Fault Phase Selection And Fault Location Method Of Single-Phase Grounded For Double-Circuit Transmission Lines On The Same Tower

The DCTL(Double Circuit Transmission Lines on the same tower)are widely used due to a series of advantages such as large transmission capacities,low occupancy of land and cost savings.Distance protection has been the main protection of the DCTL due to advantages such as less influence of system operation mode and high sensitivity.The distance between the two circuit lines of the DCTL is relatively close,and there are strong electromagnetic coupling between two circuit lines.The electromagnetic coupling phenomenon makes the electric quantities of the two circuit lines no longer independent.However,the distance protection currently applied to DCTL uses one circuit line as a protection unit.When a fault occurs,distance protection cannot obtain the electrical quantities of the adjacent line.At the same time,in order to ensure the independence of the protection,the protection at the local line is not allowed to introduce the electrical quantities from the adjacent line.So the influence of electromagnetic coupling cannot be eliminated,which leads to the failure to accurately select the fault phases and locate the fault.At present,the above mentioned problems are still unsolved,so it is of great significance to develop methods for accurately selecting fault phases and locating fault in DCTL using only the electrical quantities of single-ended single-circuit line.Traditional fault phases selection methods for DCTL are Model-based methods,and these methods use the proposed complex criteria based on the established DCTL mathematical model to select fault phases.Model-based methods only solve parts of the problems in fault phases selection.At the same time,a large number of thresholds are used in the criteria of model-based methods,which are determined based on engineering experience.In actual projects,these thresholds are difficult to set and are easily affected by line parameters,loads and other factors.Therefore,this thesis proposes a fault phases selection method based on CNN(Convolutional Neural Network),which belongs to the field of Data-driven research.The fault phases selection method based on the CNN does not need to establish any mathematical model or propose any fault criteria.The proposed method can automatically extract the fault features contained in the fault data,and can use the fault features to identify the fault phases of DCTL.First of all,in order to make CNN fully learn the fault features of different fault types,a fault simulation model of DCTL was established in PSCAD to simulate massive fault data in batches.Secondly,the fault data are processed into a format suitable for input to CNN,meanwhile,the processed data are used to train the CNN and optimize the hyper-parameters.Finally,the performance of CNN under the optimal hyper-parameters is tested.The test results show that the CNN-based fault phase selection method has a high accuracy rate,and can only use the electrical quantities of single-ended single-circuit line to accurately identify the fault phases when a fault occurs at any position of the whole line.Compared with the Model-based fault phase selection methods,the performance of CNN-based fault phase selection method is significantly improved.The zero-sequence coupling inductance of adjacent lines can reach 50%-70% of the self inductance.When faults occur in DCTL,especially single phase grounded fault,the zero sequence coupling inductance seriously affects the accuracy rate of fault location.In order to accurately locate the fault when the single phase grounded fault occurs on the DCTL,firstly,this thesis analyzes the zero-sequence decoupling network of DCTL and represents the zero-sequnence current of the adjacent line with the fault location and the electrical quantities that can be obtained by the protection.Then,under the two cases of considering the transition resistance and not considering the transition resistance,the fault location equations based on Fixed Point iteration are derived first,and the fault location equations based on Newton iteration are further improved.Finally,the proposed fault location method is tested by digital simulation.Simulation results show that the proposed fault location method has fast iteration speed and can accurately locate the fault,which is not affected by the working conditions and the transition resistance.At last,a virtual online verification method in the simulation environment is proposed,which can online verify the performance the proposed method.Taking fault location method of single phase grounded as an example,the implementing process of virtual online verification in the simulation environment is described.Firstly,a virtual protection device is built by using a custom module in PSCAD.Then,the fault location method based on Newton iteration proposed in this thesis is realized in the protection device using C language.Finally,virtual online verifying of the proposed method is carried out.The proposed virtual online verification method can be effectively used to verify the fault location method based on Newton iteration in this thesis.