kW vs kVA – An Introduction to Power Factor
Why is it in the data centre business, (and in electrical engineering in general) we constantly jump between kilo-Watts (kW) and kilo-Volt-Amperes (kVA). On one hand you have an I.T. load, say 20kW, and on the other, a 30kVA UPS. How is this correct? What is the correlation between the two completely different values? How much I.T. load can I actually connect to my data centre’s UPS?
When it comes to explaining any engineering principle there is both a long, or complete, answer and a short one. You will be delighted to know that here; we will deal in the short version or at least an engineer’s version of a short one. Just keep in mind that there ARE other factors at play here, however this is a basic introduction to power factor.
This blog post is intended for the layperson, in order to put into perspective, the terms banded around by their data centre designers and installation engineers.
The Nuts and Bolts
I would, however, be remiss in my duties as an engineer not to give SOME FORM of background. In terms of electrical power (measured in Watts) there is Active & Apparent power. Active power (or REAL power) is essentially the “useful” power distributed throughout an electrical network. This is the power (Watts) used to satiate your electrical equipment’s need for electricity in order for it to operate. Apparent power (Volt-Amperes) is the same amount of active power, but includes the amount of “effort” it takes to get that power around the network. That “effort” is as a result of the resistive, capacitive and inductive properties of a circuit, the additional power required to build up magnetic fields to engage an electrical process. The inductance and capacitance combined are known as Reactance, with Reactance and Resistance known as Impedance. Power Factor is always denoted as a value between 0.0 and 1.0, and is a ratio of the Real Power (kW) to the total Apparent Power (kVA).
So how else can I explain this?
Below we have a diagram showing the relation between Real, Apparent and Reactive power, and where power factor sits between all three. You may even remember the calculation above from those trigonometry classes from back in the day. Though, if you are like me, you like a more real world comparison, so let’s get the 13:00 train from Paddington to Bristol Temple Meads.
Here we are on the train, speeding through the countryside. The track is flat and level, but the train will encounter resistance (i.e. wind). In terms of power factor, this is known as Unity power factor, or a value of 1.0. In this case, there is no reactance. Here, you only encounter resistive losses.
A short while into the journey however the train has to travel uphill. In order to cover the same amount of distance in the same amount of time as it would on a flat track, the engine will need to do additional work. In fundamental terms, the higher the incline, the more work is required to travel the same distance. In Power this “incline” is power factor, and the additional “work” required is to power its own electrical processes. In electrical terms, if the Real Power requirement of the load is 80kW (I.T. equipment) and the power factor of the circuit is 0.8 (output side of the UPS), then the Supply requirement is 100kVA (size requirement of the UPS).
What does that mean for you the Data Centre or IT Manager?
In essence, the lower a power factor (a figure less than 1.0) any piece of electrical equipment has, the less “usable” energy you get out than you put in. As in our example above, a 100kVA UPS @ 0.8pf will deliver 80kW, however, a 100kVA UPS @ 0.9pf (closer to unity) will deliver 90kW (per se). Other factors such as efficiencies at particular % loadings are also involved, as well as the power factor of the load but that is a lesson for another day.
Power Factor is not exclusive to UPS’s however, it is present in some form or another in all electrical equipment that requires a magnetic field to operate, be they Generators/Motors, Pumps or Fans. It is common for some facilities to install Power Factor Correction units in parallel to a reactive load to ensure supply cables do not need to be oversized to allow for the reactive load.
Now that we have covered the bare essentials on power factor, my next blog will cover:
“How you can calculate how much power is needed through your data centre’s electrical infrastructure if the power rating of your IT equipment is in kW, kVA, or Amps?”
Within the following blog I will tackle Nameplate Ratings and dual power supply servers. Both of which, when misunderstood, result in data centre power systems being oversized.
Keep the following in mind,
- The closer a Power Factor is to 1.0 the better. It results in smaller cables and through other factors, reduced losses. There was a time when the Electrical Utility would charge you for poor power factor, and while they still will for very bad power factor, they don’t actively measure your power factor.
- Most manufacturers publish their UPS’s ratings in kVA, however some have now begun publishing them in kW rating for ease of reference, but also because they claim their UPS’s are hair-thin close to unity. It is still worth enquiring what their units Power Factor is.
- Efficiency is not the same as Power Factor. Power Factor is the same no matter what the load is, while efficiency can change given a particular load.
- A clamp meter will measure current “flowing” in a cable, but will not give you the power factor, so it is a reading of “x” amps Apparent and does not give kW.
- The following calculation shows the ratio: Power Factor (pf)= (Watts (W))/(Volt-Amperes(VA))