Analysis of SCADA data for early fault detection, with application to the maintenance management of wind turbines

• Published in: Renewable energy Vol. 115; pp. 521 - 532
• Main Authors: Bangalore, P., Patriksson, M.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2018
• Subjects: Annan elektroteknik och elektronik; Artificial neural network (ANN); assets; case studies; Condition monitoring system (CMS); environmental factors; income; Maintenance scheduling; mathematical models; mathematical optimization model; Other Electrical Engineering, Electronic Engineering, Information Engineering; renewable energy sources; Supervisory control and data acquisition (SCADA); Wind turbine; wind turbines; Artificial neural network (ANN); Condition monitoring system (CMS); Maintenance scheduling; mathematical optimization model; Wind turbine; Supervisory control and data acquisition (SCADA)
• ISSN: 0960-1481; 1879-0682; 1879-0682
• DOI: 10.1016/j.renene.2017.08.073

Wind turbines are, generally, placed at remote locations and are subject to harsh environmental conditions throughout their lifetimes. Consequently, major failures in wind turbines are expensive to repair and cause losses of revenue due to long down times. Asset management using optimal maintenance strategies can aid in improving the reliability and the availability of wind turbines, thereby making them more competitive. Various mathematical optimization models for maintenance scheduling have been developed for application with wind turbines. Typically, these models provide either an age based or a condition based preventive maintenance schedule. This paper proposes a wind turbine maintenance management framework which utilizes operation and maintenance data from different sources to combine the benefits of age based and condition based maintenance scheduling. A mathematical model called Preventive Maintenance Scheduling Problem with Interval Costs (PMSPIC) is presented with modification for the maintenance optimization considering both age based and condition based failure rate models. The application of the maintenance management framework is demonstrated with case studies which illustrate the advantage of the proposed approach. •A framework which provides tools for utilization of data from various sources for optimal maintenance strategy is presented.•An optimization model is presented and compared to a frequently used simple model and the implications are discussed.•The optimization model is modified for a hybrid maintenance plan, considering age-based and condition-based strategies.•A novel approach for utilizing the proportional hazards model is presented for condition based maintenance scheduling.