Michael V. Bernblit
St. Petersburg Marine Tech. Univ., Lotsmanskaya St., 3, 190008, St. Petersburg, Russia
Statistical energy analysis (SEA) is a powerful tool for vibration and sound radiation prediction at high and medium audio frequencies. The objective of this work (SEA based) is to evaluate acoustic power radiated by a large scale complex structure when its vibration is defined only within limited surface area in the vicinity of broadband sources. A tested structure is divided into a set of energy finite elements (EFEs) and power balance matrix (PBM) equation is solved to obtain the ratio of EFEs rms normal velocity components. Coefficients of the PBM equation derive from transmission, radiation, and internal loss factors. Acoustic power radiated by the whole structure is calculated as the sum of partial acoustic inputs of EFEs. Each power input is determined by radiation efficiency and surface averaged normal velocity of a structural element. This approach enables one to define generalized radiation efficiency of the total structure which links acoustic power and rms vibration of the specific EFE with preliminary measured or given at ``drawing board studies'' vibration. Damping factors provide additional information needed to study the effect of design variables on radiation efficiency of a whole structure. A typical for aerospace and shipbuilding multi-sectional structure comprising cylindrical rib-stiffened shells with transverse partitions separating them was selected to utilize the developed approach. One of cylindrical subsets was driven by a normal broadband point force and its vibration was measured in one-third octave bands and averaged over its surface area. Drops of surface averaged vibration levels between structural elements were computed using PBM equation. Assuming omnidirectional pattern of sound radiation the total acoustic power and far-field sound pressure were calculated. It allowed one to evaluate the total structural radiation efficiency. Radiation loss factors for ribbed cylindrical shells in light (air) and heavy (water) acoustic media were investigated to obtain reliable data for PBM coefficients. Experimental data were in a good agreement with theoretical prediction.