AI-Enabled Digital Twins Optimise Green Hydrogen and Data Centre Cooling Designs

The role of #AI-enabled #digitaltwins in the design, operation and optimisation of both #greenhydrogen #electrolyser setups and #liquidcooled #datacentres cannot be underestimated.

Traditional simulation methods are slowly giving way to innovative tools such as #NVIDIAOmniverse, #SimcenterAmesim and #Modulus, which offer more accurate, efficient and scalable visualisation capabilities.

Market leaders in this space including Entopy, Cadence Design Systems, Siemens and AGILOX plus NVIDIA's recent presentation '𝐃𝐢𝐠𝐢𝐭𝐚𝐥 𝐓𝐰𝐢𝐧𝐬 𝐢𝐧 𝐃𝐚𝐭𝐚 𝐂𝐞𝐧𝐭𝐞𝐫 𝐂𝐨𝐨𝐥𝐢𝐧𝐠: 𝐌𝐞𝐞𝐭𝐢𝐧𝐠 𝐀𝐈 𝐃𝐞𝐦𝐚𝐧𝐝𝐬 𝐰𝐢𝐭𝐡 𝐒𝐦𝐚𝐫𝐭 𝐃𝐞𝐬𝐢𝐠𝐧 𝐚𝐧𝐝 𝐈𝐧𝐧𝐨𝐯𝐚𝐭𝐢𝐨𝐧' is well worth a watch. Kudos to OFFIS - Institute for Information Technology for their continued research supporting this important field.
https://lnkd.in/duc_Ees9

Key steps to consider when creating your visual model include: -

➊ 𝐄𝐬𝐭𝐚𝐛𝐥𝐢𝐬𝐡 𝐚 𝐛𝐚𝐬𝐞𝐥𝐢𝐧𝐞 for your digital twin that can be upgraded over time as new information becomes available, develop a database by integrating real and simulated data sets. For preliminary modelling, even small amounts of data can be useful. Remote monitoring tools linked to existing installations, small-scale prototypes or pilot systems can also be developed to collect data in the real world. Predictive modelling (AI and machine learning algorithms) can also generate additional data based on established parameters and patterns.

➋ 𝐀𝐧𝐚𝐥𝐲𝐬𝐞 𝐯𝐚𝐫𝐢𝐨𝐮𝐬 𝐬𝐜𝐞𝐧𝐚𝐫𝐢𝐨𝐬, i.e. in 𝐞𝐥𝐞𝐜𝐭𝐫𝐨𝐥𝐲𝐬𝐞𝐫𝐬, by combining subsystem variations, such as numbers of solar panels and wind turbines, their corresponding geometries, subcomponent losses, and transient operating conditions, such as power output vs wind speeds, sun position, ambient environmental conditions. In addition to performance vs load variation, energy consumption, reaction rates, daily hydrogen production and storage tank fill rates, etc.

In 𝐥𝐢𝐪𝐮𝐢𝐝 𝐜𝐨𝐨𝐥𝐞𝐝 𝐝𝐚𝐭𝐚 𝐜𝐞𝐧𝐭𝐫𝐞𝐬, simulations can visualise several cooling configurations to understand performance vs key metrics (PUE, etc.) taking in account heightened demand, varying heat loads, ambient weather conditions etc. In addition to the systems' ability to withstand failure situations (such as power loss or cooling system failure).

➌ 𝐎𝐩𝐭𝐢𝐦𝐢𝐬𝐞 your model by modifying design parameters such as system layout, component placement, energy flow based on simulation outcomes. Automated adjustments can be made via AI-driven controls, with the ability to continuously run simulations using the respective parameters to unlock the 𝐨𝐩𝐭𝐢𝐦𝐚𝐥 𝐝𝐞𝐬𝐢𝐠𝐧.

➍ 𝐕𝐚𝐥𝐢𝐝𝐚𝐭𝐞 𝐚𝐧𝐝 𝐃𝐞𝐩𝐥𝐨𝐲. I.e. validate the optimised design through small-scale real-world tests or pilot projects. Once validated, implemented the optimised design in full-scale operations.

Photo by Steve Johnson on Unsplash