As the concept of “big data” becomes more prevalent in the oil and gas industry, IT providers are scrambling to come up with solutions. Improvements in sensing technology mean that operators are gathering more data faster than ever before, and managing, acting on, and storing those data can produce major headaches.

In a recent Hart Energy webcast titled “Big Data Analytics in Oil and Gas,” SAP’s HANA (which stands for high-performance analytic appliance) in-memory data platform was showcased as a potential remedy. According to SAP literature, the HANA platform overcomes traditional barriers by eliminating data preparation, pre-aggregation, and tuning, in turn providing real-time answers to complex problems.

“Big data growth is forcing the enterprise to look into efficient ways to drive the organization,” said Yuvaraj Athur Raghuvir, senior director and SAP HANA platform solutions manager. “It is stressing the existing infrastructure.”

Yuvaraj pointed to four “pillars” that are driving the dominant aspects of big data – scalable storage, unstructured text analysis, predictive analysis, and streaming data such as sensors and social media. “We want to look at unstructured analysis,” he said. “These are data that we have not traditionally considered as important and relevant to business. Once we start looking into information in the form of text or video or media or geospatial, we do it in a general way.

“But these data are structured differently, and we want to do analysis to make other contexts more relevant.”

He added that social media and mobile data, data generated outside of an organization, are becoming relevant as well.

The HANA platform addresses these trends by moving the analytics to where the data are stored rather than the historical approach of moving these data to where analytics happen. Typical, historic IT infrastructures located the operational data in one system and moved them into an application system for analysis. “We can bring these two things very close together,” he said. “We can start thinking about data in a raw form and offer analytics on top of them where they sit almost immediately.”

This is accomplished through engines that send the code to the data rather than sending the data to the code. “When you send the code to the data, you’re sending a smaller amount of information to a larger dataset, and you’re doing the processing at the place where the data exist,” Yuvaraj said. “This saves time, effort, and money. Developers [also] now have the ability to create a variety of engines that are able to handle transactional information and analytical information at the same time.”

Case studies

In one study, a gas field was experiencing unplanned production losses. Because the large dataset was in HANA, an analyst was able to rapidly drill down to determine that compressor failure rates were contributing to these losses.

By analyzing the maintenance issues with these compressors, the fuel filters appeared to be the prime cause of failure issues. It was determined that the 3400 series of compressors, which use smaller filters than the 3500 series, showed higher failure rates than their larger cousins. “After making this discovery based on analysis and purposely building a model with this information, our analyst moved on to a recommendation to determine maintenance planning,” Yuvaraj said.

Since all the data were integrated in HANA, the analyst was able to find wells with a combination of high production and high compressor failures. The maintenance frequency on the compressors associated with those wells was increased. This prioritized maintenance reduced the failures on the high-priority wells, and over the next six months the production losses were significantly reduced. Production is now on target.

Another scenario involved drilling problem prevention. By using the rich information coming from the sensor data such as formation and mechanical parameters and contextualizing that by augmenting it with operation logs, the operator had the ability to analyze drilling data from several wells. This analysis offered the operator the opportunity to characterize the issues, classify them in terms of severity, and identify causality patterns to provide alerts and recommendations. This led to improved efficiency and the avoidance of stuckpipe.

In a workover probability optimization study, gas wells producing both gas phase hydrocarbons and liquids were encountering liquid loading due to water condensation in wells with low gas flow. The customer used HANA to track and manage its production assets worldwide. The solution involved analyzing root causes for equipment failure on sites, performing predictive maintenance to reduce cycle time, improving asset yield, and analyzing the payout for each well for deciding remediation operations like swabbing and workovers.

Enabling technologies

Also participating in the webcast were John Thomas, technical solution architect and SAP ambassador for Cisco, and John Karagozian, global SAP principal for NetApp. Thomas agreed that the explosion of data makes it challenging for oil and gas companies to make the right decisions in a timely fashion.

“Big data is opening up a whole range of possibilities toward driving better business decisions,” he said. “How do you host all of this? How do you maintain an optimized infrastructure and cost profile for hosting a range of today’s applications while planning for the growth into the big data space?”

Cisco’s answer is the FlexPod, an integrated and optimized collection of IT infrastructure to run applications. The FlexPod combines NetApp’s network storage solutions, Cisco’s networking technologies and its unified computing system, and of course SAP’s HANA.

“One of the things that stands out about FlexPod is where we can place it,” Thomas said. “It can scale to be a large data center platform for hosting a multitude of applications, or it can be put in remote locations. Obviously in the oil and gas industry we have both situations.”

Because of security and regulatory issues, Thomas said it’s important to be able to put infrastructure in a variety of locations. The FlexPod can scale up and scale out to accommodate this infrastructure without requiring a multitude of different platforms, providing a great deal of flexibility.

Ultimately FlexPod provides a scalable solution that makes computation independent of storage and vice versa.

Karagozian said that the combination of Cisco and NetApp working in the SAP environment results in time savings. “We can cut the time by 25% to 40%,” he said. “If we can cut time off of the application side, there is typically billions of dollars in savings.”

Part of this is NetApp’s snapshot technology, which allows users to save their work throughout the project. “When you work with SAP, it’s a linear process,” Karagozian said. “You do A, then B, then C, and if you hit D and D fails, you have to start at A again. That’s when you lose productivity.”

The pointer-based snapshots, however, allow users to save their work any time during a mission-critical process. “If D fails, they go back to C,” he said. “All of that productivity is saved.”

These types of innovations have the potential to allow oil and gas companies to reshape their database approaches to support real-time intelligence and decision-making.

To view the webcast, visit secure.oilandgasinvestor.com/webinars. To download the white paper, visit epmag.com.