| Direct current system is an important working power source for power plants and substations, it supplies reliable, uninterruptible power supply to secondary equipments, affecting its normal work and operation. Circuit branch grounding is the most common faults in DC system, how to detect it effectively is a main issue for DC system's safety operation. So, it's of great significance to study ground fault detection for the safety and reliability of DC system and even the steady operation of the entire power system.Based on the analysis of the basic principles, applications and related algorithms of DC system ground fault detection, focused on the difficulties such as the effects of earth capacity, mode of connection and ripple voltage when using Low-frequency Building method and earth resistance unmeasurable under strong noise, according to the anti-noise performance and high-sensitivity characteristics of Duffing oscillator, this paper presents a new method for detecting DC system ground fault.This method remove the DC components in fault characteristics using double-Hilbert transform firstly, then determine the status of the system with Lyapunov index and the phase diagram of Duffing oscillator system as a comprehensive criterion. And then by the condition of phase transition to calculate the amplitude and phase of characteristic quantities, get the value of earth resistance, in order to determine the grounded branch. By comparing with Fourier method and Morlet wavelet method, the validity of this method is verified. Compared with traditional methods, this method can not only improve the precision of detection, but also prevent misjudge or missing detection, and the detection time is determined by the time needed for judging the phase state of system. Finally, the effects of the Duffing oscillator detection system parameters (discrete step h, the evolution time T, damping ratio k and the non-linear resilience) on the response characteristics of the system is discussed.
|
PDF Full Text Request