Wind turbine fault detection based on expanded linguistic terms and rules using non-singleton fuzzy logic. (15th March 2020)

• Record Type: Journal Article
• Authors: Qu, Fuming; Liu, Jinhai; Zhu, Hongfei; Zhou, Bowen
• Is Part Of: Applied energy. Volume 262(2020)
• Publication Date: 2020-03-15
• Subjects: Fault detection -- Wind turbine -- SCADA data -- Non-singleton fuzzy inference system -- Expanded linguistic terms and rules; Power (Mechanics) -- Periodicals; Energy conservation -- Periodicals; Energy conversion -- Periodicals; 621.042
• DOI: 10.1016/j.apenergy.2019.114469
• Languages: English
• ISSNs: 0306-2619

Abstract

Highlights: A non-singleton FIS fault detection method is proposed to detect early wind turbine faults. A non-singleton fuzzy input generation method is proposed. A mechanism of expanding linguistic terms and rules in FIS is put forward. Based on the defuzzified result, fault factor is designed to measure fault severities. Real wind turbine SCADA data are used in the experiments of this paper. Abstract: Wind power generation efficiency has been negatively affected by wind turbine (WT) faults, which makes fault detection a very important task in WT maintenance. In fault detection studies, fuzzy inference is a commonly-used method. However, it can hardly detect early faults or measure fault severities due to the singleton input and the limited linguistic terms and rules. To solve this problem, this paper proposes a WT fault detection method based on expanded linguistic terms and rules using non-singleton fuzzy logic. Firstly, a generation method of non-singleton fuzzy input is proposed. Using the generated fuzzy inputs, non-singleton fuzzy inference system (FIS) can be applied in WT fault detection. Secondly, a mechanism of expanding linguistic terms and rules is presented, so that the expanded terms and rules can provide more fault information and help to detect early faults. Thirdly, the consequent of FIS is designed by the expanded consequent terms. The defuzzified result, which is defined as the fault factor, can measure fault severities. Finally, four groups of experiments were conducted using the real WT data collected from a wind farm in northern China. Experiment results show that the proposed method is effective in detecting WT faults.