Fault-ranging Research On HVDC Transmission Lines Based On WNN Optimized By Improved SSA

HVDC transmission engineering plays an important role in long-distance transmission engineering because of its large capacity and low cost.However,due to the influence of complex topography and environmental climate,transmission lines become the part with the highest failure rate in HVDC transmission system.After the fault occurs,the line fault must be handled timely and accurately to ensure the stable operation of the power system,and the precise and efficient determination of the specific location of the fault is the premise and important protection to complete the line fault cleaning.To address the problem of large errors in traditional ranging methods,the extraction of feature information and the calculation of distances are optimized by improved complementary ensemble empirical modal decomposition(ICEEMD)and wavelet neural networks(WNN),and an improved sparrow search algorithm was proposed for the problem of complex initial parameters of WNN.Firstly,to improve the accuracy of extracting the feature information of the noisy faulted traveling wave,the correlation of the kernel density function is used to discriminate the authenticity of the decomposed signal,and the true high-frequency component is selected to extract the fault features in the high-frequency component of the faulted traveling wave.Secondly,to address the problem of complex parameter determination of WNN,the sparrow search algorithm was introduced.And to address the problem of insufficient search space in this algorithm,the Torus sequence is incorporated to initialize the population and improve the population diversity;to address the problem of unbalanced convergence speed and accuracy of the algorithm,non-linear logarithmically decreasing inertia weights are introduced to improve the individual position update formula;the optimal value is perturbed by using the random wandering strategy to improve the disadvantage that the original algorithm is prone to fall into local extremes.Finally,the advantages of various improved methods are fused,and the WNN interlayer parameters are optimised to construct a fault ranging model for high-voltage DC transmission lines.The HVDC transmission system model was built on the MATLAB Simulink platform,the information on the fault state characteristics at different fault distances and different transition resistances is obtained by changing the system parameter settings.The proposed method was verified and compared with ACO-Elman and IPSO-WNN ranging models.Simulation results show that the ranging accuracy and ranging speed of this method are better than other ranging models,and its stability is high and it is not affected by transition resistance.The measurement error of this method is verified by the field measurement data,which can provide a theoretical reference for the application of HVDC transmission line fault location.There are 34 figures,16 tables,and 66 references in this paper.