The present paper presents a numerical method for health monitoring and damage identification of civil infrastructure facilities. The detection and characterization of damages is achieved by extracting the electromechanical impedance characteristics of surface bounded self-sensing piezoelectric ceramic (PZT) patches. The frequency – dependent electromechanical admittance signatures of the PZT transducers are compared with the baseline signature of the undamaged structure to characterize the status level of the health of structures. The damage is quantified at first, conventionally by the well-known root-mean-square deviation (RMSD) index and then by using a statistical confidence method in system identification advanced routines of a mathematical computational software. In this paper a set of numerical simulations using finite element analysis is performed on concrete beams instrumented with PZT transducers. The numerical results demonstrate the advantages of using mathematical modelling methods of system identification as compared to previous damage detection methods, which required good engineering judgement and were insufficient for the complexity of modern structures.