Global Business Conferences 2018 IIoT & Digital Solutions for Oil & Gas, Amsterdam

Accenture on ‘compressive disruption’, the ‘wise pivot’ and fighting the corporate ‘immune system’.

Why is everyone talking about digital transformation and disruption today? Because a) the cost of technology is coming down and b) there is a ‘convergence’ of web technology/big data/AI and the internet of things (IoT). Oil and gas is also experiencing ‘compressive disruption’ with pressure on margins coming from a ‘new energy scenario’ of shale, renewables, and ‘demand compression’. The answer is to keep shareholders happy with continuous improvement, to ‘wise pivot’ into future energy transition scenarios and to ‘grow the core’ with predictive analytics, drones and cross-industry consortia (e.g. blockchain in trading). Examples are Shell’s acquisition of NewMotion an e-vehicle charging stations business, BP’s acquisition of NASA AI spin-out Beyond Limits and ‘data monetization’, a key topic in oilfield services. The digital transformation needs support from the top and a courageous CEO. Leideman likens the role of leadership to ‘fighting the immune system’. ‘When you to try something new, incumbents will fight you hard.’ Digital transformation is not just about technology, it needs cultural change and management thereof. On the topic of the proof of concept (PoC), Leideman was nuanced. The PoC is not wrong but it needs to be designed with scaling in mind. Many PoCs stall after their initial success. This has been observed in predictive maintenance where scaling is the biggest issue, as is resistance to change. ‘Everyone knows how stubborn the community is, every asset is different, scaling up is very hard. I see a lot of ‘solutions’ looking for a business problem to solve’.

Advisian - more failures than success in digital transformation.

Bradley Andrews, president of Worley Parson’s Advisian unit has been working on digital transformation internally as well as for clients. In one example, a new program was rolled out that addressed widespread use of Microsoft Excel. This failed because it made five people’s jobs easier and 5,000 people’s jobs harder – and they still use Excel! So what problem is digital transformation trying to solve? Fully-formed technology brought in from other industries is unlikely to be fit for (our) purpose. Advisian interviewed 500 individuals re blockchain, AI, VR and so on. All agree on the intent and speed of the transformation but not on the pathway to success. In fact, there are more failures than success and there is a crisis of confidence in leadership. Another issue is regulatory risk around emerging technology which is slowing adoption. You need to question your beliefs and the hype! ‘Anxiety engenders creativity’.

Digital transformation in Total. Are the GAFAs friend or foe?

For Gilles Cochevelou, digital transformation in Total means working hand in hand with the CIO, ‘this is not shadow IT.’ The transformation spans social networking, mobile and ATAWAD (any time anywhere any device). The magic word, a ‘platform’, is close to evoking a natural monopoly. Are the GAFAs friend or foe? Digital is a great opportunity to break the silos. Cochevelou cited Total’s global roll-out of Microsoft Office365 with its teams functionality (a counterpart to Slack). While this was not an IT program, working with the CIO was critical. Elsewhere Total has many digital initiatives Booster, Total Energy Ventures, Incubator 4.0 and hackathons. Training and digital ‘acculturation’ are achieved with a digital passport, MOOCs, reverse mentoring, a data science challenge and digital bootcamps. Subsurface represents a big digital native community that was doing data science before the day. A joint venture with Google has seen ten Total employees moving to Silicon Valley to partner on image processing and semantic analysis. Total has several drone flying teams using, inter alia SciAero’s CyberQuad. Another trial with Swiss Flyability uses a caged drone to inspect the inside of a refinery distillation column. Predictive maintenance of rotating equipment is a field where many vendors have a ‘magic’ solution - on slideware at least! But Total wants to keep control of its data. Alongside its 20 Smart Room collaborative environments, Total has developed a digital twin of a refinery, the Quantum digital twin/virtual plant. This 3D model acts as a single source of truth throughout the asset’s lifespan. Quantum represents a shift from PDFs to an object/tag-centric approach. This is key to Total’s relationships with vendors and is ‘at heart of the digital transformation’. In retail, Total is working on an e-wallet in a joint venture with SIXT and BMW’s DriveNow unit. Now ‘the car pays for its own gas!’ Total is also working on new sales channels, selling electricity to end users and on its Fioulmarket heating oil sales portal. This means learning new search engine optimization skills and ‘learning to play chess with Google as it changes the rules’.

Microsoft – Azure ML predicts coke explosions.

Microsoft’s Tibor Bacsó opined that although many speak of digital transformation, few (like Total) are actually doing it. Clients question its applicability to a refinery, raising issues such as data quality, the investment involved and resources needed. For Microsoft, it is applicable – and provides a better understanding of thermodynamics, bringing theory and practice together and moving from ‘predictive to prescriptive’ operations. Microsoft Azure machine learning underpins a drag and drop GUI for modeling and predicting coke explosions in a Rumanian refinery, executing an R script. A model can be tested with little investment as capacity is provisioned on the cloud.

Halliburton goes beyond digitalization and into open data future.

Halliburton’s Satyam Priyadarshy went ‘beyond digitalization’ and into true digital transformation in the era of Industry 4.0 and big data. This requires a change of mindset, ‘we still spend a significant amount of time building data models’, doing a project for a few months and dropping it. What is needed is continuous transformation, repurposing your geophysicists to do data science. Despite 60 years of research and 280 published papers, there is still no perfect model for drilling rate of penetration (ROP) prediction. An obstacle to machine learning is folks’ reluctance to share data. E&P leads in this waste of resources! Why do we create multiple copies of data? And we are unprepared for future fiber optical/IoT/cyber/cloud data. Halliburton has established a big data/data science center of excellence. The OpenEarth community (devops), iEnergy (cloud) and DecisionSpace got a plug. One of Priyadarshy’s AI examples goes straight from well data to a reservoir property model, begging the question as to whether AI will cannibalize some of its own software.

Extreme teaming whittles down Shell’s PoCs to feasible, scalable projects close to the core.

Shell chief data officer Anosh Thakkar cited a Gartner definition of digitalization as using digital technology to change a business model and provide new revenues and opportunities. As such, it is not new. But data is growing exponentially and the tools available to extract value from big data are getting better. There is the threat of disruption from new players particularly as oil and gas was ‘late into the digital space’. Currently there is no ‘existential push’ for change but this is changing with AI, blockchain and other novelties. Last year Shell’s digital strategy embraced some 550 proofs of concept, most sans a road map to value. Today, things are being approached coherently across Shell with particular attention to scalability. Shell’s team of 200 digital experts are to focus on digitizing the business core and adjacent areas. Projects must show a scalable minimum viable product within 3 months. Moreover, a project must address feasible technology. One current area of research is equipment failure and deferment issues. On the Shearwater asset, compressor steady state analysis is helping engineers to move from information overload to a consolidated insight into failures, with millions in savings. Shell’s secret sauce is ‘extreme teaming’ in an ‘accelerator room’ where team perform daily sprints and standups, bringing together data scientist, programmers, SMEs and ‘business interaction designers’.

TNO on making sense of big data.

Rahul-Mark Fonseca reported on trials of ‘the largest sensor network in the world’ deployed over the Netherland Groningen gas field. The low cost acoustic sensors are used to assess structural damage and repurposed to use as seismic sensors for interferometry. In another projects, NAM (the Groningen operator) and others are trialing neural nets to forecast oil production from mature assets. While NNs are not new, deep learning is where the progress (and hype!) is. But it is not all easy, a LSTM* auto tunes and works well forecasting some wells, then falls apart on the next one. Another issue is the question as to why NNs work, what is happening inside the black box? Oil and gas engineers are skeptical. Work in DARPA’s XAI Program sets explainability against accuracy. Some attempts to open up the NN box have found signs of cheating. One system correctly identified horses from the (horse) photographer’s name. In another Minority Report-style project with the Netherlands police looking for crooks with police mugshots and LinkedIn data. This worked great but why? It turned out that NN was good at spotting LinkedIn photographs of people wearing T-Shirts, just like the mugshots. Fonsecas advised, ‘Do not forget your domain knowledge!’

* long short term memory network.

Bentley Systems AssetWise as digital transformer.

Alan Kiraly, SVP asset performance with Bentley Systems, cited a Gartner study that found that ‘85% of oils have digitalization initiatives’ but ‘only 10% are being scaled to production’. At the heart of transformation is the digital engineering model, a bridge between IT and OT that ties all data together. Bentley’s solution in this space is AssetWise as deployed on BP’s Omani Khazzan gas project where it is the central information store holding 60,000 documents and 160,000 equipment tags. Other key AssetWise deployments include Shell’s Project Vantage and the US Geismar A04 extension. See also the video of an AssetWise/MindSphere combination, ‘MindApp’, a joint venture with Siemens.

Equinor’s GoDigital program and the OMNIA data platform.

Einar Landre provided an update on Equinor’s GoDigital program and the digital center of excellence he announced last year. Equinor has six programs that frame its digital activity built atop its “OMNIA” data platform. Landre stressed that OMNIA is not “IT digitizing the business,” rather “business driving digital improvement” in a move away from data silos to “enable all data”. OMNIA provides data ingestion and cleanup to consistent naming conventions. Portals and apps for data science, IoT and more run atop the system. OMNIA is built on Microsoft Azure (although it “could have been deployed on any cloud platform”)*. Data from US unconventional operations now flows into the OMNIA cloud. The cloud-based architecture facilitates separation of data and apps. Equinor can now manage data appropriately, work through APIs and micro apps. Operators can now see which wells need attention, reducing driving/accident risk. Equinor is now working on a Digital Twin with Kongsberg to support remote operations and trialing the Hololens and other AR/VR smart devices in a digital worker pilot. “There is huge upside in collaborating on developing technology and architectural principles”.

* We also understand that Landmark’s DecisionSpace Information Server is a key component of the solution.

OMV deploys Intel’s Universal Well Controller.

Intel’s Louis Desroches teamed with OMV’s Sasa Blazekovic to report on a wellhead analytics pilot with OMV in Austria on a brownfield site with around 1,000 stripper wells. The field was originally developed in the 1940s and now covers a 2,400 KM2 area. Surveillance was previously performed with twice weekly site visits. Intel has used an ‘open source approach’ to software that allows OMV to pick and mix sensor types. Alarming and soft sensing software was developed by Ipcos. Other components include Siemens’ Simocode smart motor control and Intel’s IoT gateway. Multi-mode connectivity blends WLAN, 3/4G, LoRa and WiMax. Data from beam pumps now streams into the scada control room and Siemens’ MindSphere for remote software upgrade and central shutdown. Edge analytics support local operators and minimize backhaul bandwidth requirements. Edge devices provide load cell calibration, intrusion detection (camera) and Dynacard-based soft sensing of pump performance. A video-enabled inclinometer also ran as did a near real-time surveillance dashboard/web portal. Closed-loop control start/stops pumps and the pump off controller. The low cost solution (a box costs under $1,000) is now extending to gas lift wells. The solution is now being rolled-up into a cross-vendor universal well controller and mesh network to chat with neighboring wells. The UWC embeds open source software for ‘zero touch’ provisioning and cloud-based management. Intel is working to build an ecosystem of ISV/OEMs. Other components include the Akraino Edge Stack and the Linux Foundation’s ACRN project. Wind River also ran.

Gamified learning for Repsol’s refiners.

Iñigo Ribas Sanguesa presented Repsol’s ‘gamified’ learning for its refinery personnel. Apparently, ‘face to face training is very inefficient’. The ‘Legend of Zelda’ inspired course invites participants to solve refinery enigmas by talking to different characters in the game. The somewhat infantilizing cartoon is said to force team work, pitting a ‘sulfur team’ against a ‘catalyst team’. The solution was developed by VirtualWare. Watch the online demo.

More from Global Business Conferences.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.