A model-based measurement procedure for the characterization of frequency-dependent material properties of piezoceramics using a singleton specimen

点击翻译按钮获取中文摘要

The increasingly simulation-based design processes of ultrasonic transducers require the best possible descriptions of the used materials for realistic results.Piezoceramic materials can be modeled using at least 10 parameters.The determination of these parameters is done according to current standards using the evaluation of resonance frequencies of different specimen geometries.However, since these different specimens are always subject to different process and polarization conditions, implausible or inconsistent parameter sets can result when the individual material parameters are combined.The characterisation of a piezoceramic on a single specimen is difficult, since there are usually no sufficiently high sensitivities for all relevant parameters.An increase of the sensitivities and thus a determination of a complete material parameter set was realised with an optimised electrode topology on piezoceramic disks in the previous project.Since the more complex electrode structure can no longer be represented by analytical approximations, the material parameters are determined by an inverse method.In the process, a digital twin is created in the form of a finite element method model, which contains the material parameters as input variables.Problem-adapted optimisation algorithms can then be used to influence the material parameters of the model in such a way that the behaviour of the real system is mirrored as well as possible.However, several impedance measurements on a piezoceramic disc are currently necessary for this.A more dynamic mathematical optimisation of the electrode topology should reduce this to a single measurement.In addition, a causal damping model is to be found which better maps the physical material behaviour and thus results in a more realistic simulation.Again, a reliable solution to the inverse problem must be ensured by sufficiently large sensitivities.