Debonding monitoring in adhesive layers of wind turbine blades using embedded FBG sensors

Abstract

Recently, adhesive bonding is considered as the excellent technique for the connection of the composite components of wind turbine blades. But, the more use of adhesive bonding method rise, the more frequency of adhesive failure increase. Of particular interest is that the greatest part of wind turbine blades failure is concerned with the adhesive failure, especially adhesive debonding and crack. Some researchers have studied debonding monitoring by using optical fibers, but they focused on the limited applications such as composite repair patches. Additionally, few studies have been reported on the debonding monitoring in thick adhesive layer with lots of voids, especially in wind turbine blades. In this study, FBG sensors embedded in adhesive layer, were used for debonding monitoring of wind turbine blades. However, unexpected split problems can occur, when FBG sensors are embedded in adhesive layer. Three embedding skills, PAC technique, Recoating technique, and Packaging technique, were proposed or applied to overcome unexpected troubles. The signal characteristics of the reflected spectra of FBG sensors for each technique were investigated, and PAC technique and Recoating technique, were considered as the suitable method for the embedding of FBG sensors. Meanwhile, two main experiments, Experiment A and B, were designed to verify the feasibility of FBG sensors for debonding monitoring, and the specimens were fabricated considering the position of sensors and the location of debonding. Three point bending tests, simulating the structural deformation of the specific parts of wind turbine blades, were carried out. PAC technique was applied to all specimens, and artificial debonding using release agent was applied for replicating real debonding. The strain changes of embedded sensors as well as the reflected spectra of FBG sensors were investigated, when the specimens were deformed. The signal characteristics such as peak splits, bandwidth changes, sudden stra...