The Research On Cross Line Fault Of Double Loop On The Same Tower Based On ARM

The double loop on the same tower has been widely used in high voltage transmission due to its advantages of land saving,short construction period,strong transmission capacity,and investment saving.However,due to the short distance between adjacent lines of double loop on the same tower,cross line faults are prone to occur,and there are many kinds of faults and more complex fault features.Faults on the same line double loop can be divided into single line faults and cross line faults according to the type of fault.There are a total of 120 types of faults,of which 22 are single line faults and 98 are cross line faults.The occurrence of cross line faults is more harmful to the power grid,and 84 kinds of cross line faults are synonym phase cross line faults.Fast and accurate fault location can effectively reduce the fault repair and improve the reliability of power grid operation.Aiming at fault location of synonym phase cross line faults of double loop on the same tower,this paper proposes a two terminal asynchronous fault location method based on BP neural network algorithm.The core of the fault location mathematical model presented in this paper is calculating the voltage of fault point independently by the two terminal voltages and the line’s transmission equation to get a voltage equation and the expression of fault location.Because the asynchronous problem of data in the two terminal fault location cannot be avoided,the two terminal asynchronous fault location method is adopted in this paper.The asynchronous angle in the two terminal asynchronous fault location is adjusted by using the BP neural network to train the existing sample data.And the variety of cross line fault makes it impossible to do aforementioned calculation by using the voltage and current of two ends directly.Six-sequence-component is used to decouple the mathematical model of double loop on the same tower and calculate the six sequence transform matrix and its inverse matrix.The same sequence positive sequence voltage and current are calculated on the basis of the measured voltage and current of the two ends,and the fault location is solved by substituting them into the ranging equation.For verifying the fault location method’s scientificalness,it was simulated on the electromagnetic transient simulation software EMTP-ATP and MATLAB.The simulation results show that using BP neural network to train the distance measurement model is feasible,it can effectively improve the accuracy of distance measurement;moreover,the more training times,the higher the precision of distance measurement.According to the above principle of distance measurement,taking the STM32F107 development board whose core belongs to Cortex M3 of ARM as a platform and using the off-line training mode of BP neural network,a fault distance measuring device is designed.The device mainly includes functions such as voltage and current detection,data storage and distance measurement result display.In order to guarantee the independence,real-time and sequence of each task,the μC/OS-Ⅱ operating system is transplanted in STM32F107.The experimental data shows that the fault location hardware device designed this time can display the distance measurement results in the form of a table and the online calculation time is about 1 minute,which meets the real-time requirements of the distance measurement device;the measurement accuracy of the device is less than 400 m to meet the requirements of the national grid Regulations and project requirements.