McLaskey research group

Civil and Environmental Engineering, Cornell University

Earthquake Mechanics

and Nondestructive Testing

Our research is focused on the mechanics of earthquakes, but our work also includes the study of damage processes such as fracture and friction in a more general sense.

We use the analysis of waves and deformation in order to study material behavior such as earthquakes (which are often modeled as shear cracks), friction, impact, and the interaction between fluids and faults and fractures. In order to make good use of the data, we must have a good understanding of the way solid bodies deform in response to stresses, the way elastic waves propagate, the way seismic sources can be represented, the way deformation can be measured, and how to manipulate recorded data (signals) for enhanced understanding.

This work is made possible by sensors. A few of the sensors we have used are shown below:

The Panametrics sensor (left) is sensitive to surface displacements down to about 10 picometers in amplitude and can detect ground motions ranging from about 1 MHz down to less than 100 Hz. The capacitive sensor measures the distance between the sensor probe and a target in a frequency range from mHz (hours) to kHz and has resolution less than 1 micron. When glued to a sample, strain gages (far right) provide a measurement of the distortion or strain that the sample feels averaged along the gage length. Strain is a tensor quantity, so a rosette of strain gages is needed to characterize each component of the tensor.

While the principal strength of my research group is in laboratory experimentation, associated instrumentation, and data analysis, we also utilize numerical models as tools for better understanding the laboratory results. This includes (1) finite element models to solve elastodynamic problems such as elastodynamic Green’s functions, (2) finite difference models to solve simplified pressure diffusion problems, and, (3) with collaborators, more sophisticated models that utilize the boundary integral method to simulate dynamic fault rupture and arrest.

Sometimes model results are directly compared to data acquired during laboratory experiments, to constrain certain physical parameters. In other cases, the models are used to extend the laboratory results to larger and more relevant length scales. The combination of controlled and well-instrumented laboratory experiments and matching numerical models is powerful, and it is my goal that graduate students who leave the McLaskey research group will have familiarity with both the experimental and numerical side of the research.

- Digital signal processing - spectral estimation, Fourier theory, filtering, array processing

- Sensor design and calibration - electro mechanical design of sensors, characterization of recording systems including preamps, static and dynamic calibration techniques.

- Wave propagation in solids - the physics of attenuation, theoretical and numerical calculation of Green's functions

- Seismic sources - their characterization, classification, and physical origins

- The physical and chemical behavior of interfaces and surfaces - friction, fault healing, dynamic fault rupture, multiscale surface roughness

Fluid injection and earthquakes

Fluids interacting with faults and fractures in the Earth are important for many energy applications including geothermal energy, CO2 sequestration, and wastewater disposal associated with oil and gas operations. We have studied how fluids injected directly onto a laboratory fault can trigger earthquakes at both the 0.76 m scale and on 3-m granite samples. We hope to better to understand under what stress conditions fluid-triggered slip will remain aseismic or initiate a dynamic rupture. Consistent with prior modelling studies, we have found that initial stress plays an important role in determining if slip will remain slow, will propagate only under the forcing of the fluid-induced stress changes, or will be triggered by the fluid and “runaway” (propagate unbound) under the background initial stress.

Earthquake Nucleation and Foreshocks

The large sample size we use at Cornell facilitates the study of how an earthquake begins, or “earthquake nucleation”. Understanding this process is important for designing an earthquake early warning system (such as that in development in California), for seismic hazard assessment (so we know what types of faults are capable of initiating earthquakes), and for earthquake forecasting (so we can intelligently interpret any warning signs such as foreshocks).

Earthquake Scaling

Earthquakes are observed at a wide variety of scales, ranging from M 9 events that rupture 100s of kms to tiny microtremors M -7 that rupture less than 1 mm. A specific set of scaling behavior has been observed for different earthquake parameters. We utilize measurements of small laboratory experiments to better understand the scaling of how earthquake ruptures initiate, grow, radiate seismic waves, and arrest.

Fast and Slow Earthquakes

Some faults are known to slip slowly and only radiate weak tremors while others are locked and periodically slip rapidly to produce large earthquakes. In the lab, we observe both fast and slow earthquakes and can study the origins of the differences and the resulting seismic wave radiation.

Aftershocks, Asperities, Creep Fronts, and Earthquake Triggering

Aftershocks are a manifestation of delayed earthquake triggering, and though aftershocks are extremely common, their precise mechanics are still poorly understood. Sometimes an aftershock grows larger than the preceding earthquake, so physics-based predictive models of these processes have important hazard-management implications. Historically, it has not been possible to directly study aftershocks from laboratory experiments, only through clever extrapolation of laboratory-derived constitutive relationships (i.e. Dieterich, 1994). However, with large hybrid samples that utilize a combination of compliant forcing blocks and a shear zone composed of a realistic geologic material, we can observe delayed triggering on a meter-sized laboratory sample.

Teaching and research on vibrations can be found here.

--- Ball Impact and System Calibration Project

Ball impact is used as a calibration source for absolute system calibration. While studying this source, I was able to test the limits of Hertzian impact theory.

.

.