| The single-phase earth fault is the most common fault in coal mine high-voltage power supply system. If the fault line cannot be effectively distinguished and timely removed, it easily develops into more serious two-phase short circuit fault, which will endanger the reliability of power network. The fault may cause over-grade trip and affect the normal operation of forcing fan, which will cause gas accumulation and induced explosion easily. So it poses a serious threat to personnel security. Therefore, in order to guarantee the coal mine safety production, power supply system should have the ability to accurately judge and rapidly isolate the single-phase earth fault. But for the neutral non-effective grounding system, single-phase ground fault current is usually small, and the steady-state characteristic is not obvious, which causing the detection methods of small current grounding fault cannot be well solved for a long time. Now the technology of fault line selection has almost matured. But the difficulty of fault location has not yet been resolved very well, especially lacking practical application in underground scene. This paper analyses the steady-state and transient-state of single-phase earth fault to find effective methods to select fault line and locate fault point. These methods will use line models and fault signal characteristics based on network type intelligent power supply system.In order to effectively solve the horizontal and vertical tripping accident caused by single-phase earth fault, this paper gives the network configuration based on WAMS and IEC61850 communication protocol. The research is mainly focused on several key technologies, mainly about how to apply the wide-area measurement technology in coal mine. Timing system should maintain high precision synchronization time service and hot backup function, so the system determines using GPS and Beidou system backup work mode. The scheme uses the mode of hierarchical collection and centralized monitoring in the protection system structure, which divided into station control layer, spacer level and process level based on the function and contact form. Protection configuration is unified by adopting the combination of local criterion and network criteria, which greatly improves the whole system reliability. By describing the modification project instances of this scheme, the fault diagnosis and remote real-time monitoring process is given according to the topological structure of power supply system.Reasonable line model can effectively improve the accuracy and efficiency of failure analysis, which is the basis and key of simulation to verify and locate the fault line. This paper firstly analyses three commonly used line models and introduces the modelling principle and application scope of these models. According to the external characteristic of two-port network, the H parameter and T parameter matrix equations of ?-type model, T-type model and distributed parameter line model are deduced, and the mutual relation between voltage and current on both line ends are given. In view of the actual mine high-voltage cable length is much smaller than the frequency wavelength, so it can be equivalent to a lumped element. Considering from the scope of application of different models, this paper determines the multi-? model as the equivalent circuit for single-phase earth fault simulation model. Because H mixed parameter can effectively reflect the short circuit and open circuit characteristics, frequency response analysis method is to be used to comparative analysis the H element frequency characteristics of ?-type model and T-type model, and the suitable equivalent frequency range is determined, which provides a theoretical basis for multi-? model is converted into multi-T model. By analysing the relation of both end parameters of multi-? model, a voltage three element difference equation is established, which is solved to get the line resonant frequency expressions.The steady-state line selection method requires the identifiability of fault signal amplitude. In order to measure the degree of steady-state signal attenuation, the "end power" comprehensive criterion of decay rate is proposed. Contrasting the united number and length of line models, multi-? model and distributed parameter model have the same attenuation level for steady-state signals. The fault line selection principle using characteristic frequency band is analysed. Comparing the first capacitive frequency of multi-? model and the maximum oscillation frequency of longest line, the number of ? model in transient analysis can be determined. Simulation verifies the double-? model can be equivalent to the distributed parameter model in the steady-state and transient analysis. According to the meaning of zero-sequence current at fault line selection points, the fault location method based on distributed zero-sequence current transformers is proposed. Fault distance formulas are given and the condition to realize the localization method is pointed out.By using differential protection theory, the fault line selection and positioning method based on differential variable of zero-sequence current is proposed. Firstly the distribution law of steady zero-sequence current in different feeders is analysed at different mine neutral grounding modes, and the zero-sequence current differences of polarity and amplitude between perfect line, fault point upstream line and fault point downstream line are discussed. This paper analyses the rule of fault zero-sequence voltage following the grounding resistance, and gives the fault phase criterion at two kinds of grounding models. The differential variable of zero-sequence current0?I is defined as the difference of zero-sequence current vectors at both ends. The grounding conductivity will influence the distribution of zero-sequence current vector, which makes an obvious difference in phase angle between perfect section and fault section. Combined with the amplitude relationship, the fault line selection and locating criterion is in established in different out-of-resonance degree, and the fault section judgement processes are given, which overcome the disadvantage of fault selection and location method based on steady-state signal isn't apply to the arc-suppression coil grounding system.Because mine high-voltage power system has radial multi-branch structure, which natural divided cable lines into several sections,the method has the way to practical application. The paper describes the method implement process in the network-type mine relay protection system. The result of MATLAB simulation verifies the zero-sequence reactive power direction selection principle is correct. The amplitude of zero-sequence current differential variable has an obvious difference between fault section and perfect section in neutral ungrounded system, so it doesn't need to use the phase relationship. But in arc-suppression coil grounding system, out-of-resonance degree will change the current amplitude of fault section greatly, even less than the perfect section, so the phase should be the main criterion. When the grounding resistance increases, the amplitude of each0?I will be significantly reduced, therefore the sensitivity of testing equipment should be improved accordingly. According to different grounding modes, the fault line selection problem at the end of branch lines can be solved by simulation results.Single-phase earth fault destroys the original power system steady state, and makes it from a short transient transition into a new steady state. Comparing with the fault steady signal, the characteristic of the transient signals contains more information, and its amplitude can reach more than several times with steady signal. Through the deep research of the fault transient process, the fault selection and location method based on transient signal is focus. According to the actual structure of mine high voltage power network, the equivalent circuit model of single-phase earth fault is established. Before failure occurs, zero-sequence voltage expression is given under unbalanced three-phase circuit parameters, and the equivalent conversion model is established by using Thevenin's theorem. There is a qualitative analyses for formation mechanism of zero-sequence current transient pulse, and the preliminarily conclusion is that the fault line and perfect line have opposite transient pulse. According to the equivalent circuit model, this paper establishes the zero-sequence voltage nonhomogeneous equation and the simplified expressions of zero-sequence current under the condition of overcompensation state can be solved. The analysis conclusion is the same as the qualitative result. This paper also analyses the distribution of zero-sequence current transient pulse polarity in the each level tie line switches, which is the foundation of the matrix positioning algorithm. Simulation results show that the pulse polarity of zero-sequence transient current in fault section with all higher levels are the same, and in contrast with other perfect lines. Due to the amplitude and width of transient pulse determine the accuracy and data window size of the signal detection device, simulation analysis the influencing factors for pulse amplitude and width and concluded the change rule following the fault angle ? and grounding resistance.This paper details the principle of fault location matrix algorithm based on net base structure and net shape structure and compares the merits and demerits of the two algorithms. The matrix algorithm based on net base structure needs normalized processing and the large amount of computation affects the timeliness of fault positioning. However, the matrix algorithm based on net shape structure doesn't need normalized processing and its criterion principle is clear, which can realize feeder terminal fault and multiple faults. The general matrix optimal positioning algorithm based on the polarity of zero-sequence transient current is proposed, which has the same advantages with net shape structure matrix algorithm. A network describing matrix that depicted grid topology was constructed, and fault feature matrix was proposed based on the transient pulse polarity of feeder switches. Finally fault sections could be fast and accurately judged by a new definition of optimal judgment criterion. There are four kinds of example simulations: the tie line fault between substations, feeder terminal fault, multiple grounding fault and the ability of algorithm error correction. The superiority and feasibility of this algorithm was proved by case simulation results. If the software based on the single-phase earth fault positioning algorithm is installed in the host computer of ground control center, it can realize fault section in the online monitoring scope, which can improve the ability of multiple line selective protection for mine high-voltage power network. Finally, the work is summarized and the next research direction was prospected. |
PDF Full Text Request