Big Data More Than Just About Numbers

“Google made the web accessible to answer queries with best documents and the best websites,” said Thompson. “Maana has to do things like search for which pieces of equipment have the most unplanned downtime, answers that can’t be found in a single document. Maana guides the user through the set of results to answer those questions.”

By using machine learning and natural language processing capabilities, Maana’s software can link and elevate raw data sets into concepts by bringing together fairly well structured data and unrelated data sets and detect things that are interesting from an oil and gas perspective, such as American Petroleum Institute’s well numbers and other well data.

Jeff Dalgliesh, who serves as Maana’s search enabled solutions leader for oil and gas and spent 18 years working at Chevron Corporation, said Maana’s technology takes a different approach to anything he had previously seen before – it has figured out how to ingest the technical data sets of drilling applications used by major oil and gas companies.

“Let’s say that you want to search for the word ‘kick’ in a technical data set,” Dalgliesh explained. “The Maana software is able to ingest the technical data set into their application, then ingest additional data sets not connected to the initial data set, as well as data from sensors and downhole subsurface logs.”

Lots of companies build algorithms to uncover events.

“What’s unique about Maana is that data can be uncovered and reintegrated back into the data set. In essence, it creates a conceptual model of a well and can relate it from the well to the kick,” said Dalgliesh.

In addition to learning the structure and relationships in the data, Maana also performs a variety of natural language processing tasks on unstructured content to recognize and resolve entities, learn term associations and jargon, and incorporate reference datasets, such as industry-specific coding systems. Once the word or phrase has been integrated into Maana’s glossary, documents such as International Association of Drilling Contractor timelogs can be searched and mined for specific problems, giving operators a better idea of what types of well problems might occur.

With existing databases, someone looking for information on a well would have to already know what well it’s part of. With Maana, a word or phrase can be typed that can pull up information.

“In one case, the name of a person could be entered, and would show which wells that person drilled.  The program understands that the person is related to the wells. The software program allows for searches of how the wells a certain person drilled performed, or incidents that might have occurred during a shift on which that person worked.

In the past, gathering this data could have taken days or weeks, but can now be found with a single query in seconds, Dalgliesh said.

“We can figure out how to do the analytics, but being able to search for little nuggets of information can make a difference.”


12

View Full Article

WHAT DO YOU THINK?


Generated by readers, the comments included herein do not reflect the views and opinions of Rigzone. All comments are subject to editorial review. Off-topic, inappropriate or insulting comments will be removed.


Most Popular Articles