Research on real-time location method of DC shock residual current based on SVM

To address the issue of the AC residual current positioning method's incompatibility with the DC system and the challenges in accurately investigating personal electric shock accidents caused by line leakage, the characteristics of the residual current in the low-voltage DC system are summarized according to the distribution law of human impedance and the influence of line impedance on the residual current, and a method of locating the residual current fault based on Bayesian optimization of Support Vector Machine(SVM) is proposed. The residual current generated by human electric shock can be located within 5ms, with a relative error of less than 2% in locating faults along a 1000m line. Based on the proposed method, an integrated prototype for low-voltage DC residual current acquisition and location has been developed. The effectiveness of the proposed positioning algorithm is verified by simulation and the test results of the prototype, and the real-time acquisition and fault location of residual current in DC distribution lines has been achieved, which provides a new technical method for fault location of low-voltage DC residual current. It has great application value and guiding significance.