More from Intelligent Energy (May 2014)Maersk trials Eclipse in the Amazon cloud. Exxon ‘cyber security is hard and getting harder.’ Yokogawa on safety, humans and automation in oil and gas and aviation. BP’s ‘basket analysis’ of oil production. Molten on multi terabyte fiber monitoring. Idmog on novel ‘big data’ technology.
Morgan Eldred showed how Maersk has been trialing s Schlumberger’s Eclipse fluid flow reservoir simulator in the Amazon web services cloud. A major issue is data transfer from the corporate network to the cloud fixed by Panzura’s Quicksilver technology. Inside the cloud, data transfers between simulated offices used OpenVPN on standard Linux Amazon machine images. Eclipse was run on a Linux C3 ‘xlarge’ size server with solid state drives. The proof of concept worked but challenges remain—notably licensing issues. The authors noted the ‘agility’ of working in the cloud, for instance benefitting from the latest technology refresh à la Intel Xeon E5.
Darrell Pitzer (ExxonMobil) observed that cyber security is hard, and will get harder, but business goes on. Scada vulnerabilities are on the rise, ‘the bad guys have found us.’ Control systems are hard to secure because of the 10 to 20 year equipment lifetime and the fact that they are designed for transparency and efficiency rather than security. Also, while many experts are reaching retirement, the hacker’s knowledge is ‘comprehensive and constantly building’ particularly with tools loke Metasploit and the Shodan search engine. Companies need to reduce risks to a ‘comfortable’ level—e.g. by banning removable media and changing default accounts and passwords. Training is also important. Also, verify that procedures are followed, for instance by throwing some USB drives around the parking lot and see who plugs them in! You also need a disaster recovery plan in case a site gets wiped out by a virus. Oils are not the most respected and loved companies, ‘there may be folks out there trying to shut you down.’
Maurice Wilkins (Yokogawa) investigated human behavior in critical situations such as the Indonesia Airways Airbus 380 incident (where the pilots saved the day) and the Texas City refinery explosion where the operators were overwhelmed with contradictory information on the plant’s status. With refinery losses trending upwards, is more automation the answer to process safety? For Wilkins, ‘having humans in the picture at the time of an emergency can be beneficial. But we need to prepare for abnormal situations where lots of confusing data and warnings are created. The answer is better guidance and support for operators. In the Texas City example, a ‘procedure assistant’ could have triggered actions that might have saved the day.
Richard Bailey (BP) showed how market basket analysis, as used in to analyze the contents of shoppers’ supermarket trolleys, can be repurposed for production data analytics. The oil and gas ‘basket’ has relatively few items—just start/stop events on injector and observations at producers. These are analyzed with ‘directed pattern search’ to look for cause and effect. Bailey reports that this is very successful and has helped BP optimize its gas injection strategy and ‘sell high.’ The technique is now embedded in a user-oriented toolkit.
Jeff.Liddle (Molten) presented on behalf of BP work done on distributed temperature sensing with fiber for well surveillance. Continuous monitoring with fiber can produce very large amounts of data. A single comprehensive monitoring operation of, e.g., a frac job plus flowback, can generate 150 terabytes. Fiber is however easy to deploy downhole and behind casing and requires no power or comms and it works over long distance, up to 25km. Interpretation tools are primitive and pertinent use cases are so far elusive.
Aymeric Preveral (Idmog) provided a round up of ‘big data’ technologies likely to impact oil and gas, making an analogy between the problems that Google and Amazon have and those of streaming data from the oilfield. Current SQL-based data stores and applications scale poorly and require a significant administrative burden to maintain. Data gets replicated and causes more work for IT. The answer is the NoSql as deployed in Google’s BigTable and Amazon’s SimpleDB. Preveral proposed a distributed architecture involving these technololgies but observed that connectors for Prodml, Ppdm and OPC-UA will have to be built. The IS presentations can be acquired through the OnePetro website.
Click here to comment on this article
If your browser does not work with the MailTo button, send mail to email@example.com with OilIT_1405_13 as the subject. Web use only - not for intranet/corporate use. Copyright © 2014 The Data Room - all rights reserved.