Success of a wind energy project relies on the reliability of a wind turbine system. Poor reliability will directly result in the increase operation and maintenance (O&M) cost and the decrease of the wind turbine system lifetime. To improve the wind turbine system reliability, it is important to identify critical components and characterize failure modes, this will allow the maintenance staff direct their monitoring, and focus on monitoring methods. As the blades are the key elements of a wind turbine system and the cost of the blades can account for 15–20% of the total cost, extensive attention has been given to the condition monitoring of blades. The majority of wind turbine blades are made of fiberglass reinforced with polyester or epoxy resin, the types of damage include debonding, delamination, fibre breakage and matrix cracking. Fatigue damage progression is the main failure factor. To study the contributions of these different types of damages at different fatigue stages, a tool with the ability to detect the damage initiation and to monitor failure process on line is needed. The acoustic emission (AE) testing has become a recognized suitable and effective non-destructive technique to investigate and evaluate failure process in different structural components.
Pattern recognition (PR) techniques have been used for the identifications of failure modes in composites from AE data. The principle is to separate AE signals into a number of clusters (representative of k damage mechanisms), each AE signal can be represented by a vector composed of multiple relevant descriptors. A major challenge in the use of AE technique is to associate each signal to a specific AE source to a damage mechanism with a huge noisy amount of data originating from fatigue loading tests. There are many sources of AE other than the crack of interest. These sources thus constitute noise which can be both random and coherent and often this noise far exceeds the crack signals. Such noise is to be expected in the operating wind turbine blades. An approach of using PR methodology to characterize the damage mechanisms based on the AE data collected from operating turbine blades under realistic loading conditions in service is presented. A laboratory study is reported of fatigue damage growth monitoring in a complete 45.7 m long wind turbine blade, AE data was collected throughout 21 days of cyclic loading which simulated realistic loading conditions in service.