Verne Global

1 September 2017

HPC for Oil and Gas

Written by Nick Dale

Nick is Senior Director at Verne Global and leads our work across HPC and specifically its implementation within discrete and process manufacturing.

Multiple serious challenges face the oil and gas industry today. The output of mature oil and gas reservoirs are diminishing in volume, pressuring companies to seek new methods of extracting the maximum potential value from existing resources. At the same time, as exploration for new oil and gas resources moves into new and more remote areas, the cost of discovering these resources, and the technologies needed to extract them, are increasing rapidly — tripling over the last decade. Compounding these exploration issues are the stagnant per-barrel oil prices, which have added tremendous economic pressure to the already risky process of oil and gas exploration.

To date, High PerformanceComputing (HPC) has had a dramatic increase on the efficacy of discovery, production, and management of hydrocarbon resources in the oil and gas industry; thereby, helping to relieve the financial pain caused by these downward pricing pressures. Seismic imaging technology, a vital part of the industry’s effort to increase the success rate of exploration and extraction efforts, has only become feasible because of the increased availability and power of HPC systems. Despite these important contributions, the need for more robust HPC capability is pressing.

Only a few years ago, the datasets created during oil and gas exploration were gigabytes or terabytes in size. Today, high-density, wide-azimuth seismic recording capabilities are gathering 3D data in far higher detail than engineers a decade ago could imagine, resulting in datasets frequently measured in petabytes. Furthermore, processing the raw seismic data collected by receiver arrays and turning it into actionable geological intelligence depends on increasingly complex algorithms. As the demand for greater detail and specificity has become more urgent, newer technologies like full waveform inversion (FWI) and modeling have emerged to replace older techniques, placing vastly greater demands on computing systems. As the search for new oil and gas reserves probes into remote or untapped areas of the world, the dependency on these imaging technologies to help ensure drilling accuracy increases.

It’s not just the location of new resources that has driven increased HPC adoption in the oil and gas industry. Mature oil and gas fields currently account for 70% of the world’s oil and gas production, and as these fields move through their secondary or tertiary extraction phases, recovering oil from them becomes a greater challenge too. Here as well, HPC has played a crucial role in maximising the returns from these existing resources.

In order to derive continuing value from these assets, oil and gas companies have employed a variety of technologies to help maximise their investment, which includes creating detailed models of hydrocarbon reservoirs in order to predict the flow of the oil and gas in the Earth’s subsurface. Reducing the many uncertainties involved in hydrocarbon extraction from mature reservoirs requires a large number of simulations to be performed for each model, often supplemented by geostatistical modeling to increase accuracy. These technologies are all computationally-intensive; therefore, they have been dependent on HPC systems to deliver high-quality and timely results.

Despite these many advances, those working in the field are in constant pursuit of higher levels of accuracy and detail, and the ultimate goals of geoscientists and reservoir engineers remain far out of reach. Currently available technologies are insufficient to provide the level of certainty that is needed to ensure reliable recovery, and a poor understanding of geological structures often leads to expensive errors that can drive up investment costs and negatively impact the natural environment. To mitigate these risks, geoscientists are eager to employ more complex sensor arrays for higher resolution raw data, develop new algorithms and techniques to better represent subsurface interactions, and to introduce more physics to improve their existing models. This means that models, simulations, and maps are all poised to become larger, more intricate, and more difficult to process to create actionable information.

This constant need for more and higher-quality information is driving an unquenchable thirst for HPC power in the oil and gas industry. As the compute capacity of existing HPC systems is increased, engineers hungry for better data are quick to find ways to utilise that capacity. The voracious appetite for, and reliance on, compute ability has kept the oil and gas industry on the cutting edge of HPC development and application, and also ensures that HPC will play an even more central role in the industry’s development as it forges ahead into the geographic and scientific frontier.


Sign up for the Verne Global newsletter

Opinion, thought leadership and news delivered directly to your inbox once a month.