Case Study: Dynamic Heat Map for Level 5 Commissioning

Visualizing airflow is a difficult problem. In data centers, where proper airflow can save energy and protect the assets, it is pivotal to understand how the air is (or is not) moving. Typically, Data Centers have to rely on mathematical modeling, such as a CFD model, to hypothesize potential problems, and predict how the airflow might be in the live data center. While these models are helpful, there are limitations. 

Purkay Labs’ Dynamic Heat Map uses real data, across multiple points in a facility, to show what is actually happening in terms of airflow and heat distribution. In this case study, Purkay Labs will show how the Dynamic Heat Map played a pivotal role in a Level 5 test of a newly commissioned data center. 

Introduction

In April 2019, Purkay Labs managed a series of environmental tests during the Level 5 testing of a newly commissioned data center. During this assessment, Purkay Labs used their tools & software to validate the original CFD, and show what happens to the data center environment during full-power outage. Purkay Labs’ data showed that the Client- a fortune 50 Commercial Bank- to confirm the time needed to be fully operational after a total outage, and which aisles were most at risk. This information is crucial in confirming that the newly built data center was ready to serve its customers. 

The Test:

Purkay Labs deployed twenty four AUDIT-BUDDY stands across the facility. There were three stands per hot and cold aisle, spaced four and a half feet apart. This layout allowed for maximum coverage across the 4000 square foot space. 

Each stand is equipped with three sensors, at 6”, 36”, and 72” inches off the ground to cover the room at multiple heights. Each sensor was set to capture temperature and humidity data every minute for a period of 12 hours, while the Client performed multiple different systems test. 

Isometric Transparent Data Hall with AB.png

AUDIT-BUDDY Placements in the Data Hall

Isometric Transparent Data Hall with map.png

Dynamic Heat Map at 80” Level

Multi-Height Isometric Data Hall.png

Illustration of Heat Maps at Three Different Heights

 

The Client used Purkay Labs’ wireless-enabled AUDIT-BUDDY to view real-time data across the room, in the form of a time-trend graph. This data was used to determine that the room had a baseline environment of 80°F.. During each test, the Client used the time-trend graph to determine when the room reached a baseline temperature, and as a clock to determine when to turn on and off the CRAC Units. 

During the Level 5 Testing, the Client performed a series of systems tests. In this case, Purkay Labs will focus on a single test, where all CRACs were shutdown simultaneously to mimic a complete power failure. The entire test took 36 minutes. 

Time-Trend Graph Test.png

The Data

During the data collection, Purkay Labs generated both a time-trend graph as well as series of Dynamic Heat maps for each test conducted. The following is the example of the heat Dynamic Heat Map that the environmental impact when all CRAC Units were turned off, and then re-started. As you can see, the dynamic map shows how the heat map changes from steady state, to 10 minutes after all the CRAC units are shut off.

Dynamic Heat Map.gif

At 12:59AM, the Client cut power to all the CRAC units. The screenshot below shows the heat distribution across the room at the start of the test, where the average cold aisle inlet temp is 80°F and the average hot aisle temp is 100°F.

Detail Image 1.png

 10 minutes later, at 1:09AM, the power to the CRACs was restored. The following dynamic map shows the impact on the cold and hot aisles, as cold air starts flowing again. Note that steady state has not been reached as yet.

Detail Image 2.png

14 minutes later, the room returned to its baseline temperature, with an average 80°F in the cold aisle and 100°F in the hot aisle. The following dynamic map shows all the aisles returning to base line over that time. During the test, Purkay Labs collected the data for the Dynamic Heat Map, and used their real-time time-trend graph data to monitor the overall environment.

Impact of the Dynamic Heat Map

Mapping the room performance during a complete power failure is a critical enabler to certify cooling system redundancy for a data center. This effectively certifies the N+1 or N+2 rating. Traditionally, commissioning engineers place a few sensors at random points of the data center, in an doc hoc manner, and draw conclusions from those readings.   It is hard to understand how each aisle performs, the direction of airflow and how effective the cooling system is upon recovering. 

With the Dynamic Heat Map, Purkay Lab allows their clients complete visibility on how each and every aisle performs from a cooling standpoint during an emergency. This is achieved in a structured, scientific way of measuring the actual data hall performance, instead of relying on a few sensors or a model which may or may not represent the data hall in its entirety. 

For more information on Purkay Labs’ Dynamic Room Heat Map, please visit https://www.purkaylabs.com/room-heat-map-service