Bender element transducers are used to determine the small-strain shear stiffness, G0, of soil, by determining the velocity of propagation of mechanical waves through tested samples. They are normally used in the laboratory, on their own or incorporated in geotechnical equipment such as triaxial cells or oedometers.
Different excitation signals and interpretation methods are presently applied, each producing different results. The initial assumptions of unbounded wave propagation, generally used in bender element testing and inherited from seismic cross-hole testing, are quite crude and do not account for specific boundary conditions, which might explain the lack of reliability of the results.
The main objective of this study is to establish the influence of the sample and transducer geometry in the behaviour of a typical bender element test system.
Laboratory and numerical tests, supported by a theoretical analytical study, are conducted and the results presented in order to achieve this goal.
An independent monitoring of the dynamic behaviour of the bender elements and samples is also carried out. Using a laser velocimeter, capable of recording the motion of the subjects without interference, their dynamic responses can be obtained and their mechanical properties verified.
A parametric study dealing with sample geometry is presented, where 24 samples with different geometries are tested. Synthetic rubber is used as a substitute for soft clay, due to the great number of samples involved and the necessity of guarantee the constancy of their properties.
The numerical analysis makes use of three-dimensional finite difference models with different geometries. A regressive analysis is possible since the elastic properties of the system are pre-determined and used to evaluate the results. A numerical analysis also has the benefit of providing the response not only at a single receiving point but at any node in the model.
Partilha