Electric grid failure times

Status
Not open for further replies.

tallgirl

Senior Member
Location
Great White North
Occupation
Controls Systems firmware engineer
So, while I was gone I spent a lot of time learning everything there is to know about photovoltaics and started to write software that will monitor and control a PV system.

One thing I learned is that there are a lot more very short power outages than I imagined. Which means, putting in code to notify someone when the power goes out means figuring out how long is "out" before it's really "out".

Does anyone here have any guesses as to how long the power can be out before it should be considered "out"? I see a lot of 1 and 2 second outages, and then they are all much longer -- from a few minutes on up. Would a 5 second outage be a good clue that the power is going to be gone for more than a few more seconds? Anything longer than that and I risk not being able to send an e-mail to a pager because the network infrastructure might fail as well. But too short and someone gets paged every other day.
 
Power outages

Power outages

Good stuff. Are the outages from the PV's or from the grid?

My experience with grid power outages, not including those caused by local faulty equipment, is that they are pretty unpredictable. Some utility protection schemes allow a short pattern of restarts after a high resistance fault, but most where I live rely on a simple fused disconnect.

I have been and am currently involved with backup to critical communications systems. My current recommendation to my customers is to put all critical communications on a UPS, even if it's an off the shelf computer backup from an office supply store.
 

GeorgeB

ElectroHydraulics engineer (retired)
Location
Greenville SC
Occupation
Retired
tallgirl said:
One thing I learned is that there are a lot more very short power outages than I imagined. Which means, putting in code to notify someone when the power goes out means figuring out how long is "out" before it's really "out".

Does anyone here have any guesses as to how long the power can be out before it should be considered "out"? I see a lot of 1 and 2 second outages, and then they are all much longer -- from a few minutes on up. Would a 5 second outage be a good clue that the power is going to be gone for more than a few more seconds? Anything longer than that and I risk not being able to send an e-mail to a pager because the network infrastructure might fail as well. But too short and someone gets paged every other day.
It has been MANY years since I worked in a generating station and saw, but didn't work, with the automatic reclosing on faults. My memory is that unless the lines were being serviced (we blocked the reclosers then), it reclosed after 6 cycles, again after 6 cycles, then a 3rd time after 20-60 cycles. Most faults are lightning caused and the 1st trip is the only one. I used my friend google and found an interesting website that has some info, http://www.basler.com/html/dwntech.htm with a 35 page pdf I could get after "registering" ... deep link here in case it works ... http://www.basler.com/downloads/autorecl.pdf.

From this link and my memory, I would GUESS that the transmission protection will be off for a while if it is off for a second. Depending on the utility and their capabilities, I'd guess they swap from one to another transmission circuits automatically at the distribution level in most cases. I'd guess your 5 seconds is longer than needed, and that 2 or 3 is indicative of longer term problems.
 

tallgirl

Senior Member
Location
Great White North
Occupation
Controls Systems firmware engineer
Thanks for the input.

The code presently has a 30 second timer that starts when the AC input drops (that's the answer to HydeBay1's post). I'll change the code to a 5 second delay, then put some logic in to prevent it from sending too many notifications if there are multiple failures in some short interval.
 

kingpb

Senior Member
Location
SE USA as far as you can go
Occupation
Engineer, Registered
I would think that the programming would notify of a power outage, and that the signal could only be cleared or reset by a tech. That way someone has to physically go look and see what was the cause of failure. Even if it was utility related, surges can other events can cause other problems, and someone would need to look at the historian to see the sequence of events.
 

petersonra

Senior Member
Location
Northern illinois
Occupation
engineer
kingpb said:
I would think that the programming would notify of a power outage, and that the signal could only be cleared or reset by a tech. That way someone has to physically go look and see what was the cause of failure. Even if it was utility related, surges can other events can cause other problems, and someone would need to look at the historian to see the sequence of events.

I'd be more inclined to keep track of all events, but only notify someone if the power outage lasted some period of time - 10 or 15 seconds sounds reasonable. If it is off that long, it will be off a while.
 

tallgirl

Senior Member
Location
Great White North
Occupation
Controls Systems firmware engineer
kingpb said:
I would think that the programming would notify of a power outage, and that the signal could only be cleared or reset by a tech. That way someone has to physically go look and see what was the cause of failure. Even if it was utility related, surges can other events can cause other problems, and someone would need to look at the historian to see the sequence of events.

The risk of requiring manual intervention is that the first failure occurs, the failure resolves itself, the event remains in the triggered state, and when a second failure occurs there is no notification.

The historical data is fairly large, and if there is an outage between the system and wherever the technician is, it would be impossible to examine the system remotely anyway. I'd think it would be best to clear the error after the power is restored and have the technician connect to the system, then look at the error flags and see what was reported and decide from there.
 
Status
Not open for further replies.
Top