David Smethurst’s (Hitachi Ventura) presentation encapsulated the promise and the hype of the IoT. Seemingly, ‘Capital markets now … routinely ask energy companies what they’re doing to prepare for digital’. The 2014 collapse in the oil price meant that ‘many oil and gas and service companies were unable to invest in business improvement’ and the industry is ‘well behind in figuring out how to leverage digital innovations’. The next five years will see the cloud storing the new flood of data, enabling new disruptive business models and providing the foundation for other digital innovations. ERP systems will continue to provide the commercial underpinnings for the industry, while artificial intelligence will ‘read and interpret all the data’, supporting key human decision-making functions. Sensor technologies and the internet of things will unlock remote asset monitoring and maintenance and process efficiencies, while generating ‘vast quantities of data to store and analyze’.
But what exactly is the IoT? Is it new? Is it real? Eric Neason (Detechtion) presented on ‘Accelerating asset performance management (APM) with the IoT’. Citing various sources, Neason has it that the IoT involves ‘machines, computers and people enabling intelligent industrial operations using advanced data analytics for transformational business outcomes.’ But no, it’s not exactly new. The IoT has evolved out of prior art including the PLC, M2M devices, the Ethernet, the Internet and the cloud. The IoT can be thought of as ‘the next generation of scada’. IoT connectivity benefits from the expansion of cellular networks including private LTE networks and emerging protocols like 5G. The oil and gas flavor of the IoT is different from other industries as it often operates in remote, sparsely populated areas with multiple participants across the value chain. The ubiquitous connectivity of the IoT allows for situational awareness of remote assets and enables optimal servicing of ‘underserved’ assets such as gas compressors. Notwithstanding the potential benefits, Neason warns that ‘74% of surveyed companies report that IoT initiatives were not successful’ while 60% reported that they ‘looked good on paper but proved to be more complex than expected’. APM success means ‘starting with a business problem, defining the finish line and creating a roadmap’.
Blaine Mathieu presented Vantiq’s development environment for IoT connected applications. Vantiq claims a significant speedup in development and reduction in code using its technology over ‘vanilla’ IoT/cloud app development. One satisfied user is Thierry Baron who presented Total’s TADI (Total anomaly detection initiative) at the 2019 Vantiq GPS user group. TADI uses Vantiq to perform early detection of equipment failure or gas leaks using next generation sensors and real-time data analysis. TADI was developed at Total’s decommissioned Lacq gas field, an EU Seveso 3-regulated facility which can reproduce large gas leak flow rates, from 0.5 g/s to 300 g/s. The technology is now deployed in Total’s ‘Operations Center of the Future’ testbed. A demonstrator unit is planned for delivery in 2021. Watch Baron’s GPS talk on Youtube.
Mike Boudreaux from Emerson’s Roxar unit explained how its 2600 MPFM (multiphase flowmeter), touted as a replacement for the test separator, has been connected to the Microsoft Azure cloud via a ‘secure first mile’ using the Azure IoT Edge gateway and Modbus connectivity. Microsoft’s Bobby Lee also presented a use case involving edge-deployed pattern recognition to determine pump condition in remote locations. Lee observed that ‘continuously inspecting thousands of dyno cards individually can be costly’. The IoT solution can detect pump issues at scale and in real time. If necessary, a pump can be stopped, and field technicians alerted. More from the project minisite.
We already reported on what was then Anadarko’s ML/AI project. Since then, Anadarko has been acquired by Oxy whose Dingzhou Cao provided more chapter and verse on the flagship project, carried out with help from IPCOS and Apex Systems. As we reported previously, Cao’s team is using a spectrum of ML/AI tools to derive real time drilling information from WITS0 and WITSML data streaming from the wellsite. One key function was to correctly identify drilling states and change points in directional drilling. A dataset of 10 rotary steerable system wells and 21 mud motor wells was used to build the change point detection algorithm. This was developed by converting time series data to an image and using pattern recognition technology (UNet, ResNet and transfer learning). The system proved ‘highly accurate’ with a 99.93% success rate.
Paul Bonner presented Apergy Spotlight, a real-time, event-driven application for oil and gas. Spotlight combines IoT, edge computing and automated analysis and machine learning. One use case is continuous monitoring of high-speed engines and reciprocating compressors to predict the onset of failures and enable timely maintenance. Spotlight is a Class I Div. II/IP67 add-on monitor for industrial hardware along with an edge controller and wireless gateway. Successful analytics requires domain expertise in compressors and good monitoring with the right data points at the right frequency. In this application Spotlight provides pressures, temperatures, crankshaft position and crosshead vibration for every degree of crankshaft rotation. Spotlight analytics predicts valve leaks, piston ring leakage, loss of rod reversal and other issues. The system has been trained on many compressors with a variety of features. Some 14 features were used to build the valve leak model. ‘Feature engineering’, leveraging domain knowledge, composes features into explainable models.
Michael Lewis cited IoT use cases in Chevron as monitoring process control, information gathering from connected assets and data analytics. IoT opportunities include predictive maintenance to help prevent unplanned outages and reduce number of scheduled repairs, optimizing transportation schedules and improve safety by spotting worker fatigue. IoT data is not used directly within process control networks because of the susceptibility of IoT devices to attack. IoT devices may lack standard cybersecurity solutions, they may be insecurely designed or expose a complex architecture that is hard to secure. Such vulnerabilities are magnified by sheer numbers of devices. Protecting Chevron’s extensive networks, from wells to gas stations, is an exercise in risk management, with protection that is appropriate to the intended use. Preventing a catastrophic event is key. IoT sensors can only read process control equipment data sources, preventing denial of service attacks. IoT is a peer to peer network where everything, from computers, cell phones and tablets, to monitors, windows, light bulbs, cars, watches is networked and capable of communicating to each other. Threats, either malicious or accidental, may exploit vulnerabilities or other aspects to cause loss events. Lewis gave a pointer to the NIST Cybersecurity Framework and work by NCCOE on IoT control selection. For Chevron, segregation of the process data network is the principal control as it precludes a compromised IoT from affecting the process network.
Ametek’s Skybitz commercial telematics provides real-time information on the location and status of assets. SkyBitz delivers end-to-end solutions for enterprise and local fleets, tank monitoring and petroleum logistics. Truck monitoring optimizes truck visits and provides visibility across assets.
Appian’s ‘low code’ platform for modernizing enterprise and operational applications connects field and device data to the front office for preventative maintenance, incident management pipeline inspection and safety systems. More from Appian.
Hawkiiii low cost, low energy wireless solutions for rod pumped wells.
Neudesic Insights analytic framework for merging AI and IoT. A unified data platform enabling load, store, analyze and retain knowledge. Clients include BP and Hess.
Onica’s ‘IoTanium’ rapid prototype board offers multiple pre-integrated connectivity options, including Wi Fi, BLE and LTE and exposed contacts for easy prototyping. Data can link to AI/ML analytics in the AWS cloud. One application showed a bespoke downhole sensor, feeding MQTT data to a Bluetooth surface device and on into the AWS Greengrass gateway.
Swim DataFabric is an open source platform for building data-driven applications. DataFabric replaces a bare bones Hadoop/Spark programming environment and is used ‘by supermajors’ for OT cyber monitoring and refinery production optimization.
Most futurologists only look a shortish time into the future and see stuff that is rather similar to what we already have! Clariant’s Paul Gould was more adventurous, imagining the oil and gas business in 2035 while acknowledging that ‘predictions are almost always wrong!’ As an oil country chemicals provider, Clariant sees oil operations of the future as centered on unconventionals, with intensive development of the factory drilling (and fracking) paradigm. The future will see autonomous drilling rigs, fracking operations and completions. New ‘micro fracking’ will target micro formations with precise lifecycle recovery plans ‘delivered with AI’. Giant well clusters of hundreds of wells drilled in a circular pattern will extend with 4 to 5 mile laterals into multiple production zones. ‘Super depots’ will service the clusters with collaborating robots operating 24 x 7. This means that the graduates of 2025 will need to be 40% robotics engineers. The workforce will be under one third of today’s. There will be ‘many, many more small, reliable low-cost sensors’ replacing drones. Scada suppliers will need to transform into AI and Robotics software companies. Due to the high density of wells, well life will be very short, but will yield greater production. Infill wells will no longer be needed. There will be only a few ‘mega sized’ operators with very few mid-sized and small operators. Lifting costs will be down ‘40%-60%’ compared to 2019.
The 2020 ECN IoT in oil and gas will be held at the Hilton Americas in Houston on the 28th and 29th of September 2020. More from the Energy Conference Network.
* The IoT, the internet of things, is referred to by some authors as the IIoT, the industrial internet of things. We use IoT throughout this report as a synonym for the IIoT.
© Oil IT Journal - all rights reserved.