Studies On Novel Principles And Novel Methods Of Self-diagnosis And Self-healing In Response To Distribution Network Faults

The security problems of distribution network have not attracted enough attention for a long period of time. Therefore, the level of distribution automation lags behind the development of substation automation. In view of increasing of complexity of structures and components of power systems, it becomes more difficult to identify the single-phase earth fault of distribution networks in terms of high correct rate. Furthermore, it still needs in-depth study on how to eliminate the fault automatically by means of arc-extinguishing equipment. The researches on discriminating the faulted section in a distribution network with complicated structures need to be reinforced, especially in the condition of uncompleted information of breaker and protection tripping, even the mal-operation of breaker, mal-tripping or fail-to-trip of protection. Therefore, a series of work on faulty feeder selection in a neutral un-effectively grounded system (NUGS), a all-in-one technique to integrate the functions of arc-extinguishing and faulty feeder selection, and novel fault diagnosis scheme on distribution network based on intelligent theory are conducted, as. outlined as below:A novel selective protection scheme for single-phase earth faults occurring on the feeder of a distribution network based on local measurement is proposed. The fundamental of this scheme is that the incident wave travels to the busbar and penetrates to the system side and sound feeder when a single phase earth fault occurs on a NUGS. In this case, the transient zero-sequence current and negative-sequence current through the faulty feeder will all have maximum magnitudes. To overcome the difficulty of the kind of methods, that is, all feeder currents need to be collected together to compare, a novel method is proposed. On the basis of zero-sequence overvoltage protection with inverse time delay operating characteristic, a revised factor that reflects the magnitudes of the transient zero-sequence or negative-sequence current is designed to dynamically regulate the inverse time-delay characteristic curve. The selectivity of the protection is achieved according to different tripping times. Based on above designs, the coordination between the zero-sequence based criterion and negative-sequence based criterion for improving the reliability of feeder selection is discussed.A novel faulty feeder selection method is proposed based on zero-sequence compensated admittance. Taking advantage of parameter information of the feeders of the distribution network, the susceptance component of the measured admittance is compensated. In this case, the discrimination redundancy between the faulty feeder and the sound feeder is highlighted. In this case, a single-phase fault protection without depending upon the collective measurement of all feeder currents is implemented. The correct rate of this method is not influenced by the system structure, earthed mode of the neutral and fault resistance.A novel scheme to detect the single-phase earth fault and select the faulty feeder is proposed, which is implemented with mathematical morphological filter and the identification of transient zero-sequence current wavehead. By means of the in-depth investigations on the fault scenarios, it is pointed out that the criterion simply based on the transient zero-sequence current polarity may fail due to the interferences and sampling error. Therefore, based on the fundamentals of mathematical morphology, a generalized morphological combined filter combining open-close and close-open operators is designed to detect the real polarity information of the transient zero-sequence current. The faulty feeder can be selected correctly by means of the polarity and the magnitude of the wavehead. The proposed morphological filtering operator has strong ability to resist interference. Therefore, it is also suitable for preprocessing of the data used by all types of transient based criteria.A novel self-diagnosis and self-healing scheme of distribution networks by integrating arc-extinguishing and faulty feeder selection functions is proposed. For the NUGS whose neutral is earthed via Peterson coil, the changing law of zero-sequence current with respect to the coil inductance is studied. Based on this, the change of the system zero-sequence parameters is proposed to be monitored on-line, and it should be the base to equip the arc-extinguishing system with appropriate parameters. As soon as a fault occurs, the inductance tap of the coil should be switched dynamically and the faulty feeder will be detected according to the change of the zero-sequence current. Afterwards, the coil is regulated to the resonant state to extinguish the fault arc. Therefore, the dual functions of arc-extinguishing together with faulty feeder selection are implemented in a single device.A novel method for fault diagnosis in a distribution network based on AI combined technique is proposed. The objective function of conventional fault diagnosis model ignores the coordination between the main protection and the backup protection, which will lead to the non-unique solution. Aiming at solving this problem, the objective function is revised and optimized. On the basis of this, the problem-solving efficiency is improved by virtue of the combined algorithm integrating GA together with Tabu Search, namely, GATS. In theory, the global optimization solution can be obtained by mean of GATS, and the searching efficiency can be enhanced as well. This novel algorithm is insensitive to the original solution, and therefore more robust.