Page 25 - MASHRAE 35th Anniversary

This article was published in ASHRAE Journal, October 2020. Copyright 2020 ASHRAE. Posted at www.ashrae.org. This article may not be copied and/or distributed electronically or in paper form without permission of ASHRAE. For more information about ASHRAE Journal, visit www.ashrae.org.
means deploying real-time calibration tools are possible.
Whether a manual or automated method of calibration is used, the digital twin must be continuously maintained and validated to represent the real-world conditions within the data center. This is a fundamental precept if the viability of preestablished use cases is to be realized. Ultimately the cost of the calibration efforts must be considered as part of the ROI of the digital twin.
Digital Twin Ongoing Commissioning
Digital twins are functional beyond the “what if” modelling scenarios, particularly when considering real-time calibration methods. Digital twins can serve as ongoing commissioning tools to fine-tune and update a data center’s control system and HVAC over the life of operations as well.
Changes in IT loads often necessitate changes in the HVAC and control system settings and sequences of operations. The necessity of these updates is frequently ignored until some failure mode occurs or an audit is conducted.
Today’s commissioning processes are largely manual and are typically initiated when major changes to infrastructure occur, such as power or cooling system changes. However, rarely is the commissioning process initiated by major changes to IT infrastructure alone. Examples where IT activities should initiate the com- missioning process may include:
• IT hardware additions or removals;
• IT hardware relocations;
• IT hardware density changes;
• IT software modifications (higher CPU
utilizations, etc.); and
• IT software modifications (change of
software on the servers, etc.).
The above changes can have profound implications on the loads profile or density distribution within a data center and could go completely unnoticed without a tool that continuously monitors these trends. Typical control systems are reactionary to modifications in the system operations, not predictive. Even with artificial intelligence of control systems data sets, the results are historical trend analysis of what has occurred, not predictive of what will
occur given major changes to the IT infrastructure.
Digital twins offer the unique ability to perform a vital role in predictive and ongoing commissioning of rapidly changing data center environments. IT infrastructure changes much faster than its brick-and-mortar counter- parts of HVAC and power infrastructure. As noted, often those IT changes occur without changes to the HVAC and power infrastructure.
Further, IT changes can occur at the software layer and not even include a physical and visible change to IT hardware. Such changes can still have a profound change on load densities, load profiles and load distribution within the data center. These “invisible” changes can often materially impact the efficient operations of HVAC and control systems managing the cooling of the data center.
Last, a means of ongoing commissioning can serve a benefit in the current COVID-19 environment. Travel restrictions, recommended reductions in occupancy and an increasingly remote workforce provide a natural benefit to deployment of a virtual environment with off- site management capabilities for digital infrastructure. The digital twin provides an invaluable virtual environment, which enables continued analysis and operations of built infrastructure without the need for having a physical presence on site.
Closing Comments
Digital twins offer the ability to incorporate a new world of predictive data center management through modelling of real-world environmental conditions, operational parameters and continuous ITE change.
Through diligent baselining efforts, a reliable model of the data center environment can be created by cata- logging data from deployed ITE, existing building infra- structure and in-situ power and cooling equipment. Right- sizing the modelling efforts requires appropriately ranged boundary conditions, mesh densities and modelling assumptions to create a tool with the necessary level of detail to model “what if” scenarios that are commonly expected.
Calibration of the digital twin is a necessary and ongoing activity for realizing the full opportunity potential of the changing data