RESUMEN
When a piezoelectric element is inserted in a medium, the viscoelastic properties of this medium can be deduced from the measurement of the electrical impedance of this insert, by means of optimization techniques, if all the concerned interfaces exhibit a strict parallelism. However, in practical cases, this assumption is not always respected. We have developed a model to evaluate the consequences of such a parallelism defect. Comparisons between experimental results and simulations have been carried out to check this theoretical approach. For different parallelism defects, the validity of the acoustical parameters deduced from an optimization method have been tested using criteria based on the Kramers-Kronig relations. The maximum parallelism misfit that can be tolerated for a correct application of this technique has also been evaluated.