Workshop to disseminate research on wave propagation and soil stiffness draws worldwide interest
More than 70 delegates attended a recent two day workshop organised to mark the end of an EPSRC-funded research project on Micromechanics of seismic wave propagation in granular materials.
Researchers from University of Bristol and Imperial College London have been jointly working on the project since 2009 and the Wave Propagation and Soil Stiffness workshop held in Bristol last month was aimed at sharing the results.
The first day of the event focused on experimental studies, while the presentations delivered on the second day considered theoretical and numerical studies. The workshop was jointly supported by TC101 and TC105 of the International Society of Soil Mechanics and Foundation Engineering, the British Geotechnical Association and the Institution of Civil Engineers.
With one exception, all of the presenters in the experimental session considered aspects of the used of bender element tests (see box) to determine the small strain stiffness of soil, which is also sometimes called the elastic stiffness.
Session chair City University, Hong Kong, professor Matthew Coop highlighted that the difficulty in interpreting bender element test data has been known for a decade and that the results can be dependent on the particular interpretation approach adopted.
University of Sydney professor David Airey urged use of a cross-correlation, frequency domain approach to interpret the test data, while other speakers advocated use of time-domain methods such as peak to peak. A consensus emerged that there a 5 to 10% level of uncertainty in estimates of stiffness obtained using bender element tests. Coop added that this is not news, and queried whether the bender element user community is stuck in a “ground hog day”.
Event organisers said that the session did reveal significant novelty and progression, rather than the lack of a step change in analysis.
University of Tokyo professor Reiko Kuwano reported on a survey by the Japanese Geotechnical Society that found bender element use is now becoming more common in industry, as well as in research. This wider uptake has led to the JGS developing a standard for bender element testing and interpretation, according to Kuwano.
The uptake of bender element testing in industry was also evident from the presentation of Russell Geotechnical Innovations’ Chris Russell who discussed use of the technique from a UK industrial perspective.
Georgia Institute of Technology (Atlanta, US) professor Carlos Sanatmarina gave a range of examples of the application of bender elements testing to study soil freezing, liquefaction, soil skeleton genesis.
Antonio Viana da Fonseca of the University of Porto, Portugal gave a good example of the recent advancements in the testing technology by outlining how accelerometers can be embedded within samples to trace the progression of the shear wave.
Kuwano also reported on the development of new disc-shaped transducers that can transmit both compression and shear waves across a relatively large surface to minimise the potential local sample disturbances. This approach also overcomes the problem of bender element tests being dependant on the number of particles the bender touches; an issue that provoked some debated in the discussion periods.
University of Bristol’s Simon Hamlin demonstrated the UK developments by describing three-dimensional tests undertaken in the complex cubical cell at the university. Hamlin also discussed the use of lasers to measure the velocity of the benders themselves as well as the application of dynamic analysis to infer the real motion of a soil embeded bender, given a particular input voltage.
There was some discussion on the frequency content of the received signal, with University of Hong Kong professor Jun Yang showing that there is an upper limit to the frequency that can be transmitted through a sample and this depends on particle size. ESPCI Paris Tech professor Xiaoping Jia discussed the propagation of high frequency compression waves in dense granular systems.
The second day was chaired by the University of Dundee’s professor David Muir Wood. The session was opened by University of Twente professor Stefan Luding who drew on his training as a physicist, and presented particle scale simulations of plane compression and shear wave propagation, and showed how detailed analysis of the system in the frequency domain can reveal interesting phenomena associated with wave propagation in granular materials including frequency filtering and trapping of high frequency response close to the wave source.
Oregon State University professor Matt Evans, Buro Happold’s John O’Donovan, UCL’s Helen Cheng and University of Cambridge’s Yixiang Xu, all followed Luding’s lead by discussing the use of the discrete element method to simulate wave propagation and to study the particle-scale mechanics. These numerical analysts seemed to agree with the experimentalists in concluding that there is no agreement amongst the various available methods for bender element test interpretation.
Xu and Evans supported Kuwano’s idea that disc-shaped transmitters and receivers are preferable to point sources such as bender elements. O’Donovan showed that DEM simulation data can give good agreement with laboratory test data.
However, the session did not only focus on dynamic applications and University of Twente professor Vanessa Magnanimo and University of Birmingham’s Colin Thornton both showed the insight that can be gained by simulating static probes using DEM. George Marketos showed that particle-scale studies need not use DEM; and the stiffness matrix approach has the advantage of revealing all the vibration modes in a model soil.
This theoretical session included application examples, with Evans outlining the application of wave propagation to look at cemented sand, while Cheng and Xu considered the relationship between shear wave velocity and liquefaction resistance and Itasca’s Sacha Emam demonstrating how particle scale modelling and wave propagation can be used in rock mechanics analyses associated with mining.
During the final discussion delegates were divided as to whether the sector is now in a better position. Santamarina reminded the audience that sample disturbance effects often introduce an error in laboratory testing that significantly exceeds the error associated with bender element testing. Muir Wood also argued that a bad interpretation of a bad test can never be a good thing. UPC’s Marcos Arroyo, who was the most vocal participant in the discussion periods, argued provocatively that pollution of a bender element signal is inevitable.
Edited versions of the speaker’s slides are available on the BGA website: www.britishgeotech.org.uk
Bender element tests
Bender elements are piezo ceramic strips that can move when a voltage is applied to them, and conversely, that generate a voltage when moved. By inserting receiver and transmitter bender elements at opposite sides of a soil sample, shear and compression waves can be transmitted at one point and the arrival recorded at another point. If the distance between the two elements is known, then the wave velocity and, therefore, stiffness of the material can be determined providing that some continuum media assumptions are made.