Automated Wind Turbine Pitch Fault Prognosis using ANFIS

Chen, Bindi; Matthews, P.C.; Tavner, P.J.

Abstract

Many current wind turbine (WT) studies focus on improving their reliability and reducing the cost of energy, particularly when WTs are operated offshore. WT Supervisory Control and Data Acquisition (SCADA) systems contain alarms and signals that provide significant important information. A possible WT fault can be detected through a rigorous analysis of the SCADA data. This paper proposes a new method for analysing WT SCADA data by using Adaptive Neuro-Fuzzy Inference System (ANFIS) with the aim to achieve automated detection of significant pitch faults. Two existing statistical analysis approaches were applied to detect common pitch fault symptoms. Based on the findings, an ANFIS Diagnosis Procedure was proposed and trained. The trained system was then applied in a wind farm containing 26 WTs to show its prognosis ability for pitch faults. The result was compared to a SCADA Alarms approach and the comparison has demonstrated that the ANFIS approach gives prognostic warning of pitch faults ahead of pitch alarms. Finally, a Confusion Matrix analysis was made to show the accuracy of the proposed approach.

Citation

Chen, B., Matthews, P., & Tavner, P. (2013, February). Automated Wind Turbine Pitch Fault Prognosis using ANFIS. Presented at EWEA 2013, Vienna, Austria

• Presentation Conference Type: Conference Paper (published)
• Conference Name: EWEA 2013
• Start Date: Feb 4, 2013
• End Date: Feb 7, 2013
• Publication Date: Feb 5, 2013
• Deposit Date: Jul 10, 2013
• Publicly Available Date: Jul 10, 2013
• Keywords: Wind Turbine, SCADA, Neuro-Fuzzy, ANFIS, Fault Prognosis, Fault Detection.
• Public URL: https://durham-repository.worktribe.com/output/1156753
• Publisher URL: http://www.ewea.org/annual2013/
• Additional Information: 4-7 February 2013