Analysing and Forecasting Degradation in Wind Turbines under Transient Operating Conditions through Vibration Analysis

Abstract

In the field of wind turbines, there is growing attention towards monitoring key components susceptible to high failure rates, such as gearboxes, shafts, bearings, rotor blades, and generators. The use of vibration sensors aids in diagnosing and preventing breakdowns, ensuring reliable and efficient operation. Understanding degradation minimizes costs, optimizes maintenance, and enables accurate prediction and mitigation of failures. This study investigated the vibration signatures of two wind turbines from the same wind farm. Identical sensors were used to capture vibrations over an extended period under various operating conditions. Methods including time domain analysis, frequency domain analysis, order analysis, and envelope analysis provided a comprehensive understanding of the vibration data. Fault frequencies identified through envelope analysis were cross validated with analytical calculations. A unique degradation index was developed to examine degradation over time, revealing greater degradation in the second turbine. Diverse autoregressive models were used to forecast the degradation index for the next 15 days, providing advance notice for predictive maintenance measures.