Inverter Shutdown due to Frequency Shift Fault

[SunEagle]

I am working on a problem that is causing a 1.12 Mwatt PV system to shutdown. There are two 500 Kw inverters supplying 480 volt power to a High School in Nevada. There have been recorded events that show sometimes one and sometimes both of the inverters shutting down due to a fault listed as "frequency shift".

From what I have been able to determine this fault is part of the anti islanding protection to take an inverter off line when it detects an abnormality on the grid it is connected to so it doesn't back-feed a grid that is powered down. While this type of control is normal for a grid tie inverter it seems that an event lasting maybe 4 to 10 cycles seems pretty short to take the action of a complete shutdown. The safety of Utility personnel is important but an event that lasts less than a second and then is gone but still shuts down an inverter seems to be an over reaction.

My question to anyone that is familiar with this type of fault. Exactly what event is happening on the Utility power grid to generate a "frequency shift" and do other inverters shut down almost instantaneously for this type of event?

Vince

[Sunking]

My question to anyone that is familiar with this type of fault. Exactly what event is happening on the Utility power grid to generate a "frequency shift" and do other inverters shut down almost instantaneously for this type of event?

I am very familiar with it. Not so much a safety issue with personnel, but equipment like the utility very expensive generators and transmission equipment. Under frequency trips are the root cause of all wide spread blackouts. Whole grids are shut down. The danger for you is letting out all the magic smoke in your inverters. When there is a difference in phase angle between you and the utility huge amounts of current are flowing between you and the utility and that current is flowing in your inverters. So your inverters are tripping off line to save themselves.

The most likely cause is from the utility frequency drift caused from load changes in their system. If a big load hits their system causes th egenerators at the plant to slow down slightly and it takes a few cycles for the governor in the generator to react and correct for it. On the flip side if the load suddenly decreases the generator speeds up.

[SunEagle]

I am very familiar with it. Not so much a safety issue with personnel, but equipment like the utility very expensive generators and transmission equipment. Under frequency trips are the root cause of all wide spread blackouts. Whole grids are shut down. The danger for you is letting out all the magic smoke in your inverters. When there is a difference in phase angle between you and the utility huge amounts of current are flowing between you and the utility and that current is flowing in your inverters. So your inverters are tripping off line to save themselves.

The most likely cause is from the utility frequency drift caused from load changes in their system. If a big load hits their system causes th egenerators at the plant to slow down slightly and it takes a few cycles for the governor in the generator to react and correct for it. On the flip side if the load suddenly decreases the generator speeds up.

I understand. I thought the issue was more than likely due to something going on at the Utility generation point but I didn't want to rule out anything else. It also makes sense now why even the smallest "freq shift" detected by the Inverter would shut it down. It is isolating from the grid to protect itself.

Would you know of any type of equipment or filter that may mitigate these "freq shifts" long enough to keep the inverter from shutting down yet still protect it?

Thank you for your input. I appreciate it.

[inetdog]

I understand. I thought the issue was more than likely due to something going on at the Utility generation point but I didn't want to rule out anything else. It also makes sense now why even the smallest "freq shift" detected by the Inverter would shut it down. It is isolating from the grid to protect itself.

Would you know of any type of equipment or filter that may mitigate these "freq shifts" long enough to keep the inverter from shutting down yet still protect it?

Thank you for your input. I appreciate it.

The anti-islanding (a-i) circuitry has to be able to detect not only a complete disconnect of utility power and out of spec utility power but also a "resonant" local load combination that can consume the power generated by the GTI while appearing to provide the stable external waveform reference that the GTI requires. I suspect that this type of edge case is what leads the inverter designers to put such a tight short-term tolerance on the frequency stability.
In some cases, like yours apparently, the utility is not meeting the expected stability for the grid as designed into the the a-i circuitry. Any changes in this tolerance would have to be engineered and also tested to make sure that the design requirements for a-i protection are still met.
You may be caught between a rock (GTI design) and a hard place (POCO) unless you can demonstrate by power quality monitoring that the POCO has not met its own stability requirements.

As for "filtering" the power grid frequency, I do not know of anything that would be capable of that other than an asynchronous (written pole) motor-geneartor set or equivalent. And that would cost something similar to the inverters themselves.

When the GTIs shut down is it affecting power to the school (which it should NOT for a viable design), or is it just the loss of income that is a concern?

[SunEagle]

The anti-islanding (a-i) circuitry has to be able to detect not only a complete disconnect of utility power and out of spec utility power but also a "resonant" local load combination that can consume the power generated by the GTI while appearing to provide the stable external waveform reference that the GTI requires. I suspect that this type of edge case is what leads the inverter designers to put such a tight short-term tolerance on the frequency stability.
In some cases, like yours apparently, the utility is not meeting the expected stability for the grid as designed into the the a-i circuitry. Any changes in this tolerance would have to be engineered and also tested to make sure that the design requirements for a-i protection are still met.
You may be caught between a rock (GTI design) and a hard place (POCO) unless you can demonstrate by power quality monitoring that the POCO has not met its own stability requirements.

As for "filtering" the power grid frequency, I do not know of anything that would be capable of that other than an asynchronous (written pole) motor-geneartor set or equivalent. And that would cost something similar to the inverters themselves.

When the GTIs shut down is it affecting power to the school (which it should NOT for a viable design), or is it just the loss of income that is a concern?

I have not found any evidence that the School's power quality is affected. The shutdown of the inverters are affecting the billing. When they shut down the school has to purchase the power at about $0.053/kWh but without the PV System to offset the load it may cause the Kw demand penalty to increase which is about $9.5/kw.

Thanks for your input.

[Naptown]

When they go off line do they start up automatically again or do they have to be manually started backup?
I would connect a recording scope to the power lines and see what is happening. If it is enough phase shift to cause the inverters to shut down and it is the POCO at fault I don't see why they have to pay a demand penalty. Well that's my story and i'm sticking to it.

[SunEagle]

When they go off line do they start up automatically again or do they have to be manually started backup?
I would connect a recording scope to the power lines and see what is happening. If it is enough phase shift to cause the inverters to shut down and it is the POCO at fault I don't see why they have to pay a demand penalty. Well that's my story and i'm sticking to it.

This particular fault is non-latching. The Inverters are programmed to check the "grid" power quality after 5 minutes and if all is ok they go into their start-up sequence and then start producing.

I have a number of power quality files that recorded numerous "outages" of the inverters on certain day. The inverters have internal logging capabilities which captures all of the pertinent data and faults at 1 minute increments. The fault that is of interest is the "frequency shift".

The demand penalty for the month is based on the highest kw used on any day of the billing period. If the day the inverters go down happens to be the highest kw usage then the kw demand can increase by the total amount the PV System is supplying. The Max AC kw rating of this system is about 954 kw.

[inetdog]

This particular fault is non-latching. The Inverters are programmed to check the "grid" power quality after 5 minutes and if all is ok they go into their start-up sequence and then start producing.

I have a number of power quality files that recorded numerous "outages" of the inverters on certain day. The inverters have internal logging capabilities which captures all of the pertinent data and faults at 1 minute increments. The fault that is of interest is the "frequency shift".

The demand penalty for the month is based on the highest kw used on any day of the billing period. If the day the inverters go down happens to be the highest kw usage then the kw demand can increase by the total amount the PV System is supplying. The Max AC kw rating of this system is about 954 kw.

Look into renting or contracting for a power quality monitor which can record the actual supply side waveforms at the time of the fault instead of being limited to the inverter's own characterization of the event. Without that, you do not have a strong case to argue against POCO. If monitoring would cost more than the potential savings, see what you can "negotiate" with POCO.
It looks like the potential penalty could be as much as $9000 per month if the trip happens at the worst possible time.

[SunEagle]

Look into renting or contracting for a power quality monitor which can record the actual supply side waveforms at the time of the fault instead of being limited to the inverter's own characterization of the event. Without that, you do not have a strong case to argue against POCO. If monitoring would cost more than the potential savings, see what you can "negotiate" with POCO.
It looks like the potential penalty could be as much as $9000 per month if the trip happens at the worst possible time.

Already working with a Contractor that has a quality power monitor as well as with POCO to set up their own metering. They don't like the idea of having to pick up an additional Megawatt of load when the School's system goes down. They want to correct the issue as much as I do.

Thanks for your tips and suggestions. I appreciate it.

[SunEagle]

Latest update on my issue with Inverters shutting down unexpectedly. The data from the Inverter logs indicated a "Frequency Shift" fault. Sometimes they were occurring 12 or more times a day and sometime the inverters would not see a fault like this for days.

We determined that due to fluctuations in the Utilities Power Grid the phase angle was exceeding the capability of the inverters to maintain synchronization. While we are still trying to find exactly what is causing the Grid to fluctuate (I have some pretty good ideas but would rather not throw any blame without proof) we believe we have found a way to help reduce the number of faults.

We had the Inverter Manufacturer come out and upgrade the firmware which increased the sensing time for the Inverter to react to this issue from 4 cycles to closer to 20 cycles. Really nothing more than increasing a time delay before reacting to an issue.

Since the firmware upgrade the issue seems to have gone away due to not having the fault for weeks but I will continue to monitor the system and let you know if anything else raises it's ugly head.

I appreciate your support and comments. Thanks

Oh. I tried to upload some photos of this 1.12 megawatt system but the files were too big and not accepted. The files are huge just like the solar arrays. It is made up of almost 4900 x 235 watt panels spread out across 5 acres.

Vince

[inetdog]

We determined that due to fluctuations in the Utilities Power Grid the phase angle was exceeding the capability of the inverters to maintain synchronization.

If I were POCO, I would seriously consider refunding some load-based surcharges to the school.
Wellll, maybe not, but if I were the school I would at least ask for them.

[PowerSupplyGuy]

If I were POCO, I would seriously consider refunding some load-based surcharges to the school.
Wellll, maybe not, but if I were the school I would at least ask for them.

I suppose phase variations are within the spec so you don't really have a claim. Anyway, an interesting problem and I would like to know what you find. I have seen problems in the past when utility hangs power factor correction caps on the line. Puts a bunch of garbage on the line for a few cycles that would be difficult for a slow inverter to track. We used a Dranetz power monitor. If you have access to a decent DSO and can pre-trigger on an alarm signal that the inverter has then that might be a good way around. Just make sure you are using the correct voltage probes.