Speaking at the 2020 Energistics/Energy Conference Network Summer Digitalization Summit, Richard Zhang, who is software product development lead at Shell Hydrogen explained how Shell is developing a ‘digital backbone’ for its new hydrogen fueling stations in North America, the UK and parts of the EU. The ‘first of a kind’ development includes a fully digital hydrogen supply chain to enable remote monitoring and operations and to facilitate root cause analysis of equipment failures that are to be expected in this early stage supply chain. The platform will also allow cross-validation of vendor data and will help mature partners’ capabilities and verify performance.
Zhang observed that H2 is ‘not yet cool’, and Shell needs to bring customers and users into the discussion to identify potential growth areas and inform corporate strategy. The goals of the digital hydrogen supply chain are not new, and they echo much of the challenges that have been solved in oil and gas, aviation and other verticals. In this space, vendors are typically equipment-focused and may need help with IT and data. Vendors may lack test data and proof of reliability in the field. ‘Establishing a baseline for analytics is very difficult and labor intensive and ML/AI is not the panacea’. There are also challenges with policy and industrial collaboration where ISO standards may be deficient, as they are created through simulation and limited-scale data gathering. There is also a lack of universal, machine and customer-friendly data exchange and collection methods.
Shell is therefore planning to re-use existing solutions from oil and gas as and to get up and running as soon as possible, making sure that the infrastructure is scalable ‘from one to one thousand stations or one to one million customers’.
Shell’s Hydrogen Digital Platform (HDP) is to re-use the platform developed for Shell’s Open Subsurface Data Universe’s infrastructure. ‘OSDU has solved internal IT compliance issue and made external data partnership easy’. OSDU is built on an AWS serverless architecture that supports IoT, streaming data, and data analytics. OSDU has seen few performance issues using standard services that were created to solve 80% of common use cases. This has been achieved by offloading as much infrastructure as possible to AWS with tools such as Kinesis over Kafka or Rabbit MQ.
At the H2 Station, an AWS Greengrass edge device data allows for remote operation of station equipment and real time monitoring. Station PLC data streams to the HDP Cloud for analysis and storage in the AWS S3 data lake. Other tools of the trade include AWS IoT and Kinesis, AWS Sagemaker, IoT Warehouse Event and Amplify. An HDP data catalog (built with AWS Glue, Lambda and Scripts) crawls data sources and maintains relationship among data products. A visualization component leverages Tableau, Grafana, FRACAS and MADe to provide self-service analytics, reliability assessment and reporting.
Zhang believes in letting ‘software engineers take control of their own fate’. Their efficiency has greatly increased in the last few years enabling rapid experimentation and the freedom to use the best available tool. The agile approach has been a ‘learning journey for all the major enterprises’. Today Shell Hydrogen has been able tap into the ecosystem and learn from others.
Specific learnings include the need to check time sync across vendors’ system clocks, ‘still the most important and impactful thing to do before doing any data analytics’. A unique transaction ID across system boundaries is the best way to simply data analytics and cross-validation. Shell is now working to track baseline changes from field maintenance and integrate with reliability engineering software (FRACAS, RCFA). Shell is also seeking industry consensus on data exchanges for carbon track and trace, carbon trading and policy credits verification where Zhang opined ‘blockchain is not a good solution’. More from the Summer Digitization Summit in our next issue and from the conference website.
© Oil IT Journal - all rights reserved.