Research On Fault Classification And Location Based On Improved ANFIS Algorithm

High speed development on electric industry and the unbalance of energy distribution make the high voltage power transmission develop rapidly with its features of long distance and less power loss. It has a deadly impact on industrial production and people’s life when it comes to the fault in high voltage power transmission. So it is important to confirm the type and location of the fault.It should make sure that the type of fault be confirmed before measuring the distance of the location of fault in transmission lines. This paper simulates the fault in different locations and different types in high voltage power transmission using PSCAD/EMTDC and analyzes simulation waves which show that current has more great changes than voltage when faults happen. Thus, this paper extracts fault current and classify fault type using improved ANFIS algorithm by SCG(Scaled Conjugated Gradient). The improved ANFIS algorithm has a faster convergence speed than the standard ANFIS algorithm and it cost less training time and training steps. What’s more, this algorithm has a higher accuracy in fault classification.The next step after fault classification is measuring the location of fault. Traditional measurements include wave ranging method, single-ended ranging method, the double sides synchronous distance method and the double sides synchronous distance measuring method,but the double sides synchronous distance measuring method turns out to be the greatest method which hardware is no needed compared with the other method, this method ensure the accuracy of the fault distance. However, line parameters is not fixed in the grid and it has a negative impact on the precision of fault distance.Therefore, Robust Total Least Square Algorithm is used in the parameters estimation of transmission lines. And the transmission line parameters of three most common kinds of line faults of power system:single phase to ground fault, two-phase fault and double-line to ground fault are estimated, and the fault locations are also determined by using the obtained parameters. The calculation results show that Robust Total Least Square Algorithm have certain advantages in the research of fault location, which greatly improves the accuracy of parameters estimation and ensures the accuracy of fault location.