Innovation In A Pre-Cloud Seismic Environment #seismic #automation #bigdata #bestpractices #digitize #casestudy #collaboration #metadata
21 May 19
10:15 AM
-
10:45 AM
Tracks:
Technical Trends and Innovation
The Oil & Gas industry is currently in the midst of a digital transformation, moving data, utilities, and applications into various cloud environments. Managing and conditioning seismic data in the cloud will ultimately provide substantial benefit for the industry. However, vast amounts of legacy data are trapped in physical media formats today, and cloud-enabled interpretation platforms are at a relatively early stage of development. These realties provide opportunity for continued innovation and automation, while digital transformation progresses. This paper will present a multi-pronged approach in leveraging the latest hardware and network advancements, as well as parallel workflow design with integrated automation to dramatically improve the cycle time of delivering workstation-ready seismic volumes to end users. The first uplift was captured by simply influencing physical media delivery format. By asking for data to be delivered on a USB device, the tape data dump rate of 8 mb/s was quickly increased to a USB data dump rate of 120 mb/s. The next challenge quickly became apparent, in that the network speed was a limiting factor, and the RGFX blades used for seismic data conditioning were not optimally configured for working with large seismic files. By working with a vendor we were able to configure network and hardware combinations, leveraging a faster Ethernet card, the latest processors, and increased amounts of RAM and useable solid-state scratch space. This innovation provided significant cycle time reduction in the dumping, scanning and conditioning of our seismic products. The next challenge was looking at our linear workflow to create and load end-user products to our interpretation platform. We identified opportunities to workflow steps run in parallel, and we developed macros for both 2D and 3D automated seismic loading (leveraging the improved hardware) to our interpretation platform. Combining the improved network speeds and custom fit-for-purpose hardware, along with loading automation provided astonishing results. Benchmark comparisons of one large multi-product load showed delivery time reduced from ~83 days to ~3 days. As it is likely to take many years to achieve total digital transformation and cloud enablement, there is clear benefit of continuing to innovate and automate our on premises environment.