Research on equipment health prediction technology based on edge computing and VAE-TCN

Abstract

Abstract The Internet of Things technology is developing rapidly, and the data generated has also exploded. Traditional cloud computing technology can no longer meet the demand for efficient processing of massive data. Edge computing technology can move the amount of calculation down to the edge of the network, which can greatly improve computing efficiency. Applying edge computing to the field of equipment health prediction, the combination of strong responsiveness and computing capabilities of edge computing and high-precision prediction technology makes production operation and maintenance more reliable and efficient. At the same time, a neural network prediction model combining Variational Auto-Encoder (VAE) and Time Convolutional Network (TCN) is proposed to improve the accuracy of equipment health prediction. This model uses VAE for dimensionality reduction, extracts the hidden information in the original data, reconstructs high-quality sample data, and then uses TCN to mine the internal connection between the features and the target in the long sequence information. Compared with five benchmark prediction models on the C-MAPSS dataset, experiments show that the proposed model has higher prediction accuracy.