new residual current A+6 waveform recognition technology based on a random forest algorithm, which improves recognition accuracy by selecting 13 time-domain statistical features. The experimental results show that the accuracy of the proposed method for the recognition of alternating current, F, A0, A90, A135, and SDC waveforms and their amplitudes at 50Hz and 60Hz on the experimental bench reaches 98.6% and 99.3%, respectively. In addition, the generated decision trees can be easily integrated into embedded systems, showing high practical value.