The 2022 edition of the American Business Conferences Onshore Wellsite Automation was held earlier this year live in Houston and online. As in previous years the event was categorized by the predominance of MQTT-based solutions displacing traditional scada. Todd Anslinger presented Chevron’s ‘industrial internet of things’ (IIoT) approach that envisages an IIoT hub in the cloud. Anslinger is on the steering committee of the Eclipse Foundation’s MQTT/SparkPlug. Chevron’s definition of IIoT embraces an IoT hub in the cloud. For a large company like Chevron, this may mean an in-house cloud, possibly using the Microsoft Azure Stack either on premise chez Chevron or possibly at Rockwell where Anslinger is a customer advisor. Smaller operators may be more happy with a public cloud. Chevron uses the Azure IoT hub for ingestion and the Azure Data Explorer for analysis. Use cases include flaring reduction and methane management using IIoT connected infrared cameras. Carbon capture at Gorgon and Quest is monitored with IIoT sensors. Anslinger proselytizes for the MQTT/ Sparkplug pub-sub data exchange standard that ‘drives efficiency at-scale’.
David Blanco (Autosol) and Bryan Hendrix (PureWest) reflected on the operation challenges of traditional scada. Slow polling, inconsistent data and no remote device management meant that scada had become a barrier to the digital transformation. The solution was ACM (the Autosol Communications Manager) and Inductive Automation’s Ignition scada. Again, MQTT is the preferred protocol for data transfer (although ACM uses a proprietary system). MQTT is said to deliver a ‘30-70%’ network speedup. This is just as well since ‘managing MQTT and Edge Devices over low bandwidth networks is not feasible’. Returning to the digital transformation theme the authors concluded that while ‘an entire institution may benefit from digital transformation, but that doesn’t mean it should go all-in right away. It’s often best to start in one area, identify progress and best practices, and then expand the reach of digital transformation across your campus.’
Evan Rynearson described Koda Resources’ MQTT retrofit project that set out to keep costs low when updating its legacy scada systems. There are ‘thousands of devices out in the field that don’t know the MQTT protocol’. Switching from a poll/response system to MQTT without replacing existing devices is not without risk. The rationale for such a move is the goal of big loop optimization from business down to operations. Again, Inductive’s Ignition portfolio has enabled the transition. The Sparkplug B protocol was used to create a meaningful tag namespace to identify equipment. An OPC/MQTT converter also ran. Different devices required different approaches but, as Rynearson warned, ‘MQTT is a new communication protocol, and very different to what we’ve grown used to and standardized our whole scada lives on’. But the effort has proved worthwhile with improved alarm response, flowback monitoring and providing new high resolution data. Greenfield sites are now scada-equipped ‘in minutes’ along with a template-driven namespace. The MQTT engine sends out a new ‘tags on birth’ message. The ‘send on change’ paradigm has produced huge bandwidth savings. Initial targets for deployment could include alarms, MQTT-capable devices and fixing that ‘low frequency data your data scientist is complaining about!’
Brandon Davis showed how Red Bluff Resources (RBR) has implemented IIoT edge devices to leverage real-time data. RBR uses a Red Lion Graphite/Crimson HMI front end and wireless comms from SignalFire whose Ranger node got a shout-out as providing configurable IO*. AFTI’s Watchdog pump monitor is also deployed as is Sensorfield’s monitoring technology and Andium’s video cameras. All comes together in RBR’s Azure instance (thanks to MQTT naturellement) for processing with Stream Analytics and analysis with PowerBI. Esri ArcGIS is used to display scada data in both map and schematics. ArcGIS tools monitor tank batteries and line pressures.
* In a separate presentation, Sandro Esposito drilled down into SignalFire’s sensor-to-cloud MQTT capabilities and LTE telemetry. MQTT/SparkplugB and LTE-M Cat 1 are said to be ‘the IoT industry standard’.
Marshal Hall (Mallorn Energy), reflecting on the success of customer relationship management systems, began playing with a CRM from Quickbase and soon realized that this relational database could be repurposed for use as a … well relationship management system, the WRM. By using this low code software, Hall eliminated the need for over 20 spreadsheets, replacing a reporting tool and ‘saving thousands of dollars’. After a quick RDMS tutorial, Hall showed some of the (rather compelling) results. Quickbase seems particularly suited to generating data-driven diagrams and reports for well cards, status updates, dashboards and even wellbore schematics. Hall concluded that ‘there is a massive amount of inexpensive software out there that can be used to streamline wellsite and back-office operations’. And also, ‘stop using Excel to store operations data’.
Will Whitley described trials of AI/ML at Oasis Petroleum. The project evaluated thousands of sucker rod pump cards to obtain an ML algorithm that could distinguish between fluid pound and gas interference and to spot conditions such as tagging. Agora (a Schlumberger unit) helped build the AI model which was successfully deployed on a 6 well trial before scaling up to a 74 well pilot.
Alan Bryant (Oxy and ISA 112 committee) provided a progress report from the new ISA 112 standard that is being developed for scada systems. There is currently no good guidance out there for building and operating a scada system for a diverse equipment set. A common terminology and reference architecture is needed. Work began in 2016. In 2018 a ISA112 lifecycle and architecture diagram was released. Currently a first draft of the standard is receiving comments and publication is planned, well, sometime this decade! The process appears somewhat laborious with input from 100 members across multiple verticals including oil and gas. An initial 5 layer architecture was proposed. After consultation and debate this now stands at 11 layers. It could be that the slow pace of development has meant that technology is overtaking the initiative although ‘a subcommittee is working on an IIoT version of the diagram’. More from ISA 112.
We have not covered all the presentations made at Wellsite Automation. Here are some other vendor solutions that merit a mention as follows…
More from the ABC conference website.
© Oil IT Journal - all rights reserved.