Random decrement based method for modal parameter identification of a dynamic system using acceleration responses

This paper proposes a random decrement (RD) based method to identify the modal parameters of a linear dynamic system, when only forced acceleration responses are available for use. The RD based method evaluates response data in the time and, then, frequency domains by correlating the RD technique with fast Fourier transform based algorithms and a linear least square procedure. However, the effect of the RD signatures of input forces that is added in the RD signatures of the acceleration responses must be addressed. In response to this need, a novel frequency response function (FRF) approach for the acceleration responses, with a modification of the traditional one, is verified as a first step. Subsequently, the derivation of the RD based method follows the concept of the FRF approach to achieve the theoretical basis of the FRF defined for the RD signatures. Numerical simulations of a 2DOF linear dynamic system loaded by white noise forces and simulated wind forces foster applicability of the RD based method. Applications of the Ibrahim time domain algorithm and the vector RD technique to the forced acceleration responses are also discussed.