The Internet of Things is not just hype for Woodside, with a combination of networked sensors, cloud infrastructure and analytics already paying dividends for the oil and gas company.
Woodside CTO Shaun Gregory told the Amazon Web Services Summit in Sydney that what began as a prototype at the company’s Pluto liquid natural gas plant in Western Australia was set to be rolled out across the rest of its infrastructure.
When Woodside began operations at the $10 billion Pluto LNG plant in 2012, the plant was equipped with some 200,000 sensors that monitored operations including temperature and pressure, Gregory said.
In 2014 Woodside undertook a project to assess the potential of harnessing information from a subset of those sensors to improve the Pluto’s operation, using cloud services to crunch the data.
The subject of the project was the plant’s Acid Gas Removal Unit (AGRU), which would periodically experience a problem dubbed “foaming” (and according to Gregory resembles pouring beer too quickly into a glass).
Foaming was capable of shutting down the plant for weeks or even months and was difficult to detect because the AGRU is enclosed.
Working with data scientists, information from 10,000 of Pluto’s sensors, and 30 years of operating experience, Woodside was able to develop a model that that would not only detect forming but could produce a two-hour, 72-hour and up to a one-week forecast of when foaming might occur.
The project took just six weeks.
“This had a dramatic impact on the culture of our organisation, because we were used to technology evolving in many, many years,” Gregory said. “Now it was evolving in weeks.”
The project was carried out in 2014. In the wake of its success, Woodside last year started feeding data from all 200,000 of Pluto’s sensors into Amazon’s cloud.
“We actually got a call from Amazon,” the CTO said: “What the hell are you doing? Little bit of a spike in your data – would you like some help?”
Data from those sensors now feed into 6000 analytics models that, through 3 million calculations per day, assess everything from the chances of foaming will occur to whether a valve is likely to fail and consequently should be subject to maintenance.
Woodside is now looking at expanding the system to include its other onshore plants and well as offshore facilities and vessels.
“We have a lot of infrastructure that has a lot of sensors on it that aren’t connected yet and we’re looking to connect those to analyse the next generation of problems,” Gregory said.