A Novel Pix2Pix And Yolo V3 Enabled Traveling Wave-based Fault Location Method

With the continuous growth of the social economy and the continuous improvement of the industrial level,the demand for electricity from all walks of life is also increasing.As a carrier for the transmission of electric energy,only the safe and stable operation of transmission lines can ensure the reliability of the power system.Due to the complex environment in which the transmission line is located,faults on the transmission line also occur from time to time.Accurate fault location technology can effectively reduce the line inspection time of power grid staff,improve the efficiency of fault maintenance,and reduce the economic loss caused by faults.The existing ranging technology mainly relies on the wave recording device,and the ranging accuracy is limited by the sampling rate of the wave recording device.The recording device has the problem that the ranging accuracy can only be guaranteed at a high sampling rate,and the technology of increasing the sampling rate from the algorithm software level has the prospect of practical engineering applications.In order to solve the problem that the traditional traveling wave fault location technology is limited by the low sampling rate of the measuring device and the location accuracy is not relatively high,the main research work and results of this paper are as follows:Firstly,this paper studies the fault location method based on traveling waves.The principle of the double-ended traveling wave fault location technology is to determine the fault point by the time difference between the traveling wave head reaching the measuring device at both ends of the transmission line,the length of the transmission line and the propagation speed of the traveling wave in the transmission line.Through the discussion of the problem that the accuracy of fault location is too dependent on the accuracy of the measuring device,a feasible idea to solve the dependence of traveling wave location on hardware accuracy is proposed-adopting D-type traveling wave fault location technology and deep learning algorithm.Secondly,this paper proposes a Pix2Pix-based method to improve the accuracy of traveling wave fault location of transmission line,and deep learning algorithm is used for fault location in transmission lines.The idea of ??using data to image is proposed to solve the problem of unsatisfactory effect of time series data in deep learning algorithms.The low-sampling rate recording data is converted into a scale 1 detail component partial enlarged image containing the traveling wave head through the multi-scale analysis of wavelet transform,and the image is used as the conditional input of Pix2 Pix,and the Pix2 Pix network learns to estimate the characteristics of the potential probability distribution of the data,it can output a locally enlarged picture with an effect similar to the scale 1 detail component images containing the traveling wave head obtained from the recording device with the high sampling rate.By generating a data conversion image which is similar to a high-sampling frequency recording device,the sampling frequency of the recording device can be improved through this algorithm,and the effect of improving the accuracy of traveling wave fault location can be achieved.The Pix2Pix-based method for improving the accuracy of traveling wave fault location in transmission lines can thoroughly tap the potential of existing wave recording devices,and can effectively save manpower and material resources for installing recording devices on transmission lines,and has good economic efficiency.Next,this paper proposes a YOLO v3-based method for evaluating the quality of traveling wave faults images of transmission line.For the images generated by the Pix2 Pix network,there is currently no standard image quality assessment algorithm,and manual image quality assessment is time-consuming,labor-intensive and easily affected by subjective factors.This paper uses YOLO v3,which has accurate and fast target recognition characteristics,and uses its predicted value for target detection as an indicator to evaluate the image generated by the Pix2 Pix network.It successfully compensates the problem that the quality of the images generated by the Pix2 Pix network cannot be assessed qualitatively and objectively.The problem of local assessment.The generated images are evaluated by YOLO v3 algorithm,which can effectively improve the effectiveness and reliability of the Pix2 Pix algorithm for improving the accuracy of transmission line fault location.Finally,this paper proposes a traveling wave fault location method based on Pix2 Pix and YOLO v3 algorithms,and validates the proposed method by simulation experiments.The transmission line fault model is modeled and simulated by Simulink software.The fault location results are obtained by changing the location of the fault,the type of fault and other conditions under the conditions of different sampling frequency.The results are compared with the traditional single-ended fault location method and double-ended fault location method.The results verify that the method proposed in this paper performs well,it is proved in the method can get rid of the hardware dependence of the traditional double-ended traveling wave method,its positioning accuracy has been improved,and it is very feasible.