Distribution network fault location model based on 1 DCNN-BiLSTM

To address the limitations of current fault location technologies, this paper proposes a distribution network fault location model based on 1DCNN-BiLSTM. 1DCNN effectively extracts local features from fault data, offering unique advantages in processing one-dimensional data. BiLSTM excels at capturing long-term dependencies in sequence data, making it highly effective for time series data. By integrating 1DCNN and BiLSTM, the 1DCNN-BiLSTM model leverages the strengths of both, enabling efficient processing and analysis of distribution network fault signals. Additionally, the model incorporates an attention mechanism to focus more on critical fault features, enhancing the efficiency and accuracy of feature extraction. Furthermore, transfer learning techniques can be applied to quickly adapt the model to new fault scenarios using existing fault data and models, thereby reducing training time and data requirements.