An online ultrasonic flowmeter fault diagnosis method based on an extended belief rule base

• Jinyuan Li,
• Fan Yang,
• Wei He,
• Hailong Zhu,

Vol. 19, No. 1, January 31, 2025
10.3837/tiis.2025.01.006, Download Paper (Free):

• Ultrasonic flowmeters
• extended belief rule base
• Fault diagnosis
• online updating
• Dynamic Environment

Abstract

Ultrasonic flowmeters play a vital role in industrial processes, where their stability and accuracy are considered crucial for ensuring productivity and safety. However, many current fault diagnosis models operate offline, which results in them not adapting well to changes in the environment. To address this challenge, an online fault diagnosis model based on the extended belief rule base (EBRB) is proposed in this paper to meet the immediate need for real-time and dynamic fault diagnosis in industrial environments. The approach begins by constructing an initial fault diagnosis model that is capable of processing both qualitative and quantitative information, utilizing historical data. The domain-value-based rule update method is subsequently proposed alongside a novel online parameter optimization strategy, which is applied when new observations are made in a dynamic environment. This newly acquired data is transformed into fresh rules via the parameter optimization method, and these rules are incorporated into the initial model via the domain value-based rule update method. The proposed model in this paper was finally tested via two publicly available UCI ultrasonic flowmeter datasets. The results show that the model proposed in this paper is able to flexibly cope with the transformation of fault types in dynamic environments and achieves an accuracy of 97%.

Statistics

Cite this article