Oil and Gas Industry and Deep Learning

Baker Hughes (oil and gas software giant) developed its own software for the world’s leading oil and gas companies that require supercomputing capabilities. Furthermore, the Baker Hughes’s systems can serve double-duty for emerging deep learning workloads. 

The HPC requirements make sense for an industry with a lot of data and deep learning technology is the best solution for this industry.

In fact, HPC is using deep learning. Indeed, Baker Hughes is using an InfiniBand connected cluster of Pascal GPU-based Nvidia DGX-1 appliances to support new deep-learning solutions for the oil and gas industry. Deep learning is used for overall rig optimization, traditional HPC simulation based areas in resource discovery and production, and other applications.

Arun Subramaniyan (VP of Data Science and Analytics for Baker Hughes Digital and GE) said that “most of our models are a combination of physics augmented by deep learning or AI that requires input from physics-based models. For example, for some of the large-scale multiphase simulations the oil and gas industry has been running for decades or larger turbulence models can all benefit from parts of that workflow replaced by neural networks.”

Arun Subramaniyan also said that they “are not necessarily replacing the physics models themselves with neural networks. These models are an important part of the overall oil and gas ecosystem, but there are many pieces of those workloads that can be accelerated through a combination of deep learning and physics models running together”. Furthermore, generative adversarial neural networks (GANNs) would be better than deep learning in physics-based HPC simulations.

In fact, Baker Hughes is using its DGX-1 cluster for more standard convolutional (CNN) and recurrent neural networks (RNNs) to support new software packages. So, it leads to improving many areas such as repairs or problems, safety or environmental concerns, and more specific problems like corrosion. Deep learning could be used to improve much other software for traditional HPC problems like oil and gas discovery and extraction.

Baker Hughes trained its neural networks on the DGX-1 clusters then pushes this model over the cloud as an inferencing service to its clients with retrained updates pushed through when required. Then, these models would be used by their clients, who would be able to retrain the network with their own data in a cloud environment.

There are many deep learning platforms to build sophisticated models, but the problem is how to use these platforms in this industry. It is pretty hard to make this deep learning process easy for this company.

This company wants to use deep learning models in real use cases, but must of the deep learning applications are still a young technology.

Actually, for deploying the deep learning system at scale, it is needed to build a lot of the deployment framework and infrastructure. Furthermore, it needs the cloud and Kubernetes as the base on a mesh network.

For example, Subramaniyan that they had a deep learning model built in TensorFlow, another model built in Caffe, and another model built in PyTorch with each needing to run in inferencing mode in the same overall flow.

Baker Hughes has its own AI team to deploy deep learning at scale. In fact, the ROI is that oil and gas companies won’t have to develop their own deep learning software. Indeed, this leads to better operations. Like oil and gas companies don’t want to be business of infrastructure management, but Baker Hughes has to do it.

HPC dominated areas of oil and gas are using deep learning. Indeed, this could lead that this industry starts investing in HPC hardware, adopting of GPUs or even separated networked clusters for deep learning versus traditional simulations. In fact, deep learning systems need dense GPU systems combined with powerful CPU racks. So, they are using GPU system for that reason. They want production optimization and better operations and process management, so they are using GPU plus CPU mix.

Subramaniyan will launch Volta GPUs that will also be packed into an appliance ala the DGX-1.

Binu Mathew (VP of digital development at Baker Hughes) said that “the oil and gas industry needs exaflop and beyond levels of performance in the coming years.” Furthermore, cloud-based GPU capabilities will new ways of operating in the oil and gas industry now and in years ahead.

Bigdataguys consults fortune 500 companies on using blockchain technologies, artificial intelligence, & data science to enhance their operations. We’re a full-service machine learning, artificial intelligence & deep learning firm.

Contact us Bigdataguys offers deep learning consulting and bootcamps.