|Repsol's headquarters in Madrid, Spain. / Photo by: Luis García via Wikimedia Commons|
Global energy firm Repsol has embarked on a partnership with Google Cloud that will use big data and artificial intelligence in optimizing the management of the Tarragona refinery. The joint venture has the possibility of adding 30 cents on the dollar to Repsol’s refined barrel margin, according to Maurice Smith, reporting for JWNEnergy.com.
The Tarragona refinery is one of six refineries that Repsol operates in Spain and Peru. It is Repsol’s third-largest unit and is capable of processing 9.5 million tons of raw material yearly and distilling up to 186,000 barrels of oil each day. Its storage tanks can hold a million cubic meters of refined products.
The agreement calls for Google to make available to Repsol its data and analytics products, the experience of its professional services consultants, and its machine learning managed service, Google Cloud ML. The latter will help Repsol in developing machine learning models to be used in the refinery’s operations.
There are more than 400 variables involved in the management of petroleum refineries which are among the largest and most complex industrial facilities in the world. Such variables measure pressure, temperature, flows and processing rates, and require a high level of computational capacity and a huge amount of data control. At present, only up to 30 variables can be integrated digitally, but the joint undertaking started by the two companies aims to increase that number by tenfold.
Repsol said it chose the Tarragona facility to host the project because the online architecture of its production schematics is ideal for testing and implementation. The company added that the project is part of the ongoing digitalization, innovation, and technology initiatives for improving competitiveness and efficiency in all of its business units. It is optimistic that by using big data and artificial intelligence in all of its refineries, it could generate an extra $100 million in additional revenues per year in its downstream business.