This paper presents the identification of the local nonlinear effects on the essential dynamics of distributed parameter systems. The system considered is a simple cantilever beam with an attached cubic nonlinear spring at its tip. Nonlinear system identification (NSI) method applied in this work uses numerical simulation results and combines slow-flow dynamic analysis and empirical mode decomposition (EMD) to reconstruct the dynamics in modal coordinates as reducedorder models. The reduced-order models are single-degree-of freedom linear oscillators, which are termed intrinsic modal oscillators (IMOs), with a forcing computed through slow-flow analysis. These forced oscillators are capable of reproducing the modal dynamics, and their forcing amplitudes provide essential information about modal interactions and energy transfer. The proposed NSI method was applied to 3 main cases, corresponding to weakly nonlinear, strongly nonlinear and linear dynamics, respectively. A discrete model of the original system is used to investigate the internal resonances and nonlinearity effects in the original system, by making use of Frequency-Energy plots (FEPs).