Intelligent condition assessment of industry machinery using multiple type of signal from monitoring system. (January 2020)

• Record Type: Journal Article
• Authors: Hu, Aijun; Bai, Zerui; Lin, Jianfeng; Xiang, Ling
• Is Part Of: Measurement. Volume 149(2020)
• Publication Date: 2020-01
• Subjects: Industrial machinery -- Monitoring system -- Condition assessment -- Correlation analysis -- Hidden Markov Model; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2019.107018
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: A new strategy of health assessment is presented based on multiple parameters. The industrial big data cleaning and the correlation analysis were performed. The Hidden Markov Model (HMM) is used to process data. The health states of machine are evaluated effectively. Abstract: Real time condition assessment for machinery is used for avoiding catastrophic failures. A new strategy which combined data processing with data-driven method is presented for condition assessment of machinery based on multiple characteristic parameters of industrial equipment. Firstly, the data processing is carried out, including the industrial data cleaning, the correlation analysis using the Bin method and the condition division. The vibration parameters, which are sensitive to the state changes of the machine, are assumed as data binning reference. Secondly, the multi-parameter condition evaluation technique is proposed by using Hidden Markov Model. The industrial big data collected from monitoring system are analyzed and the site test is conducted finally. The results show that the provided technique can not only evaluate the running condition of the machinery, but also reflect the change of the operational condition. It can exhibit a potential capability in tracing further deterioration of the machine.