DC power distribution network fault positioning method based on detection signal and cosine similarity

The invention discloses a DC power distribution network fault positioning method based on detection signals and cosine similarity, and the method comprises the following steps: determining an injection strategy of the detection signals according to a topological structure and equipment parameters of a double-end modular multilevel converter (MMC) type medium-voltage flexible DC power distribution network; collecting positive and negative electrode voltage signals at the outlet of the MMC, carrying out fault identification, and judging a fault electrode; zero-mode current of the head end of each line is calculated, and a band-pass filter (BPF) is adopted to extract a detection signal; and in combination with a fault line selection criterion, completing fault line selection based on cosine similarity. The method has higher accuracy under the condition that the power distribution network is complex in structure and large in transition resistance, and has obvious advantages in the aspects of positioning time, met