A Statistical Analysis of the Effects of Mistuning on the Forced Response of Cyclic Assemblies1

A statistical investigation of the effects of random mistuning on the forced response of nearly cyclic assemblies made of mono-coupled, single-mode component systems (blades) is presented. Various probabilistic methods of analysis are applied to generate the statistics of the forced response: an analytical first-order statistical perturbation method, a purely numerical Monte Carlo simulation, and a hybrid approach that combines the first two. Means and variances of the component systems' amplitudes are obtained. The limitation, accuracy, and computer costs of the various techniques are discussed. Also, the effects of the system parameters on the expected value of the largest amplitude experienced by the assemblies at any excitation frequency are investigated. It is found that weakly coupled systems are more sensitive t o mistuning than strongly coupled ones through a greater increase in largest amplitude, and that the accuracy of the various approaches depends strongly on the relative magnitudes of coupling, rnistuning, and damping.