1. Home
  2. News and insights
  3. Energy in focus magazine
  4. The minds behind the machines part 2

What lies beneath: data-driven discoveries

Release date:
15 August 2019
In part two of our ‘minds behind the machines’ series, we meet geophysicist Xukai Shen, whose algorithm is helping BP to find previously undiscovered oil and gas

Last year, Lego enthusiasts in Hong Kong were assembling a ship made from a record-breaking 2.5 million bricks. Around the same time, BP’s Xukai Shen was using the same skill and care to construct an algorithm that would uncover a 400 million-barrel oilfield in the Gulf of Mexico. The similarities stretch beyond time.


“Think of the algorithm as made of pieces of Lego; I don’t build the Lego bricks, I just put the pieces together," says Shen. 


In this case, Shen's algorithm was built, not from Lego, but from the BP-developed infrastructure, such as seismic imaging, that helps explorers to see through the complex rock in the Gulf to uncover hidden resources.


Shen, who joined BP four years ago, after completing his PhD at Stanford, explains: “Essentially, we are trying to look through rocks and the challenge in the Gulf is that this particular kind of geologic body of salt is very different to the surrounding rock and also very complex in terms of its shape.”


High-resolution imaging

The algorithm and BP's supercomputer used the signature of seismic waves to penetrate the salt and produce high-resolution 3D images of ancient layers of rock thousands of metres under the seabed, essentially, providing a view of what lies beneath. 



“It is rather similar to using an ultrasound in prenatal care.”


Xukai Shen, geophysicist 


 This led to a major discovery of crude eight kilometres below BP’s Thunder Horse field in an area where it had long thought there was none to be found. 

The Thunder Horse field in the Gulf of Mexico, where Shen's algorithm was used to detect previously undiscovered resources

Accuracy and speed

Following the success of the technology, which is not only more accurate than previous surveying methods, but also processes data in a matter of days, compared with months or years previously, BP is looking to deploy it in other locations, including Angola and Brazil.  

Meet two more data science specialists

Subscribe to our email for the latest stories in energy, technology and engineering, direct to your inbox...