Wind turbine generator failure analysis and fault diagnosis: A review

Abstract

Abstract The large scale deployment of modern wind turbines and the yearly increase of installed capacity have drawn attention to their operation and maintenance issues. The development of highly reliable and low‐maintenance wind turbines is an urgent demand in order to achieve the low‐carbon goals, and the arrival of fault diagnosis provides assurance for its satisfactory operation and maintenance. Numerous statistical studies have pointed out that generator failures are a main cause of wind turbine system downtime. The generator, as one of the core components, converts rotating mechanical energy into electrical energy. However, the generators can hardly operate reliably towards the end of the turbine life owing to the variable‐speed conditions and harsh electromagnetic environments. This article first provides a comprehensive and up‐to‐date review of the electrical and mechanical failures of various parts (stator, rotor, air gap and bearings) of the generator. Then the fault characteristics and diagnostic processes of generators are investigated, and the principles and processes of fault diagnosis are discussed. Finally, the application of four categories of model‐based, signal‐based, knowledge‐based and hybrid approaches to wind turbine generator fault diagnosis is summarized. The comprehensive review shows that the hybrid approach is now the leading and most accurate tool for real‐time fault diagnosis for wind turbine generators. A qualitative and quantitative assessment of algorithm performance using false alarm rates is proposed. The methodology can subsequently be applied to the wind industry.

Keywords

• artificial intelligence
• electric generators
• fault diagnosis
• wind turbines