Hard Disk Fault Detection Method based on Temporal Convolutional Network

The data centre of a hard disk failure is usually irreparable for data center losses, and hard disk failure detection is also the focus of industry and academia research. Traditional methods mainly use machine learning and statistical concepts for hard disk fault detection, but due to differences in sample data and manual feature simulation, the time complexity and accuracy of fault detection are insufficient. Based on this, this paper presents a hard disk fault detection method based on Temporal Convolutional Network, builds a prediction network based on Temporal Convolutional Network, and uses real environment datasets to complete the model optimization. The experimental data show that there are obvious advantages in model size and prediction accuracy.