Method using singular value decomposition and whale optimization algorithm to quantitatively detect multiple damages in turbine blades

Renewable energy has increased in recent years with a consequential increase in equipment maintenance. Maintenance costs can be reduced by structural health monitoring techniques especially for wind turbine (WT) blade damages. However, the majority are not suitable for on-line measurements and quantitative detections. A quantitative damage detection method is developed to identify multiple damages in a WT blade under in-service operation conditions. Firstly, singular value decomposition is applied to reveal singular information in the operating deflection shape (ODS), which can be treated as damage locations. Secondly, whale optimization algorithm is utilized for a damage severity decision about the natural frequency database between damage severities and natural frequencies, which are constructed by finite element method (FEM) simulations on the detected damage locations in the WT blade. The procedure is applied to FEM numerical simulations of a single WT blade with two and three damages. By adding a certain noise to the simulation dataset, the robustness of the present method is validated. Furthermore, the laser scanning vibrometer is employed to test the ODS as well as natural frequencies of WT blades to testify the performance of the multiple damage detection method. Results show that the present method is effective for the detection of multi-damage in WT blades with a certain noise robustness.