Semilinear Random Vibrations in Discrete Systems

Random vibration problems are discussed for weakly nonlinear, multidegree-of-freedom discrete systems subjected to zero-mean, stationary random excitation. The semilinear solution technique developed involves substituting an optimum linear set of equations of motion for the actual nonlinear equations of motion. Parameters of this optimum linear system are selected on the basis of the system output so that a cyclic solution occurs. The cycles require parameter selection and response analysis until a convergence occurs in the sense that the answers from cycle to cycle are similar.