We have a 2.5MW DG in parallel with a 20MVA transformer (UAT of a 250MW CCPP) (6.6kV). We want to run the 2.5MW DG in continuous parallel with the trafo. In the event of the UAT failing, can the non critical loads be switched off and DG continue to feed the critical loads?

As of now we have provided a df/dt relay which senses the UAT supply. On sensing UAT failure can the contact of this relay be used to trip all breakers of non critical loads and also incomer from UAT?

Is it necessary to ascertain separately whether the DG can withstand the load during the time between sensing of df/dt relay and tripping of the breakers?

Is there anything else to be seen before operating the DG in parallel with UAT. DG Set controller has got the facility for synchronising, limiting load to 2.5MW etc.?

 

sangeetha,

This is a classic application for a fast load shedding system. Depending on what protection relays you presently have installed on your system, it should be possible for you to detect the loss of the UAT, and shed the non critical loads fast enough to ensure your system stays running. Typically this has to be done fast, within 100-200 milliseconds. But, given the small inertia of the 2.5MW machine that you are depending on to keep your system running, you may need to be even faster. In order to find out how fast your application needs to be would require a system study that can find how quickly the frequency will decay during an event, and that will tell you how fast your system need to react. There are a few vendors that can provide you with solutions that can operate and send a command to the breaker within 4-12 milliseconds (the breaker will take an additional 32 to 80 milliseconds to open).

I would recommend reading technical papers on the subject that explain the nature of load shedding systems and what to consider when deciding to implement such a system. You can find free technical papers on the subject here:

https://www.selinc.com/literature/TechnicalPapers/

search for "power management systems" or "load shedding"

full disclosure: I am affiliated with the above company.

regards,
nic

 

Sangeetha... There is no way the DG-set can escape serious damage if the Main-Bus load is suddenly dumped on it... regardless of how fast the load-shedding scheme operates.

I suggest the establishment of a Critical-Load bus! It should be connected to the DG-set via the DG-breaker, and also to the Main-Load bus via a Tie-breaker.

Then upon loss of the UAT, trip the DG and Tie-breaker, and all of the Critical-Load motor breakers.

When the Critical-Load bus residual-voltage has decayed to some safe value, close the DG breaker, and then reaccelerate the Critical-Load motors in steps.

The purpose of the tie-breaker between the Critical-Load bus and the Main-Load bus is to insure that all of the main-load breakers have tripped, that is, none have failed to trip!

Regards, Phil Corso

 

Phil Corso,

I respectfully disagree. I have designed, installed and commissioned several of these load shedding systems myself for large industrial customers, and have oscillography and event reports to prove that they work. Many are installed specifically for the purpose that Sangeetha is describing... a utility connection that may trip leaving too much load for the remaining generation. They work well when designed properly.

I'd be happy to talk more about the theory of operation if you're interested.

Regards,
nic

 

nic,

I would like to know more about the theory of operation.

Presently we have used a df/dt relay "woodward", model MRG3 make to sense the uat supply. I am totally at a loss as to how to decide the setting of this relay.

Would be grateful for any feedback.

Rgds
Sangeetha

 

Nic... I too meant no disrespect of your knowledge, design experience, or commissioning skills.

My logic-strategy suggestion was only meant to illustrate that with the addition of the "Tie" breaker, risk of failure was reduced by several Orders-of-Magnitude, i.e., Qf^2 (Qf represents the Probability-of-Failure-to-Trip of one of the Main-Load Breakers!)

Wishing you and all Control.Com staff, members, and visitors, a Merry Christmas and a Happy, Prosperous, and more importantly, Healthy New Year!

Phil Corso

 

Nic... amendment to my earlier post:

Failure-to-trip probability, w/o tie-breaker, is (N)x(Qf), where N represents the number of breakers that are tripped!

Probability, with tie-breaker, is (N)x(Qf^2)!

Regards, Phil

 

Sangeetha,

How much load does the UAT normally supply? The basic theory is that once the UAT connection is severed, you need to trip approximately* the same amount of load that it was providing before it tripped.

*I say approximately because the dynamics of the system and the generator may allow you to trip a less load than what the UAT was providing becasue teh generator will be able to pick up the difference, but this gets into some more complication and require some detailed systems studies to determine (i.e. system inertias and governor response time).

Things can get a little more complicated depending on where your 2.5MW machine is presently running at. If your gen is only idling and barely producing any power, and then you expect ramp it up to full power in time to save your system, you are probably not going to make it, and your system will under frequency. If your generator is running at full power and the UAT trips and you shed an amount of load equal to what the UAT was providing, then you will probably be OK so long as you trip within approximately 100ms of the UAT trip.

Be aware that there are many factors surround the above assumptions I made, but could be considered generally plausible.

The key to being able to develop an optimal system is to be able to detect the loss of the UAT as quickly as possible. In some cases this is done using under frequency and df/dt. Using only a df/dt relay might be a little tricky and make your system a little too sensitive. It's also important that once you detect your under frequency condition you can trip the non-essential loads immediately; the contact in the trip coil string should be closed in 30-40 milliseconds max. I'm not familiar with the woodward relay you are using, so you should certainly investigate the timing and how quickly the relay logic is processed. If the woodward cannot operate that quickly, and you really need this functionality, then you may want to look at upgrading your protective relaying. Several manufacturers make equipment that can operate in the time frame you require.

Below is a link to download a paper on a project commissioned a few years ago. It gives a basic background on the concept of load shedding, and the factors involved. That would be a good place to start with to learn more of the theory and details. I'll continue to try and help with any specific questions you have, but I can only do so much. If you get really desperate, I can give you the contact information of some of our engineers in your area, and they could discuss your application in more detail. Just be aware that their help may come with a price tag.

https://www.selinc.com/WorkArea/DownloadAsset.aspx?id=5351

There is also a paper that I think resembles your situation that a colleague of mine worked on. You may this of interest as well.

https://www.selinc.com/WorkArea/DownloadAsset.aspx?id=5447

Regards,
nic

 

Phil Corso,

No disrespect taken. I agree that proper system design would make load shedding systems unnecessary, to a large extent. Fortunately for me, there plenty of plants out there that didn't put in the due diligence to either do it right the first time, or continue to do it right as the plant grew and expanded. Now we have these fun engineering challenges to try and solve because people backed themselves into a corner and are suffering with terrible electrical reliability.

Cheers,
nic