Testing was performed on a 34 meter blade at a facility in DTU RisÀ Campus, featuring both wired accelerometers and lowpower MEMsbased wireless accelerometers. Testing was focused on an induced delamination area on the trailing edge of the blade, which was subject to various configurations in order to simulate different degrees of damage. Excitation was performed in two ways: near the delamination zone in a simulation of operational wind excitations, and with a bar designed to excite torsional modes of the wind turbine blade. We compare the data collected from the wireless sensors against wired sensors to demonstrate their performance. We explore methods for determining damage. We first explore results of autoregressive coefficients for indicating damage levels. Finally, we demonstrate the use of damage sensitive features from the wavelet transforms of input and output signals to provide a method suitable for nonstationary blade excitations.