UPS load capacity percentage vs apparent power output

[CHMapc]

Hello,

Last week we had an increase in power usage that caused some of our UPSs to alarm due to overload.

After checking the graphs for load percentage and also the VA output value there seemed to be a discrepancy between the values and what the UPS should be capable of. For instance, in EcoStruxure IT expert, during the power usage spike, the load capacity was about 130%, whereas the apparent power output number was 1.35kVA. The unit in question was a 1500VA unit. My current understanding is that 1350VA would be less than 100% on a 1500VA unit.

Am I possibly missing something with how these sensors track the data that they may not be adding up how I would expect? Are there any other sensors/data points I should be monitoring instead to better keep a UPS from overloading and potentially causing damage to connected devices?

[Cory_McDonald]

@CHMapc,

The values within IT Expert would be polled directly via SNMP from the end device and just displayed.  I believe what you are seeing likely is a different between the "Real Power Rating" and the "Apparent Power Rating".  Here is a screenshot from a Smart-UPS 1500's About > UPS screen that displays the values.  Note that the exact UPS model displayed may be different than what your UPSs model is:

I have transferred this forum post over to the APC forum so hopefully someone can provide additional details there on the UPS load ratings.

[Teken · Accepted solution]

As a general rule and best practices the system load should not exceed 80% of its rated capacity.

You may refer to this APC Resource Link:

https://www.apc.com/ca/en/faqs/FAQ000268376/

As it relates to what you see vs what is actually happening this is primarily a polling / update interval problem.

You have 3-4 different factors to take into consideration such as what does the EcoStructure IT service provide in terms of polling interval. 

Then, you have the interval for SNMP etc.

Based on the limited information you have provided you’re drawing 90% of the units capacity. 

This leaves no margin (overhead) for variable loads that will fluctuate based on temperature, demand, and line voltage etc.

IE.

Temperature: Various devices become warmer as such their internal fans speed up increasing power draw / current.

Processors: As demands increase the various CPU will draw more power / current.

Voltage: When ever the system engages the AVR circuit the unit itself will draw more power / current. Devices that are highly regulated will also increase their power draw / current as when voltage drops current rises.

Anything you see on the LCD, NMC, SNMP, etc in terms of values are simply averaged, delayed, and not very accurate in terms of power / current draw.

Any serious site will have a power quality analyzer and many other tools. The individual circuits (breakers) would be monitored by a energy monitor to provide real time and accurate measurements and data.

Having these important tools provides critical insight, long term data, and alerts you of potential problems not apparent.

The real world options are to remove some of the loads and place them on another UPS device. The other is to install a larger 2-3 KVA UPS that has the ability and headroom for the loads you have now. 

Choosing which path to follow comes down to finances and how critical this site is and what standards are in place / enforced.

Questions Ask . . . 👍

[CHMapc]

Thank you for the reply, this was helpful to understand the limits of the built in monitoring tools.

I do wonder though, in the case of a load that may not be super critical, is there any leeway in letting a UPS overload and accepting the risks that it may trip or damage the UPS? What are the risks in regard to solely the output load devices.

For instance, in this scenario, the UPS holds a number of end user devices. The UPS usually runs at 50-60% load. Only for about 30 mins a week does it exceed 90% and sometimes hit 120% during updates. If the UPS cut power to the devices because of overload I would not be as upset as if it were to cause damage to the output devices. Obviously, based on what you explained, the right answer is to add more UPS unit capacity, but in a scenario where that is difficult, I'm curious what the realistic risks are.

[Teken]

At a high level heat is the primary killer of electronics. As such when components heat up they operate in the extreme regions of their manufactured tolerances.

Heat causes capacitors to dry out and resistors to change values. When the internal temperature rises this same ambient heat transfers over to the battery compartment.

Heat is one of the primary killers of batteries and reduces service life and performance. 

IE. This device will require more battery replacement which is a added cost. More often than not the batteries will swell, deform, which makes it very difficult to remove besides the obvious risk off gassing.

Based on your reply it sounds like the loads are computers?!? If so allowing the UPS to hard shutdown or terminate the output to a running computer is never a good idea.

Operating System corruption is and loss of user data would be the outcome. As you stated if the various computers are receiving a software / firmware update and the UPS should abruptly power down the end device(s) are probably going to be bricked / corrupted due to a sudden loss of power.

From a safety perspective all electrical circuits are rated for 80% ampacity.

As such overloading the actual 120 VAC 15/20 amp wiring is a fire hazard waiting to happen and is also illegal, none compliance for any insurance / certification such as ISO etc.

At the end of the day allowing the UPS to operate in the extreme regions will cause unexpected operations, behaviour, and damage.

I’m not sure how your clients are going to feel once they learn their data is corrupted or gone?!? Anything that’s related to the network infrastructure such as a switch, router, firewall, etc.

Obviously won’t be online and working during this unexpected shutdown.

Questions Ask . . . 👍