capacitor assist 30,000 HP motor starting

[rconnett]

I'm doing some preliminary motor starting studies (fatal flaw analysis) for a centrifugal gas compressor addition. I'm using SKM for the modelling. We're proposing using a low inrush 12.5 kV synchronous motor. The methodology proposed (by my client) to start the motor and control utility flicker is similar to that which we've used previously on smaller machines at the site - using a capacitor to partially supply the VARS during starting.

The problem is, the options are to use a 25 or 30,000 HP motor. These require a 55 MVAR and 75MVAR capacitor bank to control flicker, switched on the 12 kV bus with the motor. (The cap bank is switched out when the motor comes up to speed). The ratio of the cap bank size is 2-3X the motor size here, which is a lot bigger than I've previously seen. The client is very insistent that this method be used. (I'm proposing a "pony motor" to spin it up to speed).

1) Are any of you aware of any objective standards or recommended practice guidelines which would give any guidance?

2) Other than doing an EMTP study (which we might do anyway) are there any other recommended practices to look at for capacitor switching?

I'm doing this for flicker control on the utility bus (115 kV). I'v checked other issues such as loading, added bus fault contribution, etc & all looks OK at this time.

Thanks

 

[jraef]

thread237-77854
has some references to IEEE papers that may prove usefull.

Have you considered using s solid state starter?

"Venditori de oleum-vipera non vigere excordis populi"

 

[Skogsgurra]

I would recommend a pony motor. Especially if you can spin the motor up to speed without having any load connected and can allow a few minutes to get up to speed. Such an arrangement will be a lot cheaper and will influence the grid very little. You will have a small lock transient when you connect to the grid (after having brought the exitation up to nominal), but a lot less than direct start plus capacitors would produce. You can add a phasing device to avoid the transient altogether.

Done that on a 12 MW chipper for pulp production in a mill with a weak grid. No problems at all. Compared with the problems we had with a similar size heat pump motor with capacitors on a much better grid, the problems were non-existent.

 

[Modula2]

Is the flicker assessed by the peak or rms? Would the cap come on first, then the motor? Has a VFD been examined considering energy costs?

 

[HCBFlash]

Is this a "what-if", or for real? On the tech side it'd be interesting to see your calc's and measurements etc., but it sounds like your client is pressing for their own choices. The "pony" is probably the best answer, but your job selling it may be doomed from the start.

There are lots of threads under "electric motors" that go into many facets of many cases of starting large motors. I've seen *skogs* around plenty of them, he just might know a thing or 2....



remember: An opinion is only as good as the one who gives it!

 

[rconnett]

No - this is for real.

My client has a pretty good mechanical / rotating equipment guy (They're a major int'l oil co) who is pushing for the cap assist. He's done it elsewhere - but I believe the utility flicker reqmt's are not as stringent.

To make matters more interesting, the fault contribution from the motor exceeds the duty rating of the swork - (750MVA).

Thanks for te input-

 

[controlnovice]

Can you try a synchronous machine? Much lower starting current! I've done a 34,000hp machine in Canada across the line without any problems. Incoming was 345kV.

I'd be very hesitant about the capacitors....

 

[controlnovice]

Oh wait. I just read that it IS synchronous. My bad.....

 

[rconnett]

Just an update:

1) My calcs show that the capacitor assist will work, (long story here - they're already doing it with a 15,000 hp motor), but only if they have their two substation transformers in service (in parallel for redundancy). If they lose one, starting is out of the question. Hence, this affects unit reliability & may be a bad idea.

2) The main rotating equip[ment engineer for the client nixed the pony motor. He is very concerned about the added complexity/difficulty of balancing the units. (can't say as I blame him thinking back to my steam turbine /generator days).

3) Active debate going on now about whether a VFD would pay for itself for varying the process loads. Other options are available for turning down the process load (a natural gas compressor).

4) If a full size VFD is not required for varying process loads, then for starting only, we may go witha VFD sized at 25-40% of the motor size, since the unit will start unloaded & requires a much smaller amount of HP to come up to speed. Once at full speed, the motor will be synchronized to the line & loaded up.

Thanks for all your suggestions - thought I'd give you all an update as I have found it interesting.

 

[jraef]

rconnett,
You should still consider RVSS starting. There are several manufacturers who have been making 15kV class RVSS starters for years now, and they will usually cost 1/10th the price of an equivalent VFD if speed control is not necessary. I know for a fact that at least one of them has done several gas compressors with success beyond the original expectations.

"Venditori de oleum-vipera non vigere excordis populi"

 

[rconnett]

jraef:

I understand & appreciate your input. I might point out the following:

1) RVSS are very poor performeds when it comes to starting torque. SS starters are no better than series reactors when it comes to reduced voltage starting, and when you are trying to maintain minimum starting torque, SS are the worst type to use - remember that torque goes as the square of the applied voltage.

2) We are working closely with the motor vendor on balancing inrush, locked rotor torque, minimum terminal voltage, and load torque requirements. Synchronous motors have the advantage of being able to treat starting torque and full load running torque as independent variables.

Consequently, doing reduced voltage starting - solid state or otherwise - is inappropriate for synchronous machines where we do have control. It is appropriate for induction machines driven from stiff systems.

 

[Modula2]

What hits the line first, cap or motor?

Will your utility accept an approx flicker calc by simple steady state phasor calculation?

What measurements will be used for cap switch off?