Oil and gas companies are generating huge volumes of data from a variety of systems and sensors in the field -- information about drilling, weather conditions, seismic activity, and other factors. In order to remain competitive in today's global market, they need to be able to quickly derive insights from this data.
A key to being able to leverage all the data is the use of graphics processing units (GPUs), which provide the computing power and high-speed memory needed for these companies to analyze petabytes of data in a matter of milliseconds.
"New parallel computing platforms and APIs are coming together to use GPUs for general purpose computing," said Chris Niven, research director, Oil & Gas, at research firm IDC.
"Special microprocessors are also now available that integrate GPU acceleration, 3D rendering, video acceleration, and wave-table synthesis into a single chip," Niven said. "The combination of GPUs and a CPU are now available that can accelerate analytics, deep learning, high-performance computing, and scientific simulations."
GPU manufacturers are creating learning institutes and working with vendors such as SAP, Microsoft, IBM, Dell, Amazon, and all the major cloud vendors to build an ecosystem focused on artificial intelligence (AI) and deep learning, Niven said.
"The future for this vision is to foster accelerated computing and continuous learning on enterprise data for companies to apply analytics, modeling, simulation, and advanced analytics like cognitive processing," Niven said. "Companies are already adopting analytics initiatives in upstream production."
For oil and gas providers, it makes good sense to develop an asset performance management (APM) environment to monitor the vast amounts of data being deployed in the field such as new equipment, devices, sensors and other Internet of Things (IoT) components.
"It is important for companies to organize and analyze this data and understand and learn what works and what doesn't, to safely operate and optimize production," Niven said. This is especially true for APM, as companies strive to automate production.
"Drilling is slow these days and the production of oil generates much needed revenues for the company," Niven said. "Oil and gas companies are developing data management platforms to manage hydrocarbon production data." They're also applying descriptive and predictive analytics and cognitive processing to monitor the health of equipment and machines, to prevent unwanted downtime and disasters before they occur, and to help optimize upstream oil and gas production.
"GPUs combined with CPU is also applicable for downstream, especially with regards to modeling and simulation of equipment and machines as the Digital Twin has become a high IT initiative for APM [application performance management) in refining and petrochemical plants," Niven said. A digital twin is a digital representation of an industrial asset, which enables companies to better predict the performance of machines.
"As in upstream oil and gas, the downstream sector also recognizes the value of parallel processing, AI and fostering a continuous learning environment to keep the plant operating at full capacity and avoid downtime and disasters before they occur," Niven said.