VAR Control of none identical generators

[Barry1952]

Is there a problem with obtaining a balanced excitation for equal VARS between generators of different kW ratings? I beleive this is done on a regualar bases but wish to know if there are any "special" precautions to take into account with the excitation system. This especially with old CVG units that act with little dealy and the new digital units. Also, does the pitch of the generator winding have an effect on the PF balancing if they are not the same?

 

[rbulsara]

Devil is in the details, but best is to wait for Bill (waross) and catserveng.

All I would say is with non-identical generators, there will always be some circulating current, key is to keep them under control i.e. within acceptable limits. The winding pitch may have indirect affect on the units power factor. Different pitch results is more likely to have different internal impedance that could cause circulating currents.

Eventually you would need a good engineer and even more importantly, a good generator field service technician/engineer. Look for companies that offer generator paralleling switchgear (not necessarily generator vendors, but they would be part of the team).

Rafiq Bulsara

http://www.srengineersct.com

 

[catserveng]

Can be done well with an active controller like a Woodward DSLC or a similar type system. Have done many systems using cross current compensation, and while during commissioning was able to achieve acceptable results, find that over time the different units drift apart. Simple things like an operator making a minor voltage level adjustment, an adjustment pot wearing, changes in generator required excitation levels due to aging, all will affect the ability to evenly share. You target VAR share percentage and the amount of voltage variation you can tolerate also need to be considered.

As Rafiq noted above, winding pitch can have an effect, but really not for the true VAR share, although large circulating currents can make it appear there is a VAR share problem. If different pitch machines are being used, you'll need to evaluate if the neutrals can be left floating or if you need to install neutral reactors, or some other mitigation technique.

Lots of details! And the commissioning process can be extensive depending on the machine differences, you may find that compromise is required depending on the dynamic response rates of the connected machines. Also need to see if transient response is required, and what allowable deviations you can have and what your time to return to a stable and balanced system is.

Some switchgear companies, like CAT/ISO, Enercon, and Point 8 have a PLC based system that if properly setup can work well, but the problems I've seen is they are targeted to standard configurations and the design engineers I have worked with are reluctant to make changes for one off systems.

I just finished a small combined system two old CAT gas engines, a small hydro and a small gas turbine using Woodward EasyGen controllers, took a while to get them to all play nice, but the overall system operation is pretty good compared to old controls, and it comes in and out of island mode pretty succefully most of the time, sometimes no matter what you do differing machines just won't play nice together.

I second Rafiq's recommendation, an experienced engineer with DG experience in the size units and application you're dealing with, and a good field guy to get all adjusted, tuned and set. I won't usually ask a switchgear company what to do anymore, since they'll likely try to sell a solution that is more their "standard" and not adaptable like the Woodward type controls.

My two cents worth.

Mike L.

 

[Barry1952]

THanks to both of you for your responses.
I realize that it is a tough issue and many factors can quickly work against you.
Not sure what you mean by "Devil in details" and "Lots of Details". Do you need more details to offer more accessment?
I'm concerned with the old Siemens CVG units that are factory tuned and have not pots to adjust just taps. It has a specific range it is set to work in +/- 2.5% of set voltage between NL and FL and the fields are connected in parallel. The new unit is a CDVR by Cat that is their latest for a digital VR unit. Concerned as to its ability to be tuned to match the response of the existing 2 units. Or even if it can be!
Will no more later next week when I get to see the system.

 

[rbulsara]

Details means that there are so many variables that it cannot be addressed remotely and without spending time and effort. A good technician /engineer can make even not so user-friendly system work and other could mess up most advanced ones.

Having said that upgrading older controls to newer ones, possibly of one manufacturer, may be a wise move just from flexibility and maintainability points of view.

I have stories to tell regarding such issues.

Rafiq Bulsara

http://www.srengineersct.com

 

[waross]

DEVIL IN THE DETAILS;
For two or more sets in parallel, the KW production depends on the mechanical input power, often stated as the throttle position.
The VAR production depends on the field strength.
Throttle for KW and field strength for VARs.
Simple as it can be. But, controlling the throttle and the field strength to achieve the results you want automatically over the operating range- Therein lie the devil and the details.
Thank you for the recommendation, Rafiq but I will defer to catserveng on this. He has more experience with load control panels. I can handle the old systems but I am not up to speed on the load control panels that catserveng works on.
I wonder when you say that the fields are in parallel on the existing machines. Interesting.
Is this a grid connection or an islanded plant? You may have to install new AVRs on the old sets to get acceptable performance.

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

 

[Barry1952]

Thanks Bill,
Yes it is an island plant (ship actually) and yes the fields get connected in parallel for the existing 2 machines but not for the new replacement machine.
We need to see if the new AVR is set to respond appropriately with the old. The old have little to no adjustments but the new may not be set to work "within" the boundaries of the existing.
In the end, perhaps the best solution is going to be to repalce the Siemens CVG units and put the same AVR's on them to enable better matching of response (i.e. gains, ranges, etc.).

 

[catserveng]

If you're putting in CAT genenrator with CDVR's, one of the first things you'll need to do is reduce the loop gain in the AVR dynamics from the default of 50 down to 10 or maybe even 5. The other settings will likely be ok for a starting point.

If your existing voltage regulators have a volts/Hz function (of UFRO as Bill and others call it) make sure the new AVR's have the same slope and knee point.

Any form of cross current compensation you try will like not be satisfactory, no matter what anyone says, you might get it to work for a while but it will always need attention to keep the VAR share proper, so voltage droop or active VAR control are likely your only reliable options. The CDVR has much the same control logic and abilites of the DECS200, and will track very well (if properly adjusted) with units of similar dynamic response.

I have no experience with your existing AVR's, so my approach would be this, get a resisitve/reactive load bank sized to 100% kW and kVA load as one of your existing units. Perform step loading and document your transient response. I would do 10 and 25% load steps, try to determine your average response characteristic. The do the same with the new gens and try to match them as closely as you can. You'll likely not get them very close, so there will be a compromise. But doing this will give you your best chance, in my opinion, of getting the system as good as possible. I've worked on lot's of shipboard systems, both house power and propulsion, and likely the desire to use a load bank will be expensive, and many try to use house loads to get the systems tested and commissioned, long term the end results are usually not what the end user will want, at least in my experience.

Consider upgrading the existing gens to a Basler DECS 200 if your tail ends will play nice with them, my experience with CAT CDVR's and DECS200 mixed plants has been pretty good as long as 5 amp secondary CT's properly sized are used on the CAT units.

MIke L

 

[Barry1952]

Thanks Mike,
I thought the CDVR seamed similar to the DECS200. Especially with the BESCOM software when I looked at the info in the CDVR manual sent to me.
I don't have the settings yet for the CDVR but am interested to see what has been set (especially the loop gain as you mention).
I've heard of v/hz but not UFRO..that's a new one to me.
Attached is the Siemens CVG unit in case you are interested.
As noted, it has minimal adjustments other than taps. But it is tuned now and has worked well todate and still does with the other Siemens unit (I'm told).
Replacing them to the DECS200 might be the end recommendation. Familiar with them.
Getting a ship to setup for controlled load bank trials is near impossible unless it is tied to a refit or situation where they will take her out of service. Ship load will be the only option I'm afraid but if there is a heavy load (large motor) than can be started and used as the common step change, that will have to do.

Thanks to all for your input