On the use of high-frequency SCADA data for improved wind turbine performance monitoring

E Gonzalez,D Infield,B Stephen,J J Melero

doi:10.1088/1742-6596/926/1/012009

Abstract

SCADA-based condition monitoring of wind turbines facilitates the move from costly corrective repairs towards more proactive maintenance strategies. In this work, we advocate the use of high-frequency SCADA data and quantile regression to build a cost effective performance monitoring tool. The benefits of the approach are demonstrated through the comparison between state-of-the-art deterministic power curve modelling techniques and the suggested probabilistic model. Detection capabilities are compared for low and high-frequency SCADA data, providing evidence for monitoring at higher resolutions. Operational data from healthy and faulty turbines are used to provide a practical example of usage with the proposed tool, effectively achieving the detection of an incipient gearbox malfunction at a time horizon of more than one month prior to the actual occurrence of the failure.

Highlights

High wind farm (WF) operation and maintenance (O&M) costs are a major concern today
wind turbine (WT) condition monitoring based on the use of data from the supervisory control and data acquisition (SCADA) system [5] is increasingly seen as a cost-effective and promising approach, as SCADA data is available at no additional cost
This paper introduces the use of Quantile Regression Forests [23, 24] and high-frequency SCADA data for modelling WT normal performance together with its related uncertainty, for the purpose of performance monitoring

Summary

Introduction

High wind farm (WF) operation and maintenance (O&M) costs are a major concern today. The maintenance requirements due to wind turbine (WT) failures, together with poor site accessibility, can lead to rates up to 30% of the levelised cost of energy (LCOE) of an offshore WF [1]. Detection of failures is desired to reduce O&M costs while improving reliability. To this end, the use of condition monitoring systems (CMS) is required. WT condition monitoring based on the use of data from the supervisory control and data acquisition (SCADA) system [5] is increasingly seen as a cost-effective and promising approach, as SCADA data is available at no additional cost. Within this approach, monitoring WT performance is fundamental, since it is undoubtedly the main feature characterising WT overall operation. Performance monitoring can play a useful role in global condition monitoring

Methods

Results

Conclusion