Traveling Wave Based Fault Location

Abstract

Transmission grid system plays a key role in the interconnection of large-scale generation of a country’s power system network supplying loads all across. In many faults statistics, it has been unambiguously indicated that around 75% power system faults occur in transmission lines. This fact has increased the need for correct fault identification and location on transmission lines. Many researchers have developed algorithms based on using current, voltage measurements, and line parameters at a single end or multi-end to locate the fault. However, each of these fault location technique has some operational limitation. This thesis presents single-ended traveling wave fault location method based on wavelet transform and practical implementation on a recently released commercial relay utilizing this principle. A sample system was created in MATLAB SimPower System and simulations carried out for assessing the traveling wave from one end for different fault distance on the transmission lines. Traveling wave propagate close to the velocity of light and it is difficult to capture them so sampling rate was kept high enough to record the traveling wave at fault point. In order to decouple three phases, modal transformation is used to carry the independent analysis of each phase. Signal processing application wavelet transform has been used for feature extraction. MATLAB wavelet toolbox is used to decompose the signal by using discrete wavelet transform with appropriate mother wavelet Daubechies at level one which yields detail coefficient and approximation coefficients. Thereafter, Wavelet transform modulus maxima of detail coefficient is used to extract the sample number of the peaks in the signal. Firstly, two consecutive peaks with their respective sample numbers are used along with the velocity of propagation. Sensitivity analysis is carried out to analyse the system performance under different conditions. Extensive simulations are performed by varying the fault inception angle, fault resistance and mother wavelets showing that the traveling wave fault location method is insensitive to these parameters. All the simulations were performed by keeping the sample rate higher since by keeping sampling rate lower it is not possible to capture the realistic traveling wave which thereby impacts on the performance of any practical relay being tested. In addition to this detailed research assessment carried out, a two bus system network has been modelled using ATP/EMTP and the generated fault signal applied to a new commercial relay (M/s Schweitzer Engineering Laboratories, SEL) using a commercial testing system. By use of synchrowave event software, the fault signal is analysed and the fault distance manually calculated. The average error was around 2% for fault location but relay trip indicated correct fault classification. Analysing the various fault cases it is clearly observed that it is difficult to distinguish between the waves reflected from fault point and from the remote end of the line. This is one limitation of traveling wave fault location method. Low sampling rate and velocity of propagation appear to affect the observed results. The velocity of propagation should be calculated by using the line parameters of the transmission line and sampling rate should be kept higher to get estimated results from modelling. In future, this particular relay should be tested in the field by using the double-ended traveling wave fault location method.