Research On Feeder Protection Method In Distribution Network Based On Signal Characteristic Identification

Distribution network is an important part of electrical power system to undertake power distribution,and its security and reliability are crucial.However,in China,the development of distribution network is inadequate and imperfect,and there are many security risks.Specifically,the structure of distribution network in China is wide and weak,so that faults frequently occur.In addition,the level of automation and intelligence of distribution network in China is relatively low,which results in its feeble ability of fault detection.These drawbacks lead to much lower accuracy of distribution feeders protection action and fault location than that of transmission lines.Consequently,it is important and urgent to research new and reliable feeder protection technology in distribution network.Supported in part by the National Natural Science Foundation of China,in part by the Science and Technology Project of State Grid,and in part by other technology projects,according to practical engineering background and feeder protection requirements,this paper investigates a series of electrical signal characteristics and characteristic identification methods,and applies them to feeder protection.On this basis,specific action criteria and protection processes are presented,the adaptability of the presented algorithms are analyzed,and reliability tests and demonstration projects are implemented.The main work and innovations are summarized as follows.1)A feeder section differential protection method for active distribution network is proposed.The proposed method is especially suitable for feeders where voltage transformers are not equipped or both real-time communication and time synchronization ability of feeder terminal units(FTU)are weak.Firstly,the fault starting algorithm is used to trigger FTUs on the boundary of feeder section to exchange fault current sampling data with each other.And then,the dynamic time warping(DTW)distance is used as the signal characteristic parameter to identify the operation state of feeder section.At last,according to the DTW distance,the protection criterion is built.Besides,the applicability of the proposed algorithm to active distribution network is also investigated.With the cooperation of fault starting algorithm and DTW algorithm with strong anti-synchronization error ability,the proposed method can be adapted to the aforementioned engineering applicationbackground,and it is suitable for active distribution network with multi-branch feeders.The effectiveness of the proposed method is proved by simulation tests,and its superiority is verified by asynchronous data tests.2)In order to deal with the contradictions between bidirectional flow of fault power flow and absence of power flow direction discriminator in closed-loop distribution network with multi-power supply,a feeder differential protection method based on phase change is proposed.On the basis of analyzing the phase relationship between pre-fault feeder current and post-fault positive sequence fault component current,the differential performance of phase change in different operation states of feeders is deduced.And then,the feeder differential protection criterion is constructed by taking phase change as a characteristic identification parameter.In addition,the reliability of the proposed protection method is demonstrated,according to the theoretical analysis about protection device asynchronous measurement and feeder current change.The proposed protection method can be realized with existing third-party communication networks,without requiring special optical cable communication channels.And it does not need to install voltage transformers on the feeder to construct direction discriminator elements,so as to facilitate its engineering application.The simulation tests verify the correctness of the proposed protection method.Besides,the comparative analysis demonstrates that the proposed protection method is superior to some other feeder protection methods.3)Aiming at the problem that feeder protection is susceptible to the disturbance of inrush current in distribution network,a distortion elimination and waveform reconstruction scheme based on waveform identification is proposed.Firstly,least square estimation method is used to calculate the instantaneous amplitude of feeder current.According to calculated instantaneous amplitude,the portion segmentation is implemented in time domain.And then,via the ratio of boundary gradient,distortion portions are accurately identified.After eliminating distortion portions,the signal waveform is reconstructed by using retained portions.Besides,in order to eliminate the interference of Current Transformer(CT)saturation,the function of CT saturation current identification is also included in the proposed scheme.Finally,the feasibility and validity of the proposed scheme are verified by the simulation evaluation tests under different operating conditions and the actual recording data tests.Compared with the existing methods of inrush identification and blocking,the advantages of the proposed scheme is listed as follows.The proposed scheme caneffectively deal with not only the inrush current superimposed with sinusoidal load current but also saturated fault current.In addition,it never block or interrupt the relay protection,which opens a new door to deal with inrush current.4)In order to improve the ability of single-phase-to-ground fault detection and the accuracy of fault feeder selection in neutral ineffectively grounded distribution network,new single-phase-to-ground fault detection method and fault feeder selection method are proposed,according to signal characteristic identification based on kurtosis parameter and skewness parameter.On the basis of analyzing the mechanism of single-phase-to-ground fault and the characteristics of transient zero-sequence current,single-phase-to-ground fault is detected by using the kurtosis parameter and the zero-sequence voltage criterion,and the starting time of fault is determined accurately.According to the distribution law of transient zero-sequence current in the system,fault feeder selection criterion is built by using the skewness characteristics of signal numerical distribution.At last,the validity of the proposed method under different fault states is verified by digital simulation tests and field recording data tests.It is founded that the proposed method can work well even in the case of unbalanced three-phase voltage,intermittent arc grounding,and strong noise interference,which indicates its reliability.5)Aiming at the problem that the accuracy of single-phase-to-ground fault location is easily affected by the asynchronous start-up between devices,a single-phase-to-ground fault location method based on DTW distance is proposed.On the basis of analyzing the difference of transient zero-sequence current characteristics at the upstream and downstream of fault point,DTW distance is used as a signal characteristic parameter to identify the difference of transient zero-sequence current at feeder section boundary.And,fault search strategy matching DTW distance criterion is proposed to realize fast and accurate fault location.According to the error tolerance of DTW distance algorithm,the proposed method is able to cope with the problem of asynchronous start-up.Therefore,the proposed method has strong robustness and is suitable for practical distribution network projects with various types of devices and uneven quality levels.Fault simulation tests and error tolerance tests verify the feasibility and superiority of the proposed method.Moreover,the proposed method has been applied to a practical project of single-phase grounding fault location,showing a higher correctness rate of fault location.