We would like to invite you and your colleagues to the April London Section evening meeting for a presentation on a machine learning approach to quantitative interpretation; and the fluid dynamics of multiphase displacement in reservoir rocks. The event will be held at Imperial College; Royal School of Mines, Prince Consort Road, London, SW7 2BP.
The Royal School of Mines is about 15 minutes walk from South Kensington tube station via Exhibition Road and Prince Consort Road.
Regards
Tim Lines
SPE London Section – Programme Chairman
Email: tim.lines@oilfieldinternational.com
Agenda:
Time: 5.00 pm-6.30 pm
Talk 1: A machine learning approach to quantitative interpretation.
Dr Ehsan Naeini, Ikon Science.
6.30 pm – 7.15 pm DRINKS AND NETWORKING BUFFET
Time: 7.15 pm 8.45 pm
Talk 2: The fluid dynamics of multiphase displacement in reservoir rocks.
Catherine Spurin & Alessio Scanziani, Imperial College.
Venue: The event will be held at the Department of Earth Science and Engineering, Imperial College London.
Directions: Please note the main entrance to the Department is via the Royal School of Mines Building on Prince Consort Road, between 10 and 12 on the campus map
Booking: All booking must be paid in advance and online please, via Eventbrite.
Email: katespe@aol.com
Cost 34 for SPE/PESGB/EI members, 44 non-members, 19 unemployed members. Non refundable 5 for students booking by Friday April 26 (19 after). All tickets have an additional Eventbrite fee.
BEFORE DINNER
5.00 pm 6.30 pm
A machine learning approach to quantitative interpretation.
Dr Ehsan Naeini, Ikon Science.
Machine learning can play an important role in making quantitative interpretation workflows more efficient and potentially more accurate. For example, petrophysical interpretation, facies classification and pore pressure prediction are fundamental for quantitative interpretation of the subsurface though they are generally time-consuming tasks. Pore pressure (Pp) prediction plays a critical role in the ability to predict areas of high overpressure and fracture behaviour, which are both correlated with production.
Any pore pressure model must be supported by petrophysically conditioned elastic logs and accurate multi-mineral volume sets calibrated to core data. Determining different facies is of course critical for reservoir analysis and moreover for seismic based reservoir characterisation. Performing either of these tasks manually requires a day or two per well for the practitioner.
With increasing number of wells, e.g. in unconventional plays wells are drilled at an unprecedented rate, performing classical workflows for facies classification, petrophysics, pore pressure and geomechanics prediction can be impractical (if not impossible) due to turnaround considerations. This, together with technical challenges in terms of complex stratigraphy, multiple play types, variable rock properties and the interaction of pore pressure and geomechanics, calls for more consistent, sophisticated, and faster analytical tools. This paper shows that a machine learning approach can potentially offer such a solution.
After discussing standard workflows, a supervised deep neural network approach is introduced as an alternative innovative tool for facies classification, petrophysical, pore pressure and geomechanics analysis enabling the use of all the previously collected and interpreted data to devise solutions which simultaneously integrate wide ranging well bore and wireline logs. The implemented deep neural networks can work in a cascaded manner such that the outputs from one are the inputs for the other.
That means one can train a neural network to predict multi-mineral volumes and also porosity simultaneously in which they can then be used to predict Pp. One can also design a deep network in such manner that allows prediction of certain properties of interest, e.g. pore pressure, at wells from elastic logs only. Then the trained model can be applied to inverted elastic properties from seismic amplitudes enabling one to compute a 3D earth model with a fast turnaround. Furthermore, the volumetric pore pressure model can be correlated consistently with cumulative production values from blind long horizontal wells leading to a robust validation of the technology. The results show a promising outlook for the application of deep learning in integrated studies such as those shown in this paper.
Ehsan is a Technical Director with more than 12 years’ industry experience, particularly in advanced analytics and machine learning, 3D/4D seismic processing, well tie and inversion. He has an MSc and PhD in Geophysics (Exploration Seismology) and a BSc in Physics. Ehsan joined Ikon Science in May 2012. He is currently the research and innovation director, responsible for the incorporation of latest techniques and developments in the area of reservoir characterisation into the Ikon Science’s RokDoc platform. He is also responsible for academic affairs in Ikon Science.
The talk will be followed by a quick demo.
AFTER DINNER
The fluid dynamics of multiphase displacement in reservoir rocks.
Catherine Spurin & Alessio Scanziani, Imperial College.
Topic: We use X-ray microtomography techniques to visualize, quantify the fluid dynamics of multiphase displacement in reservoir rocks. The combination of experimentation, imaging and modelling is transforming our understanding of multiphase flow processes and the way we characterize rock samples to design hydrocarbon recovery.
Catherine and Alessio work within the porous media flow and reaction research group at Department of Earth Science and Engineering. We study multiphase flow and reactive transport in geologic systems, with applications to petroleum engineering, carbon dioxide storage and planetary sciences.
Alessio previously graduated with a MSc in Energy Engineering at Politecnico of Milan, Italy. He is currently a 3rd year PhD student at Imperial College London funded by Abu Dhabi National Oil Company (ADNOC). He is imaging three-phase flow in porous media for Carbon Capture and Storage combined with Enhanced Oil Recovery applications.
Catherine is a currently a 2nd year PhD student at Imperial College London funded by the Presidents PhD scholarship and aligned with the Shell Digital Rocks project. She uses micro-CT imaging to explore subsurface fluid flow.