Methodology of Adaptive Prognostics and Health Management Using Streaming Data in Big Data Environment

Abstract

Prognostics and health management (PHM) has gradually become an essential technique to improve the availability and efficiency of industrial systems. With the rapid advancement of sensor technology and communication technology, a huge amount of real-time data is generated from various applications industry, which brings new challenges to PHM in the context of big data streams. On one hand, high-volume stream data places a heavy demand on data storage, communication, and PHM modeling. On the other hand, continuous fluctuation and drift are essential properties of stream data in an online environment, which requires the PHM model to be capable to capture the new formation in stream data adaptively and continuously. This research proposes a systematic methodology to develop an effective online evolving PHM method with adaptive sampling mechanism against continuous stream data. An adaptive sample selection strategy is developed to effectively select the representative samples in both off-line and online environment. Meanwhile, a probabilistic theory-based modeling approach is developed to update the model with newly selected samples. Finally, the whole methodology is validated with real-world industrial cases. The result comparison between the proposed methodology and state-of-art methods verifies the superiority of the proposed method.