The main factors affecting data centre profitability include location, connectivity, demand for data services, and operational efficiency. Scale and automation obviously affect operational efficiency, but with energy efficiency as a major driver, this Blog takes a closer look at the critically important, but sometimes overlooked, role of measurement accuracy.

Why measurement accuracy matters

Governments and energy companies are constantly reminding consumers that even small adjustments on their thermostats can save large amounts of money. The same applies to data centres, but on a much larger scale! But for data centres it’s not as simple as dialling down the temperature by half a degree.

Falsely high temperature readings can lead to overcooling, resulting in excessive energy consumption and emissions. Falsely low temperature readings can lead to undercooling, increasing the risk of IT downtime. Similarly, incorrect humidity readings, high or low, can result in damage to IT equipment.

The IT infrastructure at datacentres is highly valuable, but arguably the data that it manages is even more valuable, so it is vital that the internal environment of a datacentre is maintained efficiently and that it provides optimal conditions for IT performance. Sensor performance is therefore critical.

The cost of inaccuracy

Data centre managers do not need reminding about the costs of downtime. However, downtime is the worst-case scenario – energy efficiency is all about ongoing, effective monitoring and control. 

In a recent study, Vaisala found that overcooling a data centre by a mere 0.5 °C can significantly increase cooling energy costs. For a 10 MW data centre this translates into more than €730,000 in additional energy costs over ten years. Low accuracy of, for example, ±0.5–1.0 °C could cost operators tens of thousands of euros annually and lead to tons of unnecessary greenhouse gas emissions. This can be avoided by using instruments with a high accuracy of, say, ±0.1 °C. Vaisala Origo’s precise ±0.1 °C and ±1 %RH accuracy and stable measurements reduce unnecessary cooling while ensuring the reliable environmental control that critical facilities depend on.

Quality sensors, and what they look like

It is certainly true that data centres need a lot of sensors, but sensor cost is negligible in comparison with the energy costs that they save, and with the value of the assets that they protect. So here are 5 quality measures to look for in sensors:

• Nominal Accuracy 

Choosing instruments with the highest nominal accuracy is the first step to controlling cooling precisely and reducing energy costs. Vaisala measurement instruments provide industry-leading accuracy of as high as ±0 .1 °C for temperature and ±0 .8% RH for relative humidity.

1. Stability
   This is a sensor’s ability to maintain accuracy over time without drifting. Without this, a great deal of time has to be spent on recalibration, so stability is a key design feature of Vaisala probes. Vaisala’s humidity sensors for example, feature HUMICAP® technology, which provides high performance, chemical tolerance, and stability.
2. Calibration Interval
   An instrument’s calibration interval tells you how frequently it needs to be recalibrated to compensate for drift. Longer calibration intervals minimise interruptions and lower maintenance costs.
   
3. Calibration method
   Calibration is necessary to check that a sensor is performing to specification. In some cases, a simple field check/calibration is sufficient, while some instruments or applications require an off-site laboratory calibration. Conscious of this requirement, Vaisala’s low-drift probes are quick and simple to swap out, and certified instruments are available for field calibration checks.
4. Durability
   The repair and maintenance of data centre instruments is costly and time-consuming. In addition, confidentiality regulations can restrict third-party access to the premises. Vaisala’s probes are therefore designed for long-term, reliable deployment in the harshest possible conditions. For example, Vaisala sensors are onboard the Mars Curiosity and Perseverance Rovers and have been operating continuously for many years without disruption.

Summary

The cost of good quality sensors is negligible in comparison with the cost of inaccuracy – both financially and from a sustainability perspective. Datacentre managers looking for energy efficiency and and PUE (power usage effectiveness) are therefore encouraged to invest in high quality sensors – to save time, money and greenhouse gas emissions.

For more information on how to reduce cooling costs with improved measurement accuracy, register for Vaisala’s webinar: How much can precise HVAC measurements influence energy consumption of cooling systems?

——— 

This post was written by Anu Kätkä, Vaisala Product Line Manager.