A Double-ended Fault Location Algorithm without Being Affected by Line Parameters

This study presents a ranging algorithm which requires only voltages and currents measured from the two terminals to establish ranging equations. It divides the line into a certain parts and uses every point as failure point to solve three groups of line impedances. When the function which is established by impedance data reaches the minimum, the corresponding point is right the failure point. Simulations show that it can be used for any type of failure and it may not cause ranging error for incorrect parameters.


INTRODUCTION
High-voltage transmission lines are the lifeblood of the power system, which bears transmit electrical energy.Meanwhile, it has a fault occurs in the power system in most places and is extremely difficult to find.Therefore, in line fault quickly and accurately find the point of failure, not only for timely repair lines and ensure reliable power supply and the power system security, stability and economic operation has a very important role.For a long time, people are seeking effective and accurate fault location method for a lot of research, a variety of fault location principles and methods are proposed.With the development of new technologies, fault location algorithm deepening abroad has become one of the hot spots.
According to signal characteristics, fault location method can be divided into frequency-based fault location and traveling wave fault location.Taking into account the economy, this paper study only frequencybased fault location.Many documents make further research and put forward the improvement method.Guo et al. (2003), Qi et al. (2005), Huang et al. (2009), Huang and Sun (2003) and Yu and Gu (2001) propose some improved algorithms of filtering decaying DC component based on Fourier algorithm.Liang et al. (2004) adopts distributed parameter model and it presents a fault location algorithm based on line parameter estimation and calculates the parameters of line ends, eliminating the line parameters and non synchronous angle difference influence on ranging precision.Document (Suonan et al., 2006) presents a double-ended fault location algorithm based on parameter identification in the time domain and it overcomes the traditional fault location method for measuring error of line parameters caused by the inaccurate.
Existing fault location algorithms are assuming that distribution transmission line is uniform and line parameters are known.So the exact line parameters directly determine the final degree of error ranging.In practical engineering, Lines tend to be unevenly distributed.Due to reasons such as geographic incomplete rearrangement, parameters are affected by the weather, the line height, etc.Given circuit parameters are often unreliable and the ranging accuracy is very low.To solve this problem, this study focuses on how to improve the basic algorithm of fault location and on this basis, the use of simulation experiments verify the accuracy of fault location methods and practicality.It presents a double-ended fault location algorithm without being affected by line parameters, the principle of the method is to use the idea of constructing a new match ranging function, this algorithm requires only two-terminal voltage measured by the amount of current that can be used Any type of failure, will not lead to inaccurate line parameters ranging error.Where Z s is self-impedance and Z M is coupling impedance line: (1 ) (1 ) Equation ( 3) and ( 4) can be obtained by the fault phase A. Where R is the line to ground fault resistance: In order to eliminate transition resistance Eq. ( 5) is obtained by Eq. ( 3) and (4): (5) Z s , Z M , l are the unknown parameters, which can be solved by Eq. ( 1), ( 2) and ( 5).
Principle of two-phase ground fault and phase fault algorithm: Two-phase ground fault is similar to phase fault.It provides AB phase ground failure as an example for analysis, as shown in Fig. 2. Equation ( 6) can be obtained by non-fault phase C: Equation ( 7) and ( 8) can be obtained by the fault phase A: In Eq. ( 7) and ( 8) ˡ Ӕ is fault point voltage for the A phase to ground.In order to eliminate ˡ Ӕ Eq. ( 9) is obtained by Eq. ( 7) and ( 8): Equation ( 10) and ( 11) can be obtained by fault phase B: In Eq. ( 10) and ( 11), ˡ Ӕ is fault point voltage for the B phase to ground.In order to eliminate ˡ Ӕ Eq. ( 12) is obtained by Eq. ( 10) and ( 11): (12) Z s , Z M , l can be solved by Eq. ( 6), ( 9) and ( 12).
Three-phase fault algorithm principle: When threephase short circuit failure occurs as Fig. 3 shows, the system remains symmetrical, Eq. ( 12) and ( 14) can be obtained from the A phase: In formula ( 13) and ( 14) Z 1 is the positive sequence impedance for the whole line.Its value can be calculated in advance across the electrical quantity before the failure: Superscript pre is representative of the pre-fault electrical quantities.Equation ( 16) can be obtained by Eq. ( 13) minus and ( 14).l can be solved by Eq. ( 16): 1 1 (1 ) Algorithm realization: Ranging equations derived from earlier are composed of three nonlinear equations except three-phase earth fault.It is hard to directly solve the fault distance and usually pseudo roots appear.This study uses the search method for solving equations.Assuming that the l, s initial value, selecting two of the three equations, plugging in l, Z s and Z M can be easily calculated.Since there are three selection methods, then the solved Z s , Z M are three groups, if the value l is correct, then the three groups are the same.
In the double-end unsynchronized case, repeating the search can reduce errors.Selecting the appropriate search step between 0 and 1, calculating Z s , Z M .It finds the appropriate point in which three groups of calculating value difference is minimum.If the result does not meet the requirement, it may reselect the step size until it gets the optimal solution.By this way, it can find the most appropriate point to Cause error of the three groups of Z s and Z M to be smallest.Figure 4 shows the flow chart.The simulation time is set to 0.22 sec and the ground fault occurs at 0.12 sec and it ends at 0.18 sec.PSCAD/EMTDC simulates the failure model, MATLAB processes the data.Figure 6 and 7 shows the voltage and current maps when a phase ground fault occurs at the point of 10 kM from M-terminal with grounding resistance is 10.And Fig. 8 shows the distribution diagram of function F (l).

DISCUSSION
It is can be seen that the ranging accuracy of this algorithm meets actual project's needs.There is the analysis for the factors affecting the accuracy: Transmission line length: From the results of the simulation, when the transmission line length exceeds 200 km, error exceeding 1 km and it does not meet the practical engineering requirement.It is because the algorithm ignores the distributed capacitance in the circuit when it equivalents the transmission line to series line which is consisting of self impedance and coupling impedance, when the line is not long, the distributed capacitance in the whole line may be ignored, the error does not exceed 1 km.But when the transmission distance is too long, distributed capacitance is large, distributed capacitive current time in the proportion of the entire circuit increases, the error will be increased, which cannot meet the project needs.
Filtering effect: A system asymmetrical failure occurs, it will generate a lot of high harmonics and decaying DC component, if filtering is not very good, it will have a huge impact on measuring electrical quantities and result in reduced range accuracy.Filtering effect will directly affect the size of measuring error.This study uses a differential full-wave Fourier algorithm and differential filtering and FIR low-pass filtering may also be used.

Hardware requirements and line conditions:
Since the algorithm is based on GPS synchronous sampling, the hardware conditions are high, but the actual collection samples at both ends, generally there will be time errors, which will affect the ranging results, bringing the error.

CONCLUSION
A large number of simulation results show that this algorithm is not subject to transition resistance, system operating mode, fault type and fault distance influence.In the high-impedance grounding condition, it may still get accurate results.There is a good effect for the distance transmission line which is not too long and it has a high application prospect.

Fig. 5 :
Fig. 5: System simulation modelALGORITHM SIMULATIONIn order to verify the correctness of the algorithm, this study adopts double-ended power supplying about 500 kV EHV transmission line parameters and it creates the EMTDC model with a total length of 100 kM transmission lines by the use of distributed parameter line model, as shown in Fig.5.Line parameters:

Table 2 :
Impacts of grounding resistance and fault location on the results of BCG Failure

Table 3 :
Impacts of grounding resistance and fault location on the results of ABC Failure

Table 4 :
Impacts of the length of transmission line on the resultsFailure