Grafting log analysis method of power data based on multi-layer clustering

Power system generates a large amount of log data. Extracting valuable log information can provide effective and reliable basis for event decision-making and operation management. This paper proposes a log analysis method for power data grafting based on multi-layer clustering, which reduces the log analysis complexity by reducing the log dimension, and improves the clustering effect and the accuracy of log analysis by combining K-means clustering and deep clustering algorithms. According to the real and effective log data, the model is verified, and the experimental results show that the accuracy of log analysis is 87.64%, which is better than the existing log analysis methods.