Using Big Data Technology In Oil
When he was preparing to launch his company in 2015, Dr Nikhil Shah was confident he’d spotted a gap in the market that his technology was able to fill. “Cloud native seismic processing and in particular full waveform inversion (FWI) had not been tackled by the industry service sector,” he recalls. “The gap existed for an entirely data-driven earth model, building on the cloud with real-time monitoring of the job to enhance accuracy and automation.”
In layman’s terms, what Shah’s company, S-Cube, essentially does is translate seabed geophysical survey data into precise and highly accurate 3D models of the subsurface to pinpoint the best locations where operators should drill for oil and other natural resources.
Shah, now COO, launched the business when his team developed an adaptive waveform inversion (AWI) algorithm that has since gone on to fundamentally alter how petroleum exploration is approached. It’s enabled global energy firms, such as Ineos and Tullow Oil, to use artificial and automated intelligence intelligence to mimic field data with numerical simulations to reduce risk and uncertainty in identifying profitable deep-sea deposits.
How Does Adaptive Waveform Inversion (AWI) Work?
Shah describes AWI as, “a smart optimisation cost function within the XWI next-generation waveform inversion framework that searches for the rock 3D velocity model that best predicts seismic measurements leveraging the full recorded wavefield.”
He explains that, “AWI provides an adaptive way of scoring the prediction using a mathematical mapping between field data and model predicted data. Following through the multivariable calculus, this generates more accurate model updates when starting further from true answers. It was the breakthrough on which S-Cube was founded.”
Shah firmly believes that his approach is superior to traditional methods. “XWI on AWS (Amazon Web Services) is unique to the industry. It encompasses the most sophisticated waveform inversion techniques with the largest compute facility on the planet. This combines to give verifiable accuracy and more automation,” he states.
He also points out that, “Our additional advantage comes from having machine learning capability built into our platform to automatically identify patterns that optimise job performance and settings. The more data and user projects the framework sees, the more it self-improves.”
The Success Of S-Cube
S-Cube’s solution is cutting-edge in terms of technology and it’s entered a sector that’s known for being technically conservative and risk-averse – and been highly successful. How did Shah’s team accomplish this? “There were two partnerships we struck early that were key to overcome hurdles of being a new player,” reveals Shah. These partnerships saw Tullow Oil becoming the pilot user (working with S-Cube directly) and the AWS cloud powering the whole process from uploading the data to running the computations to viewing the results.
“This three-way partnership allowed us to quickly generate results and capitalise on the burst capability available. This was used to create value for Tullow in prospective acreage in the Central Atlantic Margin,” Shah details.
The Tullow results of XWI have been totally user driven, as Shah explains: “The results have been as a result of Tullow using the functionality of S-Cube Cloud platform to identify an opportunity to increase convergence beyond the baseline score of 59% to 75%.” These results from the pilot project were co-presented by at the ISC (International Super Computing) event in Frankfurt with Tullow saying “FWI has made automated high-resolution velocity model building a reality and now XWI on AWS is taking this to the next level. Using S-Cube, we are able to robustly deploy XWI on the cloud in the highly agile and interactive manner needed for superior quality results that give E&P competitive advantage.”
Shah describes how his technology is being deployed elsewhere in the industry: “It has enabled predictive power to replicate sonic logs from already drilled wells in blind well tests, where the sonic log readings have been held back from the data fitting.
“Starting far away from the downhole measurements, the model velocity at well locations can shift in the correct direction with no forcing required by up to 500ms. Hence the obtained models can be used to place future wells with confidence.”
The Future Of AWI
When asked what the ultimate impact of his technology on oilfield exploration could be, Shah sums this up succinctly. “Fundamentally, the aim is for all modern marine seismic data to pass through XWI on AWS to generate rock velocity models with unprecedented accuracy, resolution, automation and predictive power.
“We will be able to reduce model building time to three weeks rather than nine to 12 months; identify untapped deposits with more certainty; and minimise drilling risk and reduce dry holes.”
Not one to rest on his laurels, Shah hints that there’s even more to come from S-Cube in the near future: “We’re working on enabling real-time decision making whilst a seismic survey is being conducted.”