DECS 200 Over Excitation Test 2

[FrankTWarrens]

I am running 8 engines in parallel and want to test my protection scheme against over-excitation of the DVR. I was told by Basler to lower the Exciter Field Overcurrent setpoint down to test the function, but I want to test the effects against the other engines and if my coordination study is true to form. I was told that there is a possibility of damaging equipment if I test at the current Exciter Field Overcurrent setpoint, but I believe there is a time delay before this becomes damaging. If I want to initiate this test by increasing the increment AVR V setpoint at a X% increase and my Exciter Field Overcurrent setpoint is Y amps, is there a way to calculate how much time I will have before I cause damage to equipment?

 

[waross]

There are several ways to cause damage by over voltage.
One way is semi conductor damage. This may be over voltage destruction or the over current associated with the over voltage.
If you push the generator into instability it is hard to anticipate the time or the possible damages.
The over excitation will cause reactive current to flow. This may be greater than anticipated by a simple application of Ohm's law.
Some components may experience the onset of magnetic saturation and a disproportionate rise in current.
How long to cause damage? It depends. Anywhere from instantaneously to hours.
Again; It depends.
If you choose to ignore the advice of the experts in the AVR field (Basler) and disrespect their years of experience with AVRs then you may consider this:
Calculate a resistance value that will be required to draw the additional current needed to trip your Exciter Field Overcurrent protection.
Connect the resistor across the AVR output in parallel with the field and with a switch in series with the resistor.
Start the set. Now you may either close the switch for an instantaneous trip.
or
Drop the excitation slightly, then close the switch and activate the trip by raising the excitation back to the normal level.
Tip: reduce the throttle setting to reduce the load on the set under test. This will allow more reactive current to flow without overheating the generator during the test.

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

 

[FrankTWarrens]

Thanks Bill,
I appreciate the input for the testing scheme and will discuss with my team. I do not want to ignore the advice of the experts, but regulatory requirements demand that I test the system to demonstrate protection of the power system as this is on a marine vessel and the power plant is running the propulsion systems. There are 4 generators connected to each of two switchboards for a total of 8 generators. The two switchboards are connected through a bus-tie circuit breaker. After (or at the instant) the over-excitation fault, the bus-tie breaker must open to isolate the healthy switchboard. This must happen before the healthy generators trip on reverse power. This usually requires a form of 'Advanced Generator Protection,' something like what AKA offers, but many other manufacturers offer such protection now. This is an older vessel however and without damaging equipment, I need to prove to the people that sign the checks whether or not we need advanced power plant protection.

 

[electricpete]

That sounds like a good idea that Bill suggested. I'm not sure of the configuration, but I'd think about keeping the the temporary resistor on the voltage regulator side of the field breaker to avoid scenario where rotor field current might be switched into the temporary resistor circuit (which may be susceptible to open circuit from overheating or mechanical factors).

=====================================
(2B)+(2B)' ?

 

[Parchie]

Maybe it's just me but I have this question ringing bells in my mind!
Why do you have to "actually" do that type of test on old machines? Does the test results give you discounts from your insurer/s? Just seeing it from my own point of view though, we know it's your equipment.

 

[catserveng]

I've had to do a few of these type tests. To do this right and not damage anything you would need to provide a lot more details, like size of the units, what other protections are installed and/or enabled, if all the units are the same size (rating)and how your overall system determines how a bus is "healthy" or not and how it takes action.

I have gone with Basler's advice and demonstrated the action of the OEL by using a lower settings and showing that the OEL response followed the published curve in the manual. The published test for the OEL is in section 6 of the manual (page 19 in my copy).

I have also "bench tested" a single DECS200 into a dummy load after developing a test plant that was approved by the customer and inspector. Setting up a bench test for these regulators can be a bit of a pain, but can be done. One thing you must be careful of, DO NOT OPEN THE FIELD UNDER LOAD! It is the single biggest cause of failures for these type regulators, they don't tolerate it well at all and in most cases it kills the field output driver. 90% of the DECS family DVR's I have replaced are due to failed field outputs, always associated with a problem (sometimes intentional, like installing a field relay or breaker)in the field circuit.

Driving a generator on an island system, like your vessel's electrical system, into an actual over-excitation event without something to consume the load will cause the voltage to rise up, can be pretty high depending on where your setpoints are. Does your system have over-voltage protections installed? Will you be able to disable them for the over-excitation testing? What are the over-voltage limits of the equipment on your bus?

I tried doing a test per Bill's advice above, but with an older series of the DECS regulator, it consumed one of the regulators, and I'm not sure of the exact reason, and in the end it was unsuccessful because we couldn't hit the setpoint target to the inspector's satisfaction. If you use the suggested procedure and heed Bill's advice about reducing speed, make sure the V/Hz feature(UFRO)is disabled before you test.

A final comment, based on my experience, mostly with engine driven units between 500 and 5000kW in most all kinds of applications, over-excitation events about half the time are due to a regulator problem, and while the newer digital AVR's are pretty nice and have a lot fewer problems, if they do fail you likely also lose your protection. If you truly have a critical system then I would expect that "Advanced Generator Protection" would be desired, like a good quality multifunction relay properly set and tested.

I am a pretty big fan of the DECS family of regulators, but that said in most cases we use little or no protection functions in them, except sometimes in a "backup" type of role.

Hope that helps, MikeL.

 

[waross]

Thanks for the tips Mike.
I did not mean to reduce the speed. I suggested reducing the throttle to reduce the load in anticipation that there will be a lot of reactive current flowing during over-excitation. My idea was to avoid an over-current trip.

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

 

[catserveng]

Bill,

Sorry, I was thinking testing single unit at that point. Probably my own best experience with an actual over-excitation test was done in a test cell on a single unit with a very large resistive/reactive load bank, we found a good combination of a slightly lower setting, 100% real power load on the unit and a very low lagging power factor load. Of course we had a ton of resources, and a fair amount of time and help as we were type testing a package for the government. The DECS200 was new then and we also had a Basler engineer on hand for a couple of days. The end result was we found the regulator to be VERY good at a lot of things, but we also found that a properly setup protection relay (we used SEL 300G's)was essential as well if you really wanted the unit and system protected properly.

You brought up overcurrent trip, which is a good point, depending on how you test, if you really try to over-excite a "healthy" generator, you will generate high currents. Again knowing how the installed unit is protected is needed to develop a good test plan.

Parchie's question is a good one. After years of actually driving an engine into a real overspeed to demonstrate the protection (and sometimes causing a failure in the process), the introduction and acceptance of electronic speed switches allowed for the use of calibrators and the additions of field functional test options like the "75% verify" on many engine speed switches. Same argument for electrical protective relays, we can use a qualified test procedure and proper test equipment to validate electrical protections. Wouldn't it be way better for the installed equipment, both for commissioning and future maintenance testing to develop a testing procedure that doesn't actually involve cramming a bunch of current out of the AVR and into the generator?

For the OP, if you are using the DECS to tell the rest of your system a unit is "unhealthy", I would assume you are using some combination of the outputs from the DECS to signal the system the AVR is faulted, correct? Why are you separating a whole bus section for one generator faulting? Wouldn't it be better to just drop the unit's circuit breaker and lose 1/8 your capacity instead of a single AVR fault dropping 1/2 your capacity?

Since there a lot on unknowns about your system, I'd say in general based on what info you have provided that trying to do a test at the actual OEL setpoint on a unit installed in your system is likely a BAD idea.

My two cents worth, Mike L

 

[Drivesrock]

Hi Frank.
When you mentioned marine vessel and propulsion systems that really spiked my interest as I have just finished a study for OEL and bus-tie opening for a drillship after a recent OEL event. I wonder where in the world you are and who the regulatory requirements are from?

I can appreciate all of the areas where you may want to be checking the coordination. You may have a DP system with power chop (although they don't necessarily use kVArs to initiate this function - normally just low kW spinning reserve), phase-back, PMS and other power system monitoring systems?
Not easy to do on a working system and perhaps on location?
I think you'll have to do the OEL test in two parts - a low current setting in each AVR to first prove all the tripping circuits are working. Perhaps the safest coordination test on your old machines would be a 'type test' of an AVR similar to what has already been mentioned, to check the reaction time and then either plot all of the results or look at a way of simulating inputs to your existing systems - just contacts to initiate reactions or better still, a secondary injection test set connected to the protection relays to produce 'real' values for the upstream systems to react to.

If you are down South America way, perhaps we should meet up?