An online real-time health condition monitoring method for industrial equipment based on OS-ELM

Abstract

In recent years, equipment health condition monitoring has an increasingly important impact in the industrial environment, which directly reflects the reliability and availability of the equipment. However, most of the commonly used deep learning algorithms need to be trained after the collection of data is completed, which cannot be trained in real-time and adjust the network structure online; on the other hand, the complex network structure requires a feedback session for parameter learning, which further increases the computation volume and computation time. Therefore, this paper proposes an online real-time health condition monitoring method for equipment based on OS-ELM, which dynamically adjusts the network parameters for the batch increase of data and realizes the real-time calculation of equipment health condition values. In the experimental stage, on the one hand, the feasibility and effectiveness of this method is verified by simulating the scenario that data are acquired batch by batch as time changes through real data sets, and on the other hand, the superiority of this method is verified by comparing it with other commonly used machine learning algorithms.