改进VMD和ConViT的小电流接地系统单相故障选线方法OA

Addressing the issue of reduced reliability in single-phase grounding fault line selection in small current grounding system in strong noise interference,an intelligent line selection method based on improved combination of variational mode decomposition(VMD)and ConViT was proposed.Firstly,the polar light optimizer algorithm was used to adaptively optimize the key pa-rameters of VMD(decomposition layer K and penalty factorα),and then the noisy modal compo-nents were screened and reconstructed for effective noise suppression according to sample entro-py.Then,the one-dimensional zero-sequence current signal was converted into two-dimensional image through the Gramian augular field(GASF)to preserve the temporal correlation of the sig-nal and enhance the feature expression ability.On this basis,a ConViT line selection model was constructed,which achieves accurate identification of the faulty line by integrating convolutional local feature extraction with Transformer global dependency modeling.The simulation and field test results show that the proposed method exhibits good anti-interference and robustness under strong noise conditions,with fast iterative convergence speed and a line selection accuracy rate ex-ceeding 95%.Its performance significantly surpassed traditional CNN and Vision Transformer models,making it suitable for fault line selection requirements in actual complex noise scenarios.