An application of support vector regression for impact load estimation using fiber bragg grating sensors

Low velocity impacts on composite plates often create subsurface damage that is difficult to diagnose. Fiber Bragg grating (FBG) sensors can be used to detect subsurface damage in composite laminates due to low velocity impact. This paper focuses on the prediction of impact loading in composite structures as a function of time using a support vector regression approach. A time delay embedding feature extraction scheme is used since it can characterize the dynamics of the impact using the sensor signals. The novelty of this approach is that it can be applied on complex geometries and does not require a dense array of sensors to reconstruct the load profile at the point of impact. The efficacy of the algorithm has been demonstrated through simulation results on composite plates and wing structures. Trained using impact data at four locations with three different energies, the constructed framework is able to predict the force-time history at an unknown impact location to within 12 percent for a composite plate and to within 10 percent for a composite wing when the impact was within the sensor network region. Experimental validation is also presented on carbon fiber reinforced polymer wings showing low prediction errors even with small training sets.