The Key Technologies Of Traveling Wave Fault Location For Transmission Lines Based On The Measured Data

Traveling wave fault location is a key technlogy for locating faults for transmission lines. For the applied traveling wave fault location devices and methods, the drawbacks of lack of automatica, strict restriction in time synchronization and communication hardware, low reliability and is vulnerable to be affected by a variety of factors are fully exposed. In this dissertation, based on the measured field data, the comprehensive research were done, according to the difficulties in contingency screening of data record, identification of traveling waves’ wavefronts, fault location methods, intelligent decision of location methods and parameters, and optimal placement for traveling wave devices, in order to enhance crediting degree of measured data and comprehensive fault location performance.The contingency screening method based on constrained semi-supervised clustering is proposed, in order to overcome the difficulties in efficiently screening the contingency data from the massive multi-channel dataset. General modal current components were defined to reduce dimensionality and extract feature at first. Clustering was implemented considering the neighborhood constraints of unlabeled sample set in time domain and the sugesstion provided by a small amount of labeled historical sample. The proposed method is shown high detection rate by actual dataset’s test.The demarcation method for traveling wave based on line detection by Hough transform is proposed, in order to solve the problem in automatic detecting and identifying the wavefront among the measured field data. By analyzing the propagation of traveling waves on the faulty line, the selection of the analytical channels, and the feature of measured waveforms, Hough Transform is used to detect the wavefronts in the rational window. The disturbances caused by reflection of healthy neighbouring lines and remote busbars can be eliminated through the amplitude and slope of the surges. The confidence of wavefront identification can be determined through the interval of series fault reflected surges and the consistency of the initial wavefront slopes in different scope, which avoids the mistakes caused by irrational experienced scope. The proposed method possesses high anti-noise performance and is insentive to local defects of waveshapes.Three novel fault location methods are proposed, in order to solve the inherent technlogy limitations of current based single and double ended traveling wave fault location methods. Under the condition of available sound path between the two-side buses of the faulty line, a single-ended fault location method using the arrival time difference of fault induced dominant traveling waves between faulted and healthy lines at the same substation is proposed, and it has the advantages of independent of communications or timing synchronization with remote side, and is easier for identifying wavefronts and without dead-zone for the nearby faults. For cascade topology type transmission lines, the conditions of direct extension of single-ended method were analyzed. For the faults which not satisifie the direction extension conditons, the single-ended fault location indirect extension is proposed according to the associated arrival time of traveling waves in two sub-windows, which extends the range of traditional single-ended traveling wave location. For double-ended fault location, according to the features of sequence of arrival time for traveling waves at two terminals, the fierce and weak faulted modes can be identified through the matched sequence of initial and follow-up double-ended surges, and the separated two-ended traveling wave fault location collaborative approaches independent of two–ended time synchronization are proposed. Both of the two approaches avoid the risk of misleading fault location result caused by asynchronism of two side clock or mistakable wavefront identified by single-ended approaches.In order to improve the consistency between calculated and engineering so-called value, the cross-database or cross-platform intelligent fault location method via reusing historical cases is proposed. Faults often occur incidentally and repeatedly to a certain extent. The performance of fault location should be improved via reasonably resuing the parameters and results according to the similarity between current fault and small amounts of valuable historical cases. Based on this thoughtfulness, the case library was formluated by each case’s description using triples <features of faults, fault locating parameters, fault locating conclusions>, and the parameters of wavefront identification methods can be reused by case matching according to the feature of the proportion of the initial peak, approximate entropy, and Hu invariant moment in normalized waveshapes, then repetitive faults can be identified. Historical conclusions can be reused directly for repetitive faults. Conversion coefficient of the optimal matched case can be reused for high similarity non-repetitive faults, which is helpful to narrow the range of investigation for line’s fault.The optimal placement strategies for current based traveling wave fault location devices is proposed, in order to improve the rationality of traveling wave fault location systems under different development stages of power grids. On the basis of mathematical description of the faulted traveling wave currents’ observability in the power grid, the expanded integer programming model for traveling wave devices’ placement is formulated, considering the limitation of practical engineering and constraint conditions of the observability analysis. Based on the static optimal solution, the dynamic sequence of the each device’s installation is determined, according to the minimum decreasing weighted observable length of the lines by single and double ended fault location principles caused by each device’s retreat. The proposed method is capable of taking both the economy and coverage into account for traveling wave fault location systems.