NEWS

Big Data Algorithms Optimize Oil Wells

The global oil producing industry spends millions of dollars every week on preemptive well monitoring—sending technicians to remote sites to check wells for operational faults. Though the idea of operating oil wells with the help of the Industrial Internet of Things (IIoT) is catching on, labor costs still dominate operating expenses. Not only does the typical oil company still send field technicians out driving from well to well to regularly inspect the equipment and report back on findings, even using IIoT technologies in many cases does not resolve issues with labor costs, argues Nav Dhunay, president and CEO of Ambyint.

“If you just analyze the data, all you’ve done is push the problem around,” Dhunay says, arguing that the main expense is still labor, albeit a different kind. “Now you have people staring at screens, looking at graphs, trying to understand the data. It just pushes up the labor cost.”

Ambyint, based in Calgary, Canada, focuses on taking labor costs out of operating oil wells by removing the manual component of data analytics, using Big Data to create intelligent, high-functioning wells. Its latest product, launched last week, is AmbyControl for EPM (Electric Prime Mover), which links any oil well operated by electric motor to the Ambyint platform. Through Ambyint’s data algorithms, well managers are provided with detailed real-time well analysis along with operational recommendations to optimize the wells. Well managers are also given the ability to remotely control each of their well assets directly via the Ambyint mobile and web app.

This puts an end to the preemptive labor costs that oilfields typically face. “Through leveraging sophisticated technology to connect these machines and effectively make them smart, our AmbyControl device is able to entirely change this model, putting detailed, real-time well performance data and remote well control directly in the palm of every operator’s hand,” Dhunay says. “Using our technology, operators are able to monitor their wells 24/7 and receive instant alerts of any faults or interruption to on-site operations, which they can react to quickly, and from wherever they are.”

AmbyControl for EPM features include remote start/stop of electric-powered pumpjacks; access, via mobile or web app, to real-time torque information, including trends; access to insights into the on-time performance of individual wells; instant notifications of any abnormal pump activity or shutdowns; and instant torque violation alerts to help avoid costly workovers.

Ambyint came about last year as somewhat of a reinvention of the company PumpWell, which focused predominantly on artificial lift for oil wells. Developed on the idea that pumpjacks could be operated more efficiently, with wear and tear reduced on the equipment, PumpWell was built around mathematics and data collection from an extensive number of sensors.

Dhunay was brought onto PumpWell a few years ago to reinvent the company, reinvent the technology, and chart a new path. He gravitated toward all the data that PumpWell had been gathering for years. “My background was in consumer technology,” he says. “I wanted to build a system where we could use all this data to run an oil well.”

The Ambyint platform is built around pattern matching and machine learning, and is geared toward controlling technology automatically. “It’s consumer technology, hardened for industrial use—we’ve made it lightweight, very small, with discrete sensors that can be installed on an oil well,” Dhunay says. Information is transmitted to a cloud system, and from that it spits out an action.

“If you just analyze the data, all you’ve done is push the problem around. Now you have people staring at screens, looking at graphs, trying to understand the data. It just pushes up the labor cost.”

A typical problem in Canadian oil wells is wax buildup that attaches itself to rods in the pump, which then becomes sluggish. “So companies dump chemicals into the wells to break down the wax buildup,” Dhunay explains. “Our system can predict wax buildup and tell you exactly how much chemicals to dump in, and save chemical use.”

In one case study, an oil producer left half of its wells as is, the other half with Ambyint’s system. “Within three months, they saved $15,000 on just chemicals,” Dhunay says.

Transportation is another factor in that case, and the oil producer is also saving money on operator visits. Trucking is a large component of the cost for U.S. oil wells, one example being the trucks that unload tanks. “The trucking companies are just working off of a schedule; they show up once a month,” Dhunay says. “They should show up when the tanks are actually full. We can detect how much oil you’re producing and tell you when your tank is full.”

Ambyint also has AmbyControl for VFD, designed to work with variable-frequency drives; and earlier this year, the company launched AmbyControl IoT Bridge, which collects, analyzes and delivers all data from competing well management systems like Lufkin and Weatherford systems.

With a variety of AmbyControl and AmbySense products, Ambyint is creating what it calls a self-driving pumpjack. It’s installed on more than 600 oil wells, predominantly within Canada’s Alberta and Saskatchewan provinces. The company is also learning more about how oil wells are operated in the U.S. In Canada, for example, wells produce a lot of heavy oil, Dhunay points out. “Down in the States, there are not a lot of waxy wells, but there is a lot of sand and debris,” he says. “There will be new challenges that the system will adapt to and learn.”

https://www.automationworld.com/big-data-algorithms-optimize-oil-wells

By:
July 7, 2017