A comparison of GA objective functions for estimating internal properties of piezoelectric transducers used in medical echo-graphic imaging

Wideband piezoelectric transducers are extensively used in several areas such as industrial non-destructive evaluation, medical imaging and robotic vision. These devices are key elements in the performance of many commercial diagnostic systems as employed in medicine and industry. The knowledge of their design and operation parameters is quite important in modeling, simulation and optimization of these commercial transducers, especially when the design and manufacturing process of the probe is unknown. In this paper, we present an implementation of a genetic algorithm (GA) procedure for the estimation of some internal parameters of broadband piezoelectric transducers. The GA implementation process was developed considering the inclusion of different terms in the objective function proposed. Linear and quadratic terms, related with an important characteristic magnitude of the transducer, have been considered and alternatively included in the objective function proposed. The performance of the GA estimation procedure was evaluated for the different analyzed cases, and the implementation was applied to a practical wideband piezoceramic transducer. The estimation results were compared with experimental data based on the emitting transfer function of the practical probe.