UPS going to Bypass Mode instead of supplying through Batteries. 2

[ArsalanKhan]

Background:

We operate a gas plant with a load of approximately 7 kW. Our power supply is provided by a substation, and we also have a backup generator connected to an Automatic Transfer Switch (ATS). This setup ensures that in the event of a power outage, the ATS switches to the generator. For critical loads, we have two UPS systems with Static Transfer Switches (STS) installed to guarantee uninterrupted power supply.

Issue:

Currently, we are encountering a problem where, during a power outage, the UPS effectively switches to battery power and maintains an uninterrupted supply to the critical load. However, when power is briefly restored and then interrupted again within a span of some seconds, the UPS switches to bypass mode instead of continuing to provide power from the batteries as it is designed to do. Can someone help us identify whether this issue is due to a fault in the UPS or the power supply?

 

[LittleInch]

Can't tell from a distance or why the bypass is not a manual switch?

More details needed. circuit diagram or SLD would help.

IME, you normally run through the UPS at all times other than when you want to service the batteries or there is a fault on the UPS when you "bypass" the protection, normally by a manual switch. but maybe you have it set differently for some reason?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Heaviside1925]

As LI noted, single line or circuit drawing would be helpful.
Make and model of UPSs and STS if you want device specific information.

A few thoughts.
When you say

switches to bypass mode

, do you actually mean the UPS/STS switches to bypass mode? or do you mean it is bypassing the UPS? Bypass mode is typically a protected user selected mode.

There should be a setting in the STS/UPS for retransfer, probably 4-6 seconds. You could extend this time period provided the UPS has sufficient capacity.

 

[cranky108]

Synchronizing sources can be difficult, and maybe that is the issue. That the sources are too far out of alignment (frequency, voltage) to allow a clean transfer.

 

[Heaviside1925]

cranky108 Bring up another possible reason. Your STS/UPS should have parameters it looks for when power is restored to determine if it's acceptable to retransfer.

 

[waross]

First;
Is this a problem or is this something that someone has just noticed and may not understand?
It is theoretically possible for grid power to pass through the UPS and for a Static Transfer Switch to enable the inverter should the grid fail.
Possible but possibly not a good idea.
How does the STS know when the grid has failed?
There must be a finite time without grid power to imply a grid failure.
That very small interval of no power may lead to unanticipated transients being produced when the inverter comes on-line.
There may also be a small phase shift, again leading to transients.
Then there is the matter of re-transfer to the grid and frequency.
The UPS will be crystal controlled at exactly 60 Hz or very close.
The grid is not exactly 60 Hz.
If that is so, how did the old electric clocks keep accurate time with a small synchronous motor?
The average grid frequency over an hour or a day is controlled at exactly 60 Hz, but in the short term and particularly during times of increase or decrease in loading, the frequency will rise or drop slightly.
This is particularly likely during outages and restoration of grid power.
An immediate re-transfer from UPS to grid may easily be out of phase, leading to transients and/or breaker tripping.

So what is actually happening?
A typical UPS is based on OR logic.
The source may be the grid OR the source may be the batteries.
In normal operation the front end rectifiers will be supplying the same energy as the output inverter is demanding.
The batteries will be floating with no power into or out of the batteries.
If the grid fails the batteries seamlessly support the inverter which will continue to supply Uninterrupted Power.
When the grid is restored, the Front end will supply output power as well as power to re-charge the batteries.

So what is happening?
I suspect that the UPS is acting as designed and the indications are being misinterpreted.
What you refer to as bypass indication is more likely an indication that the grid is supplying power and that the matteries are fully charged.
I expect that there will also be indications that the batteries are charging and that the batteries are supplying the power.
A short outage may not drain the batteries enough to trigger a battery charging indication.
Expect some dead-band in the battery indication to for allow for normal float charging.

The solution?
Re-read the instructions.
I expect that the operation is as it should be and the indications are being misinterpreted.)


--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[cranky108]

That said, the grid is no longer held to 60Hz (as it was in the past). Its is within some tolerance, but not very exact.

 

[mparenteau]

Is the UPS operating in "energy savings" mode rather than double conversion? I wonder if there is some issue w/ syncing as previously mentioned. If it is operating in D-C mode, then maybe an issue with the rectifier/inverter?

Mike

 

[waross]

That said, the grid is no longer held to 60Hz (as it was in the past). Its is within some tolerance, but not very exact.

When was that.
While exact frequency control may make an exact frequency possible, previous forms of control inherently incororated so slight frequency drift.
Proportional control depends on some slight frequency drift and load sharing of distributed control is facilitated by slight frequency drifts.
The hourly and daily average frequency may be exactly 60 Hz.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[cranky108]

I don't have an exact date, but the major grid operators no longer keep the frequency at exactly 60Hz. This maybe because it cost them to pay larger plants to increase or decrease output in order to keep the frequency exact, and because of recent demands for lower cost electricity, this was one things of the past that was dropped.

Sort of like gold to backup the dollar. Doesn't happen anymore.

 

[waross]

Cranky, the grid was never held to exactly 60 Hz.
Electric clocks the used a grid reference were accurate because the variations were and are slight and lost cycles are replaced with slight over frequency so that the average frequency over time was always exact.
Since the earliest days of reciprocating steam and fly-ball governors, frequency control has always been by droop.
A change in load results in a slight change in frequency.
The frequency is then corrected to 6o Hz by integral action.
The frequency is then raised slightly to add the lost cycles and make the daily average exactly 60 Hz.
In the past it may have been assumed that the grid was exactly 60 Hz because electric clocks kept perfect time.
The time from day to day was accurate but at any one time the clock may be a part of a second fast or slow.
We can go to;
Control Theory 101 if you wish.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[cranky108]

No need for theory, and don't believe me if you want.

My original comment is that it can be difficult to synchronize two sources.

 

[waross]

My original comment is that it can be difficult to synchronize two sources.

The most difficult to synchronize are often two sources running at exactly the same frequency.
I have experienced exactly that effect.
Two sources at exactly 60 Hz.
They could not, would not synchronize until one source drifted slightly off 60 Hz.

No need for theory, and don't believe me if you want.

Your belief has no bearing or effect on reality.
It was in the late 50's when I first saw demonstrated that the grid was not exactly 60 Hz.


--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[Compositepro]

"the grid was never held to exactly 60 Hz."

What we have here is a failure to communicate.
The grid was held to exactly 60 Hz, based on a daily (or whatever) average.

 

[bacon4life]

Historically in the eastern interconnection when a significant time error had accumulated, the grid operators temporarily adjusted target frequency to either 59.98 Hz or 60.02 Hz for a few hours to fix the time error However, this has a reliability risk because it positions the system closer to the underfrequency and overfrequency tripping points. My understanding was that this reliability concern was the driver for suspending the time error correction.

 

[waross]

The grid still varies very slightly, plus or minus.
There is an site (power-grid-frequency.org) that lists actual grid frequencies from around he world.
The variation now is so slight that the error from 60 Hz is listed in mHz.
The first file that I found was for the North American Western Interconnect for a period in May, 2019.
The first reading:
2019-05-19 02:54:06 3.26219 or 60.00326219 Hz
The reports are second by second.
The highest reading that I saw in a fast look was 50 mHz.
The last reading in the file:
2019-05-25 11:58:16 9.10726 or 60.00910726 Hz.

I suspect that they are now using zero crossing timing to detect frequency errors and control the swing set.
I don't believe that that precision of measurement is possible with the original speed monitoring.
Using zero crossing timing will give two readings per cycle.
By the way the line numbers in the spreadsheet file run from 2 to 551,051.
But still variations, however slight due to load changes.



--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[waross]

Did I do the math correctly?
target frequency to either 59.98 Hz or 60.02 Hz
60.02 would be an error of 0.02 mHz correct?
A frequency error of 20 mHz and 30 mHz, and occasionally 40 mHz was often seen.
At other times the error would be in single digits mHz
With over 1/2 million entries, I did not look at every entry.
Of the entries I looked at about 50 mHz was the greatest error.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[bacon4life]

Within WECC there are several frequency events per year that drop below 59.95 Hz. Dropping to 59.90 Hz happens perhaps once a year.

This presentation shows a recording of WECC event that dropped to 59.85 Hz.

https://www.nerc.com/pa/rrm/Webinars DL/March 31 NERC Call 110916 Freq Event final [Read-Only].pdf

 

[crshears]

The Facts Re: Time Error Correction

see

https://www.nerc.com/comm/OC/Refere...ment_NERC_Balancing_and_Frequency_Control.pdf

[ 2021 document ]

Item: pg. 26, Time Control

In the Eastern Interconnection, time error correction continues to the present; at this time MISO [ Midwest Independent System Operator ] is accountable for seeing to this.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]

 

[waross]

Thanks for the support CR.
Ihave been hoping that you would join in.
That is what I have been trying to explain.
If my explanation was inadequate, that's on me.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!