8 x 15 MW diesel plant, no grid

[ters]

An 8 x 15 MW diesel plant is proposed as the only power source for a large industrial facility (no grid). The proposed generation voltage level is 11 kV which may be used to feed some large motors directly, or will be stepped down to other medim and low voltage levels to feed other loads, as needed. The plan is to connect all units to the same 11k V bus, no sections. Diesel machines would be some medium speed units.

What would be the most optimal configuration for dealing with high short circuit levels in this case? One of the proposed options would be to use step up transformers for each unit, bring the voltage up to 33 kV and then step it down to whatever various medium and low voltages will be needed in the process throughout the plant, introducing so additional impedance in the system what obviously will be much costlier.

This would also minimize transmission losses and cable sizes, although most large loads will be close to the power plant (within 200-1000 m) but few may be as far as 5 km away.

There are other propositions such as employing IS current limiters or having generator voltage 11 kV but with each unit having a step down transformer to introduce some impendence and to bring the voltage down to say 6 kV at which level most large motors could operate.

 

[slavag]

Hi Ters.
First view:
1. Yes used bus-tie, but with IS limiter in serires.
2. Connect 4x15MW DG to each bus ( 60MW, intuitivi, , no problem with SC, possible use standard MV swg).
3. For far object use 11/33kV step-up xfr.
4. For big motors (>1-2MW) with direct connection used
same 11kV.
5. For "small" motors used MCC 3.3kV or 4.16kV with step down xfr..
Regards.
Slava

 

[davidbeach]

There are various options depending on how well you want to protect the generators.

The "easiest/cheapest" solution (you get what you pay for) would be to solidly ground the generators and connect them all to the same bus. Any fault within the generator will result in the need for a restack or a complete replacement.

The next would be to low impedance ground the generators and connect them into the common bus. Better protection for the machines, an internal fault will now require a rewind and maybe a partial restack. Ground faults on the system will be seen by all generators.

Best would be high resistance ground the generators and connect them to the system through transformers with delta on the generator side and grounded wye - solidly grounded or impedance grounded - on the system side. You could use one 18.75MVA transformer per machine, or you could get 37.5MVA Ydd (or yDD depending on the voltage relationships) three winding transformers and put each generator on its own winding with a common winding on the system side.

With one transformer per generator you wouldn't need a breaker between the generator and the transformer. With the three winding transformers you would need individual generator breakers.

 

[ters]

Thank you very much Slava and David. I just roughly estimated the SC level, and it may not be that high as previously speculated. For 12 pole machines the X'd seems to be in the range of 0.3 per unit, and if so, the total SC current would be in the range of 3-4 times FLA. Since each unit FLA is somewhere about 800 A, I would be looking at about 3 kA of SC current per unit, all together in the range of 25 kA for all units + lets say so much (probably less) of the contribution of approx 20 large synchronous motors (they represent something like 60% of the total load), which brings me to about 50 kA, seems to be low enough for a properly sized standard switchgear to withstand. Am I making some totally out of the range assumptions/conclusions here? One unit will always be spare and not running anyway, which sort of gives me some more margin.

Ignoring the SC issue for a moment, I still have to deal with the normal FLA bus current which, if I use a common bus would seem be in the range of 6000 A times X factor 1.25 comes to something like 8000 A, assuming the main bus is 11 kV. This is a bit high, but still achievable.

If I split the bus into 2 sections, obviously each section does not need to be rated at 8000 A, but by doing so it sems I would somewhat reduce the supply reliability - all loads would be fed from only half of the units.

Speaking of the bus, and ignoring what the applicable standards dictate, I'm wondering why I would need to size the bus to the combined FLA capacity if I distribute incoming breakers (from gensets) and load feeder breakers in a sort of “mix and match” configuration (generator, feeder, generator...etc), whereas no place on the bus will ever see more than some 3000 Amp?

As for the grounding, we have in mind the low impedance option, but that might change.

 

[Gianoli]

I am not sure if I am reading this post correctly.

Let’s start with “The plan is to connect all units to the same 11k V bus, no sections”.
I don’t even know were to start. How would you ever maintain this line-up of switchgear? A total plant outage every time you needed to do maintenance? Do not even consider this.

“Ignoring the SC issue for a moment, I still have to deal with the normal FLA bus current which, if I use a common bus would seem be in the range of 6000 A times X factor 1.25 comes to something like 8000 A, assuming the main bus is 11 kV. This is a bit high, but still achievable.”

Let me know when you find a switchgear manufacturer that will supply you with medium voltage switchgear with a main buss ampacity of 8kA. I think I will be waiting for a while.

I my opinion you need to revisit the entire plan with respect to this project. I really don’t have enough details to comment with any authority. However, some basic principles come to mind.

1. A dedicated 11kV generator buss line-up. Each generator has a breaker. Generator buss is split with a tie breaker. No loads feed directly off this buss. Feeder breakers off the generator buss supply separate 11kV switchgear line-ups for loads.
2. Dedicated 11kV switchgear line-up(s) for loads. The feed(s) for this switchgear would come from the generator buss feeder breakers. Probably one feed from each side of the generator buss with interlocking so that only one generator feeder buss breaker can be closed at any time.


My comments represent the bare minimum requirements in my mind as far a switchgear layout. Depending on what type of plant you are dealing with here you may need to consider more busses with automatic load transfer etc…

 

[ters]

Gianoli,

Your comments are very sound. And you are not alone, I also don’t know where to start from .

I used to work with a switchgear manufacturer and the highest current rating we were able to deal with was 6000 A. So I know that 8000 A if much higher than a standard value. However, as I said, depending how (where on the bus) are loads connected, no place of the bus will ever see that current, so the bus does not necessarily have to be rated that high, but this reasoning is likely to conflict with some applicable (IEC) standards.

The comment about maintenance is also very legitimate, but at the same time, the same problem will exist (to a lesser degree) in the configuration you are proposing.

However, this is very preliminary and nothing is stone and concrete here. Someone before me proposed that the lineup would not be sectionalized for some reason. Obviously, introducing a tie breaker is something which has minor financial implications, while, for example introducing step up transformers for each unit is much more expensive.

Still, maybe the most optimal compromise is to spend money in step up transformers, have the common bus at a higher voltage level, and by doing so we would optimize current ratings, improve grounding provisions, and reduce losses in longer feeders and simplify future expansions. The awkward part of this approach is that there would be some step down transformers literally 200 m away from step up transformers...

One of the requirements is also a reliable supply, but to improve reliability, we may need to go with the double bus system, i.e, two lineups, whereas any incomers and feeder can be connected to bus A or bus B depending which one is available. If both are available, than both incomers and feeders could be split roughly 50:50.

 

[davidbeach]

Go for a more utility design than more of an industrial design. Put in both a main and a transfer bus with a breaker between the main and transfer. That will allow transfer of any load or source onto the transfer bus via switches between the breakers and the cable lugs. Split the bus into at least two sections. On each section alternate equal amounts of generation and load. Best would be single load breakers that feed load distribution boards with about 1/8 the total load each. A better configuration would be 12 or 16 sections in a ring bus and then you wouldn't need the main and transfer busses. Design the system that you can have any one source or any one load connection out of service at a time and you can maintain the whole thing. This configuration might suggest not using the three winding transformers as you would loose two generators at a time to maintain that section of the gear (but you are at least N+2 on your sizing aren't you?) and the one transformer per generator would work better with the ring. Many, many good reasons to high impedance ground the generators and separate them from the rest of the system with transformers and only reason not to, that reason would be to build a CHEAP system (penny wise and pound foolish). Select your generators at a voltage that makes the most sense for the application and select a distribution voltage that keeps load and fault currents at manageable levels. Don't let the bean counters do the engineering.

 

[slavag]

Hi.
As usaully , we see dilemma, price- best technical solution.
Utilities apllication - industrial application.
Ters, you have now all of options from cheaper up to utilities solutions.
David closed all best solutions.

To my first post, I would like add additional options:
1. I'm assume, that you'll use metal-glad swg.
2. You can use double BB solution ( air or SF6 primary substation).
3. As I remember, you have option for 12kV, BB up to 4000A
but SC wistand up to 30-40kA.
4. For my pinion: good solution is
Doble BB -Double station, half generation and half load on one station with bus-tie and station coupler to second station (section).
5. I think, you need check with mnf. , satndard SC level up to 31.5kA for 3sec., that mean add IS limiter in series with station coupler.
6. In this solution you split generation/load on 4 BB's
each-bus for example possibe order 1600/2500A.

David, I would like send your expressions to my bosses and customers, I think from this need start each tender/offer/technical evaluation. and it's answers for maqny Q's in the forum.
"You get what you pay for"
"Penny wise and pound foolish"
and very importans for my bosses
"Don't let the bean counters do the engineering"
Regards.
Slava

 

[ters]

As always, this forum is the greatest place under the sky to ask technical questions. Thanks you very much all of you.

The configuration, as proposed, may be N+1 or N+2. The pick load is about 90 MW, so depending on the unit size there could be anywhere from 6 to 12 units. Due to delivery time, the size of units is also somewhat flexible for now, introducing so one more variable to this...

If we use only 6 units, then it would have to be N+1, which is not good. For now, I’m assuming that 8x15 MW and N+2 would be the most optimal solution. The argument against more smaller units is again cost, complexity, lower efficiency, etc.

The argument against step up transformers is not only the cost but also load proximity. This is an ore processing facility in the middle of nothing. I cannot perceive reasons why the power plant should not be inside the same fence with the processing plant, whereas units will be outside while the generator switchgear will be practically under the processing plant roof.

So, if I step the voltage up from 11 to 33 kV, then I need one or two 33 kV switchgear lineups, and then, immediately after that I need to start stepping down for most of the load. Excluding a couple of remote loads (expected to be within 5- 10 km, which may be in the range of 5-10% of the total load, and which then can be accommodated using a 33 kV OH line.

Low voltage (0.4 V) is going to be in the range of 10-20% of the total load, which can be solved via say 3 -4 step down transformers and low voltage switchboards. So remaining approx 70% of load are mostly medium voltage motors, synchronous or not, ranging from 300 kW to even 15 MW, for which I need to introduce X number of larger step down transformers feeding also X (or different) number of MV voltage switchgear lineups, whereas MV can be 11 kV, 6.6 kV or 3.3 kV, and probably at least two of them will exist.

I perceive the processing plant area to be approx 600 x 1000 m. The argument some people have is that it looks rather strange that on a small area like 600 x 1000, one first steps up 8 times, and then within a few hundred meters steps down something like 8 – 10 times depending what No X is going to be.

Too many transformers they say, larger losss, more things to maintain... But it sounded from your and some internal comments that there are still plenty advantages of having this step up - step down configuration and the transformer yard full of transfomers.

 

[slavag]

Hi Ters. For industrial plant are practical use same voltage for generation and load w/o step-up/dowm xfr. But usaully we used or double BB or V connection ( that mean two CB for generator infeed)
Please take in account, motor 11kV at size up to 1MW it's special order, best solution for this motor step-down xfr. 11/3.3Kv and you will use MCC with contactors and fuses
(tripple level).

Several examples, that I remember:
Oil factory, double bus with direct connection of two 20MW
generators and two utilities 70MW xfr. 6.6kV
IPP ( Wartsila), 4DG ( 8MW each one )connection to same common bus (6.9kV)
Airport: 6DG on same bus ( Hunday project).
All application with low-impedance grounding.

Of course, better step-up xfr. for each generator, but..

About protection, in case of Wartsila project, for ground(earth) fault protection used 67N protection
Regards.
Slava