New AVRs malfunction on VFD operation. 3

[jl1495]

REf. thread237-286989,

we had to replace the old AVRs on two generators on a marine system incl. frequency converters. The old AVRs are no longer supported, but dod operate flawlessly.
The new AVRs cannot control the voltage on converter operation.
Suppose a low pass filter on th sensing lines would do the job?

 

[ScottyUK]

What type of AVR? A three-phase averaging type is usual for a machine supplying an awkward load.

If this is a bus-fed exciter have a look to see how badly the machine terminal voltage is distorted. Some firing circuits will mis-behave if the voltage is a real mess, but usually some other load will protest first.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[jraef]

When this issue first began to show up 25+ years ago on portable gensets, mfrs like Cat had to provide retrofit filter kits for their AVRs in the field. Shortly thereafter they fixed it internally and although those retrofit kits were still available for a long time afterward, last time I asked, nobody at my local Cat dealership even knew what I was talking about. It may be too late now, it's been "fixed" for too long.

At one time I even had the circuit diagram for the retrofit kit that we reverse-engineered because the company I worked for, who sold VFDs and Soft Starters, was going to make one to sell. But all that was left behind when I quit and they never did release a product. IIRC though, it was just a simple little LC filter circuit. What I don't remember for sure were the values.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[catserveng]

What kind of AVR's did you have before? What did you replace them with? What kind of generators. Are they PM or SE? When you retrofitted did you have any problems identifying things like droop circuits, voltage adjust circuits, etc? What is the voltage doing, going up, going down, going unstable?

Somewhere at home I have a couple of simple filter circuits drawn up for older CAT voltage regulators, newer regualtors, especially the new digital regulator seem to work quite well with most loads.

Maybe some more details on your end can get you a better answer, I very rarely need to resort to adding filtering on the AVR sensing these days, however I do regularly run across retrofitted systems that used a poor regulator for the loads, or are not properly installed, they kinda work until a larger load, a load with some harmonics, or some other system upset that causes them not work as expected.

Mike L.

 

[jl1495]

The generators are LETAG 10EXR 315M types, rated at 325 KVA, pf 0.8, originally with SVE S-11P AVRs;
those being ten years old started to break down, and the type was discontinued and we had to find a replacement.
Due to space limitations the Kutai EA08A was selected as being suitable to the type of generator.
Upon installation everything was tested being OK, except for to of cargo pump frequency converters driving two motors rated at 180 kW, where upon engaging these the voltage suddenly raised from 400 VAC to 440 VAC and tripped.
Moreover, we have one additional converter of same type but 20 kW higher rating, which does not give the same fault.
All three converters are basic 6-puls front end converters.

 

[catserveng]

I looked up your new regulator, it has single phase sensing, which usually means it does not perform well with harmonics. The rise in voltage is fairly typical of an AVR seeing a notch in the sensed voltage, and tries to compensate for what it thinks is low voltage.

I'm not familiar with that type generator. Do you have the excitation specs from the old regulator? A quick internet search didn't yield much, except that it looks like the SVE S 11 is still a current product. It is also single phase sensing based on the website info, but it may have been more tolerant to voltage distortion. The current model is 63 Volts 4 amps, is that what your old regualtor was?

Mike L.

 

[waross]

Mike, would it help to take the sense voltage from three small single phase transformers connected in the same zig-zag fashion that we use to take single phase from a three phase generator help? I am hoping that there will be some averaging and that the transformers will attenuate the harmonics somewhat. Is a filter circuit just a lot cheaper?

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

 

[catserveng]

Bill,

Really depends on how bad the voltage is distorted, in some cases I've just put in a PT, either 1:1 or 2:1 and that has fixed the problem. In some cases I have sat with my Scopemeter and a resistance and capacitor decade box and played with it until I got a cleaner voltage signal to the AVR.

Frankly I used to cheat a lot, carried a Basler SSR regulator and most of the time it fixed the problem. But I mostly work on larger sets, 1 MW and up most of the time and a new AVR was not too hard of a sell.

I am running into to this a lot lately with older CAT sets with either VR3 or VR4's installed, mostly after some form of energy reduction upgrade I get a call on the first load test that the voltage is high/low/unstable. Most times a VR6 (or Basler AVC63-12) does the trick.

The first thing I do is try to take a look at the voltage at the regulator with a scope, see what the distortion looks like and make a plan from there. Honestly haven't found a one size fits all solution for this. As you pointed out, the tail end itself, the source of excitation power, the type of AVR, the prime mover and the load itself all play a part in the problem.

One thing I did note about the regulator the OP has installed, according to the website its for an AREP machine, meaning it wants a source of power NOT from the generator terminals, instead from an independant set of regulator power windings in the stator. In most cases it will probably work ok hooked up on an SE machine, but the voltage distortion could play havoc with the field output depending on how well (or not) filtered it is. It's a pretty small package, I'm guessing it doesn't have much in the way of filtering or power conditioning. Some of the Leroy Somer AVR's have similar problems.

Hope that helps, Mike L.

 

[waross]

Thanks Mike. I hope this helps the OP. Your post is the kind of first hand information that it is often impossible to get from books. Eng-Tips at its best. lps

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

 

[catserveng]

Thanks Bill, hope the OP finds it helpful.

Mike L.

 

[jl1495]

Thanks Mike and Bill, all of the above certainly helps;
however, a few uncertaintie still remain.
I just got this from our expert on the business, who is in on this line of correspondance:

Quote

The voltage sensing is correctly a 2-ph sense system – no 3-ph averaging measuring trafo is included!

The generator design is per definition … a compounding shunt machine, with two (2) separate excitation fields (a SE & CE – the latter being low resistance powered by a full 3 ph bridge rectified output from 3 off star-connected current trafos, placed in the main armature)!

Depending at the degree of compounding utilized, being either under-, medium- or over compounded, a compounding machine needs in the low load area an additional AVR, since the voltage control during uploading is handled purely and only by the design itself – no active voltage control takes place!
From the test sheet, it can be seen the voltage has a raising voltage/load characteristic!!!!, which is killing in the effort of sharing and controlling the KVAr correct and it was with the old type AVR exactly what was seen, when a generator running in parallel was taken out of service!
When load was shifted and one generator was going towards zero and the other loading up, depending on the common load level on the busbar, we observed the unloaded machine (now running under the control of the SE part and “normal” AVR) taking up a powerless current, generated by the other generator, now under control of the CE part taking gradually over! It is by nature caused by the CE system now trying to lift the voltage level!

Anyhow there is no discussion of the nature of the problem, -it relates - as you correctly says – to distorted waveforms, however the question remains, why one out of three VSD runs okay, and two not!!??

However, we plan to place a small step down trafo on the sense leads (400/230) and place a good quality EMI filter on the secondary just before the sense inputs (now using the 230 VAC inputs). Our concern here is whether we due to the phase shift in the EMI filter, will cause our reactive loadsharing part to malfunction - we have to see!
Perhaps only the trafo and a few capacitors of say 0,1 micro (instead of the EMI filter), will do the job!

Un-quote

As you can see we are trying to get the correct type/size of filtering in on the sense-lines.
Naturally, a scope on the sense lines would tell much more of what causes the problem seen; might even tell why only two out of the three causes the problem.
It is however, not an easy task these daus as the vessel in question in inderway to Africa and virtually out of reach for the next month.

John L.

 

[catserveng]

It may be possible rather than a harmonics problem you have some other issue. Do the voltage regulators have droop CT's installed? How is VAR sharing sone between the paralleled gensets, by droop or cross current? Is it possible, if droop CT's are installed the polarity is backwards?

I ask because normally with harmonics induced problems we tend to see steep changes in generator voltage, you state that it appears the voltage increase is in step with load, but only with the two units retrofitted with new AVR's, correct?

If you are in droop, what percentage droop are you using, after the retrofit was each units droop characteristic checked?

Mike L.

 

[jl1495]

Just received from our expert:

Quote

Yes, - the AVR’s have droop trafo’s installed and they are phased correct (otherwise the reactive loadsharing from remaining part of vessel load would be troublesome too – hope you agree)!!
The connection is made as standard “reactive droop compensation”, no cross current compensation!

If one (or both) droop kits (current trafo’s or connections to the AVRs) were installed incorrect, you would experience an increasing current on the generator in question – hope you agree here!
It would so to speak try “lifting the voltage”, causing a blind current to be generated, which would be absorbed by the other genset and partially by the vessel load! This would have a continuous tendency, which eventually will cause the Imax protection in either breaker or monitoring circuit to trip!

Also – with an inverse connected droop trafo - we see the current raise without the same degree in voltage raise as we see here! Here we actually has a Umax (voltage max) tripping condition before the Imax tripping, probably because the Imax is delayed and Umax not!!

The voltage do have a steep change and it happens irrespective of whether a generator is running alone or in parallel with the other! Do recall we experience only problems with two out of three VSDs!!
So I agree in the harmonics conclusion!

Un-quote

I apologise for not being clear in my description of the problems seen.

John L

 

[catserveng]

Thanks for the further explanations, and it seems you have done a good job investigating, sorry to ask but 90% of the time after a retrofit I find it is usually due to a problem in the retrofit install.

Ideally taking a look at the voltage waveform would be your best next step, it likely has a pretty good notch in it.

Was the AVR wiring configuration very different? You could swap your working "old" AVR with one of the new ones and see if the problem follows the AVR, then for sure you know it is a component problem and need to find a better solution, like filtering or another regulator.

Mike L.