Real-time Processing for Large-scale Streaming Seismic Data
SEG Postconvention Workshop
8:30 AM - 12 PM, Thu. Sep. 19, 2019
San Antonio Convention Center, room 225C
Organizers: Eileen Martin (Virginia Tech) and Biondo Biondi (Stanford University)
Schedule:
| Time | Speaker, affiliation | Title |
|---|---|---|
| 8:30-8:40 | E. Martin, B. Biondi | opening remarks |
| Land Seismic | ||
| 8:40-8:50 | Gary Binder, Colorado School of Mines | Automated microseismic event detection in DAS data using convolutional neural networks |
| 8:52-9:02 | Eileen Martin, Virginia Tech | Scalable algorithms for ambient noise interferometry |
| 9:04-9:14 | Emad Al-Hemyari, Saudi Aramco | DrillCAM Integrates Wireless Geophones, Rig Sensors and Near-bit Tool and Assists with Real-time drilling decisions |
| 9:16-9:26 | Ted Manning, BP | Nimble nodes pave the way to million channel deployments and the next order of trace density – are we ready? |
| 9:28-9:55 | Moderator: B. Biondi | Discussion and Q&A with all land seismic speakers |
| 9:55-10:15 | coffee break | |
| 10:15-10:45 talk, 10:45-10:55 questions | Matthew Rocklin, NVIDIA | Accelerating Python for Real Time Analytics with GPUs, RAPIDS, Dask, and Numba |
| Marine Seismic | ||
| 10:55-11:05 | Andrew Long, PGS | The digitalization path to cycle time reduction |
| 11:07-11:17 | Alexander Goertz, Octio | Real-time passive monitoring of offshore drilling operations with large-aperture ocean-bottom cables |
| 11:19-11:29 | Dave Nichols, Schlumberger | Computation and decisions: at the edge, in town, and in the cloud |
| 11:31-12 | Moderator: E. Martin | Discussion and Q&A with all marine seismic speakers |
Description: Recent advances in many sensor data acquisition technologies paired with new computational algorithms and hardware are making real-time geophysical processing of large-scale streaming data a reality for some applications. This includes highly-instrumented unconventional and conventional fields, next-generation microseismicity monitoring, in-the-field processing of large-scale marine surveys and wireless node land surveys. In this workshop we will explore methods for real-time processing from data to decision. This includes a variety of scalable streaming algorithms including machine learning methods, adaptive pre-processing, automated noise detection and filtering, the detection of weak signals or small changes in noisy data, and fast methods for data reduction, compression or decompression. Some particular questions of interest include: How can real-time processing make a difference in operations and decisions and what are current bottlenecks? How do data acquisition systems shape real-time processing workflows (quality measures, synchronicity between sensors, batch sizes of data, etc.)? What compute resources are expected to be available in the field, and what are the tradeoffs being made to meet these constraints? How can cloud and edge computing play a role in real-time processing, and can we take advantage of new computing paradigms (e.g. map-reduce, randomized approximation algorithms, etc.)? When should machine learning or AI methods be used in lieu of more traditional computational science methods, and how interpretable or generalizable are the results?
More information about all post-convention workshops can be found here.