Power Quality Analyzer Location

[GMIEE]

Where is the proper place to install a power quality analyzer? I know it depends on what you are looking for, but the facility currently has no problems. It is a building that houses a data center and offices. It has with a single service feeder that comes into a main distribution panel (480) and then downstream in another panel has a backup generator that feeds into an ATS, then into a UPS, and on down to the PDU's for the servers.

The facility would like a power quality analysis done simply to assure them that their power is OK. I have been leaning toward monitoring both the primary of the UPS and maybe some voltage monitoring at the receptacles at the PDU secondary.

Where is the proper place to monitor this sort of thing?

 

[wirenut1980]

The one thing I would suggest is that you monitor at a PDU, downstream of the UPS, and run a simulation of a utility outage and power switching over to generator to make sure the PDU's do not lose power (if that is the intended response).

If the customer wants more, I would then monitor at the incoming main service, to check utility voltage RMS levels, maybe look at voltage and current THD.

I might also monitor at the generator output to see what voltage and current look like there, again looking at RMS voltage levels, voltage and current THD.

You could also check loading in a few places to make sure nothing is getting close to being overloaded.

 

[76nemo]

Divide and conquer IF there are problems.

 

[Cold Fusion]

The one thing I would suggest is that you monitor at a PDU, downstream of the UPS, and run a simulation of a utility outage and power switching over to generator to make sure the PDU's do not lose power (if that is the intended response).

Having been involved in a few of these monitoring issues on UPS systems, there are a few questions I have asked about the simulations and have never got anything for an answer other than a blank look.

Q: How are you going to simulate the utility failing?

A: Open the cb feeding the UPS.

Q: That puts an open circuit on the UPS input - a high impedance.

A: Yeah and ???

Q: When the utility fails, the voltage goes to zero. That's the definition of a short circuit - a very low impedance on the UPS input. How does opening the UPS feeder simulate the short circuit on the UPS input?

A - silence with a blank look, conversation usually stops here.

cf

 

[wirenut1980]

Having been involved in a few of these monitoring issues on UPS systems, there are a few questions I have asked about the simulations and have never got anything for an answer other than a blank look.

Q: How are you going to simulate the utility failing?

A: Open the cb feeding the UPS.

Q: That puts an open circuit on the UPS input - a high impedance.

A: Yeah and ???

Q: When the utility fails, the voltage goes to zero. That's the definition of a short circuit - a very low impedance on the UPS input. How does opening the UPS feeder simulate the short circuit on the UPS input?

A - silence with a blank look, conversation usually stops here.

cf

I am now giving you an E-blank look.:wink:

I don't follow how when the utility voltage goes to zero, is like a short circuit on the UPS input? Could you explain more?

In my way of thinking, there is still an open circuit where the utility protective device opens. If the point of the fault is out on the utility distribution system, that is still prety high impedance to the UPS input.

 

[charlietuna]

You need to ask your customer just what they want analized? Maybe there concerned about issues with their 480 volt equipment?? The power quality downstream of their UPS system should be monitored by their UPS unit and remain fairly constant under normal power. Power quality needs to be monitored at the service and at the effective branch circuit to determine most problems...

 

[Cold Fusion]

In my way of thinking, there is still an open circuit where the utility protective device opens. If the point of the fault is out on the utility distribution system, that is still prety high impedance to the UPS input.

Well, if the utility fails with a fault, up stream OCPD open, the fault is still connected to the customer.

The easiest way is to show this is: attachment

cf

 

[wirenut1980]

Well, if the utility fails with a fault, up stream OCPD open, the fault is still connected to the customer.

The easiest way is to show this is: attachment

cf

Ok, I see what you are saying now. So the proper way to simulate a utility outage is to probably use the transfer switch controls to switch to generator power, when it is convenient.

I guess I still do not see what the source impedance (open circuit vs. low impedance) has to do with how the UPS operates.

 

[Cold Fusion]

Ok, I see what you are saying now. So the proper way to simulate a utility outage is to probably use the transfer switch controls to switch to generator power, when it is convenient.

I guess I still do not see what the source impedance (open circuit vs. low impedance) has to do with how the UPS operates.

When the utility is on-line, all of load impedances in the facility are sitting there, in parallel, drawing current. If the utility feeder pops open, all of those impedance are still there, still in parallel and now add up to a pretty low impedance.

I'll keep my example resistive, pf = 1. Say a facility draws 1000kW, at 480V. So I ~ 1200A, so effective line to line resistance, is 480/1200 = .4 ohms. Now consider the utility cb pops open. That .4 ohms is still connected line to line and is directly connected to the UPS input.

About 20 years ago I got tasked with evaluating small UPSs ~1000va for an industrial mfg, just starting to get computers, programmable controllers, VFDs. Some of the PLCs were fed from small UPS and they were having occasional PLCs resets during power bumps.

One of the controls guys showed me how well the UPSs worked by pulling the plug and pointing out that the PLC stayed running. I had just finished a seminar on protective relays, and that got me thinking about UV relays and how far you let the voltage drop before you trip the incomer. That was the Eureka moment (or maybe a urea moment:roll

So I tried pulling the UPS plug and then shorting the plug - UPS output stayed right up just fine. humm - maybe too long from pulling the plug to shorting the blades.

So my next test involved a 10' zip cord feeding an inline fuse, feeding a receptacle. I connected a switch from the receptacle side of the fuse to the neutral. Plugged the ups into the receptacle, zip cord into a wall receptacle. Connected a storage scope to the UPS output, and a nominal load - 500W heat lamp. Threw the switch. Fuse blows - and there is a dead short on the UPS input. And there is the cutest little nosedive in the UPS output - little over a half cycle, enough to make a PLC reset.

That test method worked for a 1000va. I think I would find a different method to test a 15kva UPS. The switch gets a little big to handle the 10kA - 20kA SSC:roll:

edit: Hope this makes sense. I'm PA2B.:grin:

cf

 

[ELA]

About 20 years ago I got tasked with evaluating small UPSs ~1000va for an industrial mfg, just starting to get computers, programmable controllers, VFDs. Some of the PLCs were fed from small UPS and they were having occasional PLCs resets during power bumps.

One of the controls guys showed me how well the UPSs worked by pulling the plug and pointing out that the PLC stayed running. I had just finished a seminar on protective relays, and that got me thinking about UV relays and how far you let the voltage drop before you trip the incomer. That was the Eureka moment (or maybe a urea moment:roll

So I tried pulling the UPS plug and then shorting the plug - UPS output stayed right up just fine. humm - maybe too long from pulling the plug to shorting the blades.

So my next test involved a 10' zip cord feeding an inline fuse, feeding a receptacle. I connected a switch from the receptacle side of the fuse to the neutral. Plugged the ups into the receptacle, zip cord into a wall receptacle. Connected a storage scope to the UPS output, and a nominal load - 500W heat lamp. Threw the switch. Fuse blows - and there is a dead short on the UPS input. And there is the cutest little nosedive in the UPS output - little over a half cycle, enough to make a PLC reset.

That test method worked for a 1000va. I think I would find a different method to test a 15kva UPS. The switch gets a little big to handle the 10kA - 20kA SSC:roll:

edit: Hope this makes sense. I'm PA2B.:grin:

cf

Was the dip due to the input inpedance being low or due to the switch over time from line to battery?

Years ago we had issues with certain types of UPS's having a 4msec switch over time causing issues with equipment. It could also be influenced by the phase angle at which you turned off the line.

A think a lot depends upon the type of UPS you are talking about.

 

[Cold Fusion]

Was the dip due to the input inpedance being low or due to the switch over time from line to battery?

I highly suspect yes and yes. As I described, that particular brand only glitched when the input voltage went low as opposed to open circuit.

Years ago we had issues with certain types of UPS's having a 4msec switch over time causing issues with equipment. .

The 4ms we could stand if there wasn't a 1/2 cycle dip along with it.

It could also be influenced by the phase angle at which you turned off the line.

That I don't know. I never looked at the incoming phase angle

A think a lot depends upon the type of UPS you are talking about.

Yes, In this case, I stopped looking/testing as soon as I found a brand that did not do this.


cf

 

[robbietan]

Where is the proper place to install a power quality analyzer? I know it depends on what you are looking for, but the facility currently has no problems. It is a building that houses a data center and offices. It has with a single service feeder that comes into a main distribution panel (480) and then downstream in another panel has a backup generator that feeds into an ATS, then into a UPS, and on down to the PDU's for the servers.

The facility would like a power quality analysis done simply to assure them that their power is OK. I have been leaning toward monitoring both the primary of the UPS and maybe some voltage monitoring at the receptacles at the PDU secondary.

Where is the proper place to monitor this sort of thing?

if the facility wants to check if their utility supply is OK, the best place would be at the point of common coupling - the place where the customer equipment and the utility equipment meet. if the utility supplies the facility at 460 volts, the place to monitor is where you can measure 460V.

if the facility wants to check the power going to their UPS, then place the analyzer between the supply and the UPS