Abstract:
This talk gives an overview of the seismic wave propagation problem and how it can be solved on distributed memory supercomputers. Earthquakes usually arise when tectonic plates suddenly slip along a fault, for example the Hayward or San Andreas faults in the San Francisco Bay area. The slip along the fault generates ground motion that propagate as seismic waves through the earth. The foundation of buildings and structures is shaken when the waves reach the surface, which can lead to significant damage or destruction. However, the amount of shaking depends on many factors including the proximity to the fault, the material structure of the earth and local soil characteristics. Using high resolution computer simulations on GPU accelerated supercomputers, such as the Summit or Sierra systems, we can now perform detailed studies of wave propagation through the earth at frequencies that are important for building dynamics. The final link in the simulation is to propagate the ground motion to calculate the dynamical motion of individual buildings, for estimating the risk of permanent deformation or failure.
Bio:
Anders Petersson is a computational mathematician in the Numerical Analysis and Simulations group at the Center for Applied Scientific Computing at Lawrence Livermore National Laboratory. Anders earned his Ph.D. in Numerical Analysis from the Royal Institute of Technology in Stockholm, Sweden. Anders does research in Applied Mathematics, specializing in numerical methods for computational wave propagation and their implementation on large parallel computers. He leads the development of the seismic wave propagation code SW4, which currently is supported by the project "High Performance, Multidisciplinary Simulations for Regional Scale Earthquake Hazard and Risk Assessments", within the Exascale Computing Project (ECP). He also studies waves on a much smaller scale in the project “Scalable Quantum Optimal Control”. Here, advanced numerical methods are being developed and implemented on parallel (classical) computers for solving the quantum optimal control and characterization problems for many-qubit quantum systems.