LV Genertator feeding Transformer

[sibeen]

I have a site which is currently fed via a centre tapped 480 volts giving two 240 volt supplies (j240 and -j240). All the loads on site are single phase 240 volts. The site requires a new generator(s) system to provide back up power in the event of a main fail. The site load is approximately 40 kVA. The client is suggesting using two 20 kVA single phase generators to provide the backup power required.

What I'm thinking about is using a 50 kVA three phase generator and feeding a Scott connected transformer to provide 2 x 240 volt systems. Due to service agreements the client has in place I will be limited to using Cummins generators. So I'm thinking of using a 50 kVA which uses a PowerCommand 1.1 control system.

http://www.cumminspower.com/

http://www/Commercial/Controlsnetworks/s-1566.pdf

The transformer will be directly connected to the output of the generator. My worry is that the generator will fail to start as the inrush to the transformer will cause it to stall.

I've contacted Cummins and whilst they have come back to me it hasn't been all that reassuring. Has anyone had any experience with this style of setup and can suggest whether it will work as I propose?

 

[waross]

Forget the Scott transformers or any transformers for that matter.
You will need a 60 KVA gen-set at 120/208 volts.
Be sure that the set KVA is rated at 208 volts and not 240 Volts or the set will be undersized. (Been there and got the "T" shirt.)
The set may be reconnected for single phase.
Use the Zig-zag connection to develop 240/480 Volts.
You may have to do this yourself. Cummins' standard three phase to single phase re-connection is the double delta.
The double delta is suitable only for 120/240 Volts.
If you want to forgo the 480 Volts and go for a single 240 supply Cummins will supply a 40 KVA, 120/240 Volt set.
It will be a 60 KVA, 120/208 Volt three phase set re-connected in double delta.
If the contact person at Cummins can't fill the order, go higher up the food chain. (Been there also with another brand. Got that "T" shirt as well.)
Up to 20 KVA, and 3600 RPM (15 KVA at 1800 RPM) some sets are wound single phase. At larger sizes it is cheaper to reconnect a three phase set.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sibeen]

waross, thanks for that, but I feel I may have been a tad succinct in my opening post and did not provide enough information.

I'm in Australia where the common LV voltage is 240/415, which is now specified as 230/400 in line with European standards, but suffice to say that at this site the line to neutral voltage is 240 volts.

The client is a large company, at least by Australian standards, and has thousands of sites spread across the continent. Due to this they strive to use standard packages of equipment. Any generator installed, especially at the smaller sizes, will be standardised as far as possible to ensure ease of service or replacement under extreme conditions. In this way I'm going to be restrained to be using a standard three phase 240/415 volt generator.

Which is why I came up with the Scott connected idea in the first place.

As an aside, I don't think I've seen the Scott connection mentioned on these boards for quite a few years, and lo and behold, it was mentioned twice on another thread over at the transmission & distribution thread

 

[ScottyUK]

sibeen,

The AVR will regulate whatever winding connection you adopt to produce the required voltage on the sensing terminals. A 415/240V machine will likely be connected in series star, but could likely be reconnected in parallel zig-zag: most are 12-lead types. Each winding will develop 120V with the normal phase displacement, and the zig-zag connection will give you 240V single phase. The AVR will need to be a single-phase sensing type, but these are fairly common.

 

[waross]

Thanks Scotty.
The series zig-zag will give center tapped 480 Volts as you have now. The connection change may be done on a series connected star with only a continuity tester even if all the number markings are missing from the leads.
Typically a 415/240 Volt machine will have three line or phase leads marked L1, L2, and L3, and a neutral bus. The neutral bus will have leads number 10, 11, and 12 connected to it. Remove one lead from the neutral bus. Identify the phase lead with continuity and then connect the neutral end lead to either of the other lines and insulate the splice and forget about it. You will now have two line or phase leads available and 480 Volts across them. From either line to neutral you will have 240 Volts. True single phase.
Example:
Remove lead #11 from the neutral terminal and connect it to lead #3 (Line 3). Insulate it and forget it. Now you will have 240 Volts from line 1 to neutral and 240 volts from neutral to line 2. You will have 480 Volts from line 1 to line 2. The supplied circuit breaker will still be the correct size.
Or you can buy a single phase 40 KVA, 240 Volt machine and and use only 240 Volts.
Most 240 Volt single phase machines may be reconnected to provide 240/480 Volts but the breaker should then be changed.
A single phase 40 KVA generator will be a reconnected three phase 60 KVA machine.
If there is any advantage to leaving the field equipment on the center tapped 240/480 Volt arrangement, go with the 480 Volt 3 phase machine and change one wire to convert it to single phase.
Here is a sample alternator spec. Diagrams start on page 8.
Diagrams, FF and G show standard single phase connections.
The change that I suggested for 240/480 Volt use is based on diagram D

http://www.leroy-somer.com/documentation_pdf/4803_en.pdf

Here is a single phase 240 Volt set. Cummins should be able to match this.

http://www.olympianpower.com/cda/files/4118701/7/LEHF0215-00.pdf

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[waross]

If you want to share some details of the load profiles and duty cycles, we can suggest which options may be advantageous and which may not be required.
Would you consider this a prime power or standby power application? How many hours a year?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[ScottyUK]

sibeen, Bill -

I mis-read the original post, my reference to parallel zig-zg is wrong. Should be series zig-zag as Bill describes. I think the rest is more-or-less right. Ooops.

 

[mls1]

I don't think the fundamental question of in-rush current stalling the generator if a Scott connected transformer is used. Because the voltage will build up on the generator I don't think this will be an issue. Inrush is caused by saturation from asymmetric offset of the core flux due to the difference in remnant flux from steady state waveform flux but since the transformer will experience a voltage buildup rather than a discrete change I don't think the transient will be the same. I wouldn't expect nearly the same inrush and I certainly wouldn't expect it to stall the generator.

 

[waross]

Did I mention that I have hands on experience with single phase generators in that size range? 15 KVA to 60 KVA single phase. All the sets above 20 KVA were reconnected three phase sets. This is standard industry practice.
There is no advantage to using any transformer, Scott connection or otherwise. A transformer for this installation is an added expense and added complexity. If you use a transformer it should be connected to the generator before the generator is started. Even if the transformer is switched in after the transformer is up to full output, the inrush will not be as severe as it would be going across the line. The source impedance of the transformer is much greater than the source impedance of the grid. If there is enough real power component in the inrush to stall the generator, the UfRO feature of the AVR will reduce the voltage to maintain close to the proper Volts per Hertz ratio for the reduced frequency.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sibeen]

waross, Scotty and mls1, thank you all for your responses.

Just to flesh out the story. The site in question is stuck on top of a hill about 5 hours drive from a major city. I was initially tasked with visiting the site to implement a design for a new 48 volt DC power system to cater for load growth. When I arrived at the site the power had been down for about 24 hours and the site was being held up by two trailer mounted generators that had been carted to site after the power had failed.

The site does have an existing generator, but as it is rated at 10 kVA and the site load is about 30 to 40 kVA, it was never going to cut the mustard. I was unable to take any really meaningful power readings as the trailer generators were feeding vital equipment via extension leads and such, and all non-essential equipment had been turned off. So I wandered around with my torch and my camera and collected what site information I could.

Basically the site is a basket case. As stated, the original generator is no where near up to size, the main switchboard is from about the 60s and there is no way anything else can be attached to it, the onsite documentation is completely non-existent etc. So I drove don the mountain, called the client and explained that his budget for this job just exploded. I now have to come up with some options to give them so that the site can be bought back up to spec.

Now, many years ago when I first read about the Scott connection I thought to myself "I'm going to put one of these in one day". Approximately 30 years later I'm still waiting and this may be my one and only chance. I strongly suspected that as the generator started up, with a transformer directly connected to its output, that the ramping up of the generator voltage would limit the current into the transformer, and that the generator wouldn't have a problem starting. I just wanted to check to see whether anyone had practical experience with this sort of situation.

waross, unlike you I've had no hands on experience with generators and their winding. As long as I knew the subtransient reactance etc I just treated them as a black box at the start of the line. So thank you for those links and I'll get out my pad and pen and try to understand them. I'll also be contacting the cummins engineer come Monday and see what he can do to help. My one worry about using a bespoke wiring termination on the generator is what happens in 5 or 10 years time when the generator decides to blow a poofle valve. The client is rather notorious for going to the market every few years to get a better price on their service contracts, and knowledge and documentation dribbles out of the organisation at a great rate of knots. This is the major reason that I was trying to stuff a standard three phase generator into the design. Saying that, I'll certainly be putting this forward as one of the options for the site.

But I really, really want to use a Scott connected transformer

 

[ScottyUK]

sibeen,

It is commonplace to close couple a generator and its transformer so that the generator builds up excitation with the transformer connected. This is used on sets up to several hundred MW, and there is no inrush.

Bill's suggestion uses a standard 3-phase machine but reconnected to produce single phase. The zig-zag connection is well-documented, probably better documented than the Scott-connected transformer. Any have-decent generator tech would be able to look after it. Even I could probably manage if I had the manual.

 

[waross]

Time for tough love. Your generator needs a Scott transformer almost as badly as the water cooler needs a Scott transformer. There is no advantage to two phase and some disadvantages. I understand your desire to use a unique transformer connection but this is not a defensible application. You may want to look into the Tee connection. Tee connected dry type lighting transformers were fairly common about 20 years ago, which places them several generations ahead of Scott transformer applications. These were generally about 37 KVA and fed 120/208 volt lighting panels. Do some research and see if these are still available. This would be a much more defensible application of a Scott like connection. These transformers used two cores instead of three cores and transformed three phase to three phase with a variation of the Scott connection.
The zig-zag connection is a common and time tested connection.
Any complete list of generator connections includes the zig-zag connection.
The Collins or bar-diamond connection is a rarity, but that is only needed for the single phase connection of a 10 lead generators and they went out about two generations ago. The most common source of ten lead machines was old Delcos which had served for years as Telco backups. They generally hit the surplus market as very old machines with quite few hours.
But you don't have a ten lead machine and aren't likely to see one.
Tough love #2:
Any utilization of energy at a phase angle different from the source phase angle involves an power factor equal to the cosine of the angle of displacement.
With a bar-diamond, double delta or zig-zag connection two of the three windings have a phase displacement of 60 degrees, (cos 0.5),
one leading and one lagging.

My one worry about using a bespoke wiring termination on the generator is what happens in 5 or 10 years time when the generator decides to blow a poofle valve.

This would be a valid concern if you used the little known Collins (bar-diamond) connection. Any generator service man who is not familiar with the zig-zag and double diamond connections is not quite ready for prime time. Hopefully he has a mentor nearby.
Your three main options are:
1. Reconnect a 240/415 Volt generator for 240/480 Volts single phase. This is the best option if there is a long shared neutral between the two 240 Volt loads.
2. Buy an off the shelf 240 Volt single phase generator.
3. Buy a 240/415 Volt generator and use two phases to power the two 240 Volt circuits. (This is unacceptable in a large machine due to circulating currents and over heating in the rotor. In a small machine, particularly on standby duty, rotor heating is not an issue for several reasons.



Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[waross]

Sizing:
Most gen-set manufacturers have sizing programs available for free download.
Quick and dirty sizing:
Total KVA or Amps of all loads including motors.
Plus
200% of the KVA or Amps of the largest motor or motors that may start simultaneously. (There are economical monitors available for refrigeration compressors which monitor Volts, Amps, Voltage unbalance and a few other things. They will trip on phase loss, phase reversal and excess voltage unbalance. More importantly, for gen-set applications, after a power outage they give a random time delay restart. This reduces the block loading when the gen-set starts and picks up the load. There are other solutions also.)
Size the generator for the total KVA or Amps.

For occasional non-critical standby use use a standby rated set.
For critical loads and applications which must be test run monthly, consider using a prime rated set. (Basically 10% over-sized, but may include options such as an oil cooler and/or a larger oil sump.)
The 10% over-size helps to ensure that twenty years from now when the engine is worn and feeling its age it will still be able to pull the load.
I have seen prime rated sets for continuous duty where the engine was over-sized an additional 25% to allow an aging factor.
By the way, most manufacturers are happy to custom configure gen-sets.
After you have run through an on-line or downloaded sizing program, you may want to double check against the quick sizer I provided to check for gross errors.
Anecdote: Manufacturers want to sell sets, the bigger the better. Their sizing programs are conservative. About 20 years ago I installed about 20 sets. At last look, about 6 years ago, all were still in service. All were sized with my quick sizer. A few years ago I ran the numbers through a manufacturers sizing program and all the installations failed. The big difference is in the voltage dip when a motor starts. Most customers are happy to accept a larger voltage drop than the manufacturers recommend.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sibeen]

OK, OK, OK, waross, you've ground me into submission

One thing, when I talk about the Scott connected transformer and you talk about the Tee connection, I suspect we're talking about the same thing. For me the Scott connection is a dual core, one limb centre tapped, the other either tapped at 86.7% or just wound at 86.7 of the other primary (see diagram).

To see whether I'm understanding what a series zig zag connection for generators is all about I've included a diagram. Hopefully I've put all the leads in the right spot.

Summarising:

1. Re-wired three phase generator.

a) Easy to do.
b) Less complex than alternative.
c) Less expensive.
d) Will require oversized generator.

2. Scott - T transformer option.

a) More complex.
b) Can use standard three phase connected generator.
c) Generator does not need to be oversized.
d) More expensive option.


Does that sound about right?

 

[waross]

Almost right.
Any three phase to single phase conversion will require an over-sized generator. The manufacturers long ago determined that it was cheaper to market reconnected three phase machines than to market dedicated single phase machines in larger sizes.
I am pretty sure that you will find that the Scott connection also requires an over-sized generator.
A shared neutral is not a good option with a two phase system. In the case of balanced loads, the neutral current will be 141% of the line currents.
There is another option for your application. I don't recommend it but you may find it interesting as it is a "Tee" derived configuration.
Consider the four wire delta connection. This is most often used in distribution systems with an open delta to supply small amounts of three phase energy where the majority of the load is single phase. (The closed delta secondary with a four wire wye primary is a bad configuration for distribution for several reasons that don't affect generator applications.)
Back to the generator.
Look at the high voltage delta connection. This connection will produce 240 Volt three phase, line to line.
L2-M-L3 will produce 120/120:240 Volts.
A connection from L1 to M will produce 208 Volts. A 240 Volt auto transformer with an 86.7% tap will boost this up to 240 Volts. Now you have two 240 Volt circuits displaced 90 degrees. This is the same result as a Scott connection with one small auto-transformer rather the two transformer Scott connection.
The auto-transformer size need only be (generator KVA/3 x 13.3%) or small, or for 40 KVA single phase, 60 KVA/3 x 0.133 = 2.67 KVA.
Both an auto transformer with an 86.7% tap and a conventional tap for Scott connection use will be a custom build.
Load sharing:
All of the three phase to single phase reconnection schemes shown will share the KVA load evenly across the three generator windings.
All of the reconnection schemes will result in a 50%, 25%, 25% KW sharing.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sakaran51]

Hello Everyone,
What will be the limit of a transformer capacity one can safely switch on to a generator?In many of the schemes prevailing in industries,it is not possible to start the generator with transformer circuit breaker closed like a single unit.And the transformer switching on generator is going on, day in day out, without giving any thought or concern!So what is the cut off limit?

 

[sakaran51]

In continuation of my previous post,I forgot to mention,that this scenario is especially true and prevalent in the case of HV generators,where the auxiliary transformers(11/0.433kV) are switched on to generators.Not only in industries but also during islanding operation of power plants.

 

[waross]

sakaran51: Please start another thread with your query.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[sakaran51]

Yes,Waross,I will do that.