- Home
- Conferences
- Conference Proceedings
- Conferences
First EAGE Digitalization Conference and Exhibition
- Conference date: November 30 - December 3, 2020
- Location: Vienna, Austria
- Published: 30 November 2020
81 - 93 of 93 results
-
-
Tight Integration of Decision Forests into Geostatistical Modelling
By C. DalySummaryThis presentation will develop a novel approach to geostatistical modelling based on Conditional Random Fields. Instead of starting with a full multigaussian prior, the conditional distribution are estimated at each target location where results are required. This step does not require the specification of a full Random Function. The conditional distributions are estimated with a Decision Forest approach which is known to converge to the true conditional distribution in quite general conditions. A method for simulating realizations by correlated sampling from the distributions is shown. The advantages of this new approach are that it provides good quality estimates of uncertainty, allows the use of many secondary variables and does not require strong models of stationarity either for the target variable or for the relationship between target and secondary variables. The results of the method are compared with the classic algorithm.
-
-
-
Leveraging Scalable Cloud Computing to Facilitate Collaborative, Large-Volume Pre-Stack Data Analysis: A Frontier Basin Example
Authors D. Goulding, W. Shea, S. Cutten, S. Perrier and O. MichotSummaryAnalysis and interpretation of pre-stack seismic data has historically been limited by the compute capacity and storage required to visualize, process, and analyze large volumes of pre-stack seismic traces in a typical 3D seismic survey. Over the past decade, Sharp Reflections has developed a solution to this problem by leveraging advancements in parallelized computing and memory sharing capabilities in High Performance Computing (HPC) environments to produce a software package (Pre-Stack Pro) designed to view, process, and interpret full fidelity pre-stack seismic data volumes, in memory, in real-time.
In recent years, Sharp Reflections has adapted Pre-Stack Pro to run in public compute clouds, benefits of which have been multi-fold; access to required software and hardware resources from any location at any time; scalability options allowing customization of cluster sizes for the survey size at hand; and increased opportunity for collaboration across geophysical disciplines and geographically distant locations.
This presentation describes how OMV New Zealand and Sharp Reflections exploited the new cloud digital building block to develop a new full-fidelity workflow for pre-stack data analysis, QC, improvement and interpretation using a remote compute solution. The subject data set was a large frontier exploration survey from the Great South Basin in New Zealand.
-
-
-
Digitalization through Automated Prospect Ranking Evaluation of NFED to Improve Business Decision Process
Authors N.A. Rilian, H. Rasid, K.A. Zamri, E.B. Keong and A. JazimSummaryNear Field Exploration and Development (NFED) is a part of the ongoing effort to ensure the successful achievement of PETRONAS Upstream 2030 Blueprint, where the intent is to provide focus on monetizing prospects and leads (2U: Undiscovered Resources) within 20 km radius from existing hubs across offshore Malaysia.
Currently, there are no establish processes that provide a comprehensive review into the viability of monetization of exploration candidates within 20 km radius, meanwhile there are circa thousand prospects and leads included discovered resources (2C) lied across offshore Malaysia and all those data have been scattered randomly. The challenges for NFED team is how to manage more than 600 2U within 20 km and 400 2C for both oil and gas molecules that has been registered, there is no standardize and integrated screening evaluation tools to deal with currently.
To addresses these challenges, new application tool as part of digitalization effort should be developed and established to deal with big information data and standardize screening and evaluation process workflow by accommodating and catering the diversity and uniqueness of each resources. The new application tool shall be functioned as Dashboard, Database and have the ability for data processing capability.
-
-
-
Learning the Matrix-Fractures Transfer Rate Using a Convolutional Neural Network
Authors N. Andrianov and H.M. NickSummaryOne of the key elements in constructing of representative dual porosity/dual permeability models is to provide the mass transfer rate between the matrix and the fractures. Whereas it is possible to compute numerically this transfer rate for specific geometries, it is challenging to estimate the transfer function without running the CPU intensive computations. In this work, we demonstrate that a convolutional neural network can approximate a transfer function using the encoded fracture geometry and the precomputed fine-scale simulation results.
-
-
-
Enhancing Fault Interpretation Efficiency and Accuracy with Deep Convolutional Neural Network and Elastic Cloud Compute
By S. ManralSummarySeismic Interpretation, Deep Convolution Neural Network , cloud compute, digital
-
-
-
Automated Well Portfolio Optimization – Leveraging Digital Technologies to Accelerate Well Intervention
Authors R. Holy, P. Songchitruksa, R. Sinha, K. Vadivel, S. Ramachandran and R. GaskariSummaryWell Portfolio Optimization
-
-
-
Artificial Intelligence in Early Stage Exploration at Wintershall Dea
More LessSummarySuccessful exploration relies on quickly managing uncertainties and opportunities. Examples include bidding rounds or farm-in offers. Explorationists must quickly develop an understanding of the available data and the implications on value of the assets. Under strong time pressure to develop reliable evaluations, any technology to speed up a comprehensive overview of data is beneficial. A cognitive exploration advisor tool that intuitively provides an overview of subsurface analogues to new projects will reduce cycle times and improve effectiveness. By ingesting internal and external sources, the tool enables research with lower uncertainties in a shorter time. In 2019, Wintershall Dea performed a pilot to develop a cognitive tool to support searching unstructured data. The pilot concluded with promising results: in less than 3 months, we created a minimum viable product that could search for key words and concepts and identify these in documents used for training the cognitive engine. The tool also extracted tables and images. Search results were presented in a GIS interface with queries geographically constrained by user-defined polygons. Users testing the system experienced increasingly effective and timely searches and considered this to be quite helpful in their daily work.
-
-
-
Scaling Well Log Interpretation for Faster Results with AI
Authors J. Fowler and J. StrobelSummaryWe present a case study of applying machine learning for well log interpretation. The project started with a pilot phase using a selection of 30 wells, then expanding to 126 wells. The developed workflow empowered by machine learning provided excellent interpretation results with higher than expected quality and significant reduction in turnaround time. The workflow is cheaper, faster, unbiased (being data-driven), and able to capture uncertainty – overall, able to produce a higher quality outcome. We project cost reductions of more than 40% compared to conventional workflows, making a good business case to implement this technology routinely. The technology will now be extended to further field studies, allowing further understanding of the technology and a growing financial benefit from the advantages it delivers today.
-
-
-
Seamless Translation of Modern File Formats to SEG-Y through the File System Interface
Authors L. C. Villa Real and M. De BayserSummarySEG-Y is the de-facto exchange and archiving data file format used by the oil and gas industries. Sadly, it has failed to keep up with improvements made in the last decades in storage technologies, which includes compression algorithms for scientific data, parallel data storage and retrieval, spatial indexing, and portable metadata interfaces. Although replacements for SEG-Y do exist, the majority of geophysical software continues to rely on that file format. This paper presents an alternative self-describing container format for geophysical data and a translation layer based on the file system interface which enables unmodified legacy programs to benefit from this new format.
-
-
-
Combining Stratigraphic Forward Modeling with Multi-Point Statistics
Authors A. Miller and J. PeiskerSummaryThe combination of stratigraphic forward modeling (SFM) with multi-point statistics enables the user to generate more dynamically diverse models, while maintaining to a geological concept and its corresponding depositional patterns that conforms to observed data. In comparison with classical geostatistics, this helps to achieve a more robust history match. While classical constraints are always tailor-made for every reservoir, the generated TI can now be recycled for other reservoirs with similar depositional environments.
Although this workflow showed satisfactory results, a more effective process needs to be developed. Automatization of SFM-generation by matching the well data with an objective function based differential evolution is one of the key elements to reduce the time consuming step in the process.
-
-
-
How to Use Automation to Improve Log Quality Control Process for Multiple Wells
Authors A. Fraser, H. Beurdouche, B. Rosvoll Bøklepp and C. FraserSummaryThe first ever well-log was recorded by Schlumberger, in 1927, 93 years ago, and since then a large number and variety of well-logs have been acquired, processed and stored to characterize various properties of rocks and fluids in a formation. The right characterization of these properties is essential to make good commercial decisions to develop these hydrocarbon reservoirs. Due to challenges related to complex formations, borehole environments, data acquired by different service companies, with potentially different technologies, calibrations or vintage, the first step is to quality control (QC) these logs. A key challenge, often encountered, is working with well-log data acquired over many decades which can be difficult to use straight away due to different legacy data formats, missing or inconsistent units and/or properties associated with well-logs, and multiple repeat or duplicated logs present across the same interval. These challenges make manual quality analysis and processing tedious, a problem which is further exaggerated in a high volume multi well setting. In such a setting, detecting log types, unit issues, best data in repeat sections can be cumbersome. A robust workflow to select logs for further interpretation and analysis workflows must be put in place.
-
-
-
Blockchain Applied to the E&P Phase in the Oil & Gas Sector
By A. AbadSummaryIn a global industry where each year billions of operations move through a complex supply chain involving tens of thousands of suppliers, blockchain’s potential cannot be understated. Blockchain is particularly useful in the seismic sector, an area in which Repsol is pioneering the development of one first blockchain pilots.
Seismic rights require complicated processes to ensure the traceability of ownership and their limited use over time, and involve many different suppliers and international teams of workers. As part of the Oil&Gas Blockchain Consortium, Repsol has been responsible for developing this pilot to bring the use of Blockchain technology to the management of seismic rights.
-
-
-
Digital Experimentation to Meet the Challenges of Sustainable and Safe Energy
More LessSummaryIn the E&P sector, this is a time of decreasing margins and in which the requirements for meeting HSE (health, safety & environment) standards are becoming increasingly strict and necessary.
Repsol has opted for digital experimentation as a way of responding to the challenges of this new context in E&P. Drones, IoT, 3D printing are just some of the initiatives we are considering implementing in our facilities.
-