OTC Commentary: Big data comes to deepwater drilling

By Matthew Franchek

Matthew Franchek

An unexpected equipment failure can affect offshore producers much the way an unexpected closure of Interstate 10 affects Houston drivers. Everything comes to a stop. For drivers, that means lost time. For offshore operators, lost time means less money to the bottom line.

That’s bad news for companies and their shareholders as unplanned equipment shutdowns cost billions of dollars each year, potentially driving the cost of producing offshore oil well above today’s market prices.

The safe and economical recovery of future oil and gas resources demands operational efficiency, and this efficiency can be realized only if there are no unplanned downtimes due to equipment failures. Almost all oilfield equipment now is fitted with sensors that provide data about operations, and increasingly production companies are realizing that those terabytes of data streaming from the monitoring sensors built into equipment can be used to improve operational efficiencies and, ultimately, profit margins. Why shut down if they don’t need to?

The traditional method of interpreting the data takes months and requires that it be stored for extended periods of time. Data analytics is changing that.

Data analytics uses mathematical modeling to harness “big data,” the huge amounts of data that flow from the increasingly connected world around us. Posts on Facebook, Instagram and other social media sites can be captured and analyzed for trends and other useful information. Oil and gas companies are interested in using the industrial internet of things – the idea that “smart” machines using big data technology and machine learning are better than humans at accurately and consistently capturing data.

The goal is to perform maintenance as it is needed, rather than following a rigid, pre-set schedule – potentially losing production time for unnecessary maintenance or suffering an unanticipated shutdown when equipment fails before the scheduled maintenance time.

The system operates on the same principal as that used by your car to estimate how much farther you can drive without running out of gas, an estimate that is constantly adjusted based on your driving patterns. The potential benefit goes beyond financial savings through avoiding unnecessary shutdowns. It also is expected to reduce risks to both workers and the environment as companies gain advance warning before potentially disastrous accidents.

This technology goes one step further by analyzing overall operations, including worker performance. Complete coordination through the integration of data from both workers and equipment is now possible, with workers assured that the equipment is available and reliable. It can also reduce data storage costs – data can be stored in the cloud rather than on production platforms, where space is at a premium.

Beyond that, data can help identify future design modifications much like evolution. The reality is that the equipment can now “speak” through the data streaming from various sensors to identify when it is stressed, much like an athlete can report signs of injury before it becomes debilitating.

Companies already have begun to realize these savings; key players in the industry developed a process for using this data, which has gone through rigorous testing over the past 18 months. I worked with them to develop a curriculum to train engineers to use mathematical modeling, simulation and data processing to capture and use this data for real-time condition and performance monitoring of oil and gas production systems.

Monitoring and managing big data is a growing discipline, not just in the energy industry but also in health care, aerospace and other industries.

The large variety of data ranging from numerical to alphabetical to images, all streaming in real time from thousands of sensed values, is what makes big data analytics big. There is a renaissance underway in the oil and gas business, and model-based data analytics is its foundation.

Matthew Franchek is founding director of the subsea engineering program at the University of Houston and has developed a three-course certificate program on data analytics for engineered systems.