Exploring Vibration-Based Assessment Methods for Wind Turbine Rotor Blades:

Abstract

Blades are essential components in the energy conversion process of wind turbines, and faults in these blades may lead to catastrophic failures. Vibration signals can play a crucial role in the early detection of faults in such systems, providing detailed information across a broad spectrum that can be used to prevent injuries, system damage and associated economic losses. But detection and accurate diagnosis of faults in wind turbines using vibration measurements is inherently problematic because of the multiplicity of sources and ways whereby the phenomena are modulated. In this study a custom-designed permanent magnet wind turbine test rig was developed, containing a small horizontal-axis wind turbine featuring a three bladed rotor, each blade with the NACA 2412 airofoil. Baseline vibration data were collected remotely and analysed to compare with signals indicative of blade faults, facilitating an assessment of blade condition. One of the four blades was seeded with a series of similar cracks and data gathered for each crack separately at 150 rpm, while maintaining a constant load of 100 Ohms. Shaft vibration frequencies served as sensitive indicators of changes in blade conditions. A comprehensive dynamic analysis was conducted using a full 3D finite element method (ANSYS) to simulate fundamental vibration characteristics. Simulated and real-time vibration levels were measured and compared. This method proposed here utilizes a development of the Continuous Wavelet Transform (CWT), termed the Wavelet Transform Feature Intensity Level (CWTFIL), to compute the FIL, enabling a fast and accurate comparison of shaft signatures for healthy and damaged blades.