Remove Modeling Headaches with Black Box Cabinets

Setting up CFD models can seem like a time intensive exercise, particularly when the user is forced to include information they don’t know, or don’t care about. One such example is detailed IT inventories for Data Center designers and Co-Location operators. The black box cabinet object in Future Facilities’ 6SigmaDCX aims to remove these modelling headaches.

For data center designers and Co-location operators, the lack of detailed inventory for IT cabinets is a shared problem when performing CFD analysis on a data center. In the case of the designer, specifying a kW/per Rack is a useful design simplification, but when IT size, type and placement must still be specified, is it simple enough? Likewise, the Co-location operator may not have access to this information, simply because he cannot open the cabinet door.

A CFD model with black box cabinet objects aims to remove the need to input a high level of detail, while maintaining an accurate simulation of the room as a whole. By removing detail from the cabinet model, significant time can be saved in both setting up the model and performing the CFD computations.

Setting up a Black Box Cabinet

The black box requires only 2 pieces of information about the IT contents: the power load and temperature rise. Further detail can be added by selecting the level of blanking from a pre-defined dropdown and specifying the quadrants in which the IT equipment is installed. All of this can be done on the cabinet itself, meaning the whole thing can be set up in only a small number of mouse-clicks.

Cabinet power and temperature rise are both straightforward concepts. Power is usually a known value, and temperature rise can be taken from measured data or industry standards. Blanking settings, however, can often require a little more thought from the user. Again, the black box cabinet simplifies the matter with its predefined and tested settings, ranging from no blanking at all to a perfectly sealed rack.

How do the results stack up?

Let’s take a look at a single cabinet placed in a simple test chamber. On the left is the traditional cabinet model with generic IT spread evenly across the cabinet; On the right is the black box equivalent. When comparing the two, the streamlines show a very close match between the two simulations.

What about something on a slightly larger scale, such as a simple 6SigmaRoom model using the two cabinets above? Comparing the results at the room level shows a very good match between the two models, and we could also draw the same conclusions about the performance of the room no matter which method we decided to use.

What about time savings?

The table below shows reductions in both seconds per time step as well as iterations performed when using the black box object – making the black box model significantly faster than its counterpart.

What are the limitations?

The purpose of the black box object is to allow the user to reduce the complexity of the model when a more defined inventory is unavailable, cutting solve time. The extent to which this method can be used is dependent on the objective of the analysis, which means it is very model and project specific. It can be said however, that the deviation from an accurate solution when using black box objects increases with room disorder. For instance, if the room is extremely well-arranged, contained and maintained with similar IT and cabinets then the black box method provides a much more accurate solution than an unregulated, randomly loaded, and retrofitted data center with minimal in-rack management. This method is more suited to verification at the facility level, and a more accurate representation would require more refined data.

By: Fred Rhodes, Senior Consultant Engineer

Other Recent Posts

Data Center Operations: Adopting a Proactive Process with the Data Center Digital Twin

Whether you’re just starting out managing data center operations or you’ve been using the same appr…

Read More

Why Data Center Operators Need to Rethink the Rule of Thumb Practice

Many of us may feel we’re winning the battle against over-provisioning  of data center resources, b…

Read More

31 October, 2017

Back to entries