| In order to ensure power supply reliability and security, many countries including china adopt the compensated neutral point of arc-suppression coil in medium voltage distribution network especially in enterprise high voltage power network with flammable and explosive environments, called compensated earthed network. The technology contradiction between arc-suppression and fault-line detection has become a difficult problem due to weak fault current, unstable fault arc and complex boundary conditions in this mode. So, the protection of fault-line detection hasn't been settled satisfactorily and become one of key for long time in compensated earthed network. Thus, it is necessary to investigate the basic theory and key techniques for enhancing integrated electrical safety in the environment of explosive gas.The protection method of fault line detection using transient signals produced has better sensitivity and reliability than that using steady fault signals. Considerable work has been done to study how to use transient signals to detect fault-line. However, these studies mainlytake a rather simplified fault model and emphasize the advanced mathematical tools, which result in the lacking of depth analysis to the fault mechanism and transient characteristics.In view of the fundamental study lack in fault line detection by transient information, the electromagnetic-transient model is set up for single-phase-to-ground fault, and then the transient characteristics and fault essence are analyzed using the high-order circuit dynamic equations. The fault mechanism and universal regulars are fundamentally revealed with the phase-mold transformation technology and fault component principle. The functions on transient characteristics and fault conditions (fault time, transition resistance, line parameters and length) are analyzed to improve fault model accuracy.For the problem of the complexity existing in boundary condition uncertainty and strong-coupling fault conditions for single-phase-to-ground fault, three universal applicable parameters, including energy ratio factor, impedance factor and energy entropy, are defined. The degree of uncertainty and complexity of fault transient are quantitatively measured by means of mining abundant fault data with defined parameters. The spectrum energy distributions are analyzed and described according to energy entropy measure and saturation entropy. The results indicate that there are some laws based on multi-scale energy statistic, and they can be used to cluster fault model. At last, the single-phase-to-ground fault is divided three classes: strong fault, small angle fault and weak fault.Aiming at overcoming the influences of the fault conditions on fault-line detection based on the transient components, the paper presents a new adaptive method conforming to the different fault model. The characteristic band is defined and adaptively taken out for each line based on the spectrum energy distribution. And then, the adaptive faulty line detection method is proposed. When the strong fault occurs, the multi-frequency bands correlation analysis is adopted to detect fault-line; when the small angle fault occurs, the low-frequency band energy comparison is selected to detect fault-line; when the weak fault occurs, the wavelet energy entropy principle is chosen to diagnosis fault-line. All information of the transient zero-model currents including low frequencies, fundamental frequencies and high frequencies are utilizedin proposed method and the reliability can be effectively guaranteed for protection of fault-line detection in compensated earthed network.The short-circuit arc in medium voltage distribution caused by single-phase-to--ground fault burn in free air, called as air arc at low current in long gap with complex characters. The nature of this type of arc characters is difficult and extremely nonlinear. In the paper the form mechanism and the process of the arc waveforms are explored with the arc theory and the physical doctrine. Dynamic volt-ampere characteristic of arc in time-frequency are explained according to the energy balance principle. The black-box model for arc is further constructed according to the device-molding method of nonlinear circuit theory. Electromagneticsimulation and theory analysis have proved the proposed protection algorithm is not influenced by unstable arc, harmonic component and compensated degree.Vast simulations are carried out by the EMTP software. The security and reliability of the proposed method of detection fault line can be further proved by the test results from 10kV high voltage physical experimental system and the field measured waveforms. |