ASD on a generator 3

[hydrae]

Hi, I do not normally visit here, seeking your help

My client has a well pump that has been operating fine for the past year. 150 hp submersible, powered via an pulse width modulated ASD, includes load reactors due to the distance between the drive and the motor, We added a 230 kW minimum generator to the site, with specifications that the primary load would be the pump via the drive. Just did the startup test of the generator, when we added load from the drive, the flouresecnt lights started to flicker, and the piping started to vibrate, the pressure also started vibrate, with reluctance we continued to increase power to the normal operating range, where the vibrations lessened, we proceeded with the generator proof test and at the end of the hour, we allowed an automatic transfer to normal power, the transfer switch has a phase monitor to allow transfer under load. At transfer there was a slight bump and all vibrations went back to normal on normal power. I have asked the generator manufactorer, whats up?

I have put generators on drives before without issue.

Any comments?

Thanks, Hydrae

 

[jbartos]

Suggestion: The generator winding connection and system grounding method have not been mentioned yet.
The ungrounded generator might cause overvoltages. A generator delta winding would call for a system grounding transformer or corner delta system grounding.

 

[jOmega]

The answer to your dilemma is simple: HARMONICS, as Busbar said.

The impedance of the generator is much higher than that of the ASD and the interconnecting distribution system.

Consider the generator and a source impedance resistor and the ASD's and wiring thereto as another impedance resistor, all in a series circuit. Make the generator resistor several orders of magnitude greater than the ASD's & wiring impedance. Then, introduce significant current harmonics and you can see that the significant voltage drop across the generator impedance results. This equivocates to the voltage distortaion present at the output of the generator.

Now, while the current distortion caused by the non-linear load characteristics of the ASD's has a direct relationship ONLY with the generator impedance; the consequent voltage distortion is present at every device connected to the generator output bus.

I see this weekly as hospitals and water treatment facilites and such add ASD's to replace older constant speed driven Air Handlers & pumps.

Virtually all hospitals have emergency generators that are tested on a weekly basis.

While the ASD's may operate without any noticable ill effects on transformer distribution power, surprise-surprise when they switched over to em.gen. power.

A rule of thumb is that the non-linear load on a generator should not exceed 15% of the generator capacity. This doesn't mean that the percent non-linear loading can't be greater... but if it is ... a Harmonic analyis of the system will show that you will need to had some harmonic mitigation equipment to prevent the THD levels from causing problems such as you have experienced and described.

Some months back, a Hospital in Northern California was experiencing short life on the electronic coil cirucits of their motor starters when the em.gen. power was tested (weekly schedule). Also had a lot of noise on the bus bars and vibration.

The resulting voltage harmonics consequent to the current harmonic interacting with the impedance of the generator, was the culprit.

The solution, after a harmonic analyis was performed, was to add a trap filter at the input bus to the ASD's. Some such filters are designed to attenuate the 5th harmonic; other may also attenuate the 7th harmonic as well. MTE Corp.'s Matrix Filter guarantees either a maximum 12% THiD (or 8%....depending upon which filter is selected).
This will surely solve the problem... but... an harmonic analysis of the system should be done.

Recently, some software for performing an harmonic analyis became available from a company called HarmonicPro. The software is called HA Lite. A 30 demo is available for free download. After that, a license needs to be purchased. Very inexpensive for the results that it serves up. The software allows you to analyze both transformer sourced distribution to the ASD's as well as Induction Generator sourced power to the ASD's .

If interested, see

http://www.harmonicpro.com

and if you do pursue the software, be sure to read all of the HELP files.

HTH - jOmega

 

[jraef]

I too run into this all the time, and like the others above, am convinced that it is the AVR giving trouble, caused by harmonics, switching noise (in the case of SCR based systems) or both. Also, EVERY genset mfgr representative denies, denies, denies that they have ever heard of this problem. Then they put in a different AVR and the problem goes away. Several of them have slipped up and admitted to me that they see it frequently. Cat has been the worst up until recently when they started filtering the inputs to the AVR sensing circuit as a standard practice.

Quando Omni Flunkus Moritati

 

[ScottyUK]

Hydrae,

I agree with most of the opinions here that harmonics are the problem.

The generator should be rated for rectifier duty. This should not be an unusual request for the manufacturer, even though it might be for their agent.

The generator sounds too small, given the size of VSD and the lack of much rotating plant to calm things down. Adding a fairly large (unloaded) delta-connected motor would make life easier for the generator as it provides an energy store on the system and a circulation path for the triplen harmonics.

The transformer will help IF it introduces a delta winding onto the system which will allow the triplen harmonic currents to circulate, thus keeping them off the generator. The VSD will almost certainly have a large third harmonic content which at 150Hz (180Hz?) is a low enough frequency for a fast electronic AVR to respond to.

if you are stuck with the generator, can you change the drive to one with an active power-factor-controlled front end? Or consider adding an active harmonic filter in parallel with the dirty load, which will suppress everything except the fundamental current.

 

[hydrae]

Thanks for all your inputs

We did a second load test today, swapped out the 1% AVR with a 2%, it helped but not enough, The isolation transformer to the AVR helped only marginally. The Generator tech had a harmonic evaulater, the fifth harmonic exceeds the ieee spec as predicted by cb2 (thanks) and the current and voltage plots look simular to a plot by the program jOmega referenced. The drive manufacturer is recomending a 5% filter on the line side of the drive and to see if it will work, and if not then add a harmonic trap on the fifth since that device is expensive.

Hydrae

 

[jbartos]

Suggestion: If you do not treat the generator harmonics, as recommended by the generator tech support, then there will probably be much bigger expense than the harmonic filter. How much would the passive harmonic filter cost?

 

[jOmega]

Hydrae,

Thank you for posting the latest developments and tests results.

What is the magnitude of the 5th and 7th
current harmonic and voltage harmonic ?


From experience, I think you'll save time and money going with either a 5th harmonic trap filter from someone like TCI (

http://www.transcoil.com)

which includes a 5% line inductor.

Or ... the Matrix Filter by MTE will do you well.

I do not think that the 5% line inductor alone will do you much good. HA Lite allows you to try the 5% line inductor for harmonic mitigation in the simulation... you'll be able to assess the results before committing any $$$ to it. HA Lite will also allow you to try a 5th harmonic trap filter ... and again, you can assess the benefit therefrom.

jOmega

 

[jbartos]

Suggestion: Active harmonic filters or hybrid harmonic filters might also be considered for their superiority over passive harmonic filters or reactors.

 

[dadfap]

jbartos

I would have thought that the passive filter would have been sufficient on a generator. On a mains system I would go for the active filter as the passive filter becomes a harmonic "black hole" and sucks up all the harmonics and may leave the "target" harmonic untouched.

dadfap

 

[jbartos]

Suggestion: Visit

http://www.squared.com/us/products/...2485256736006bf339/$FILE/D5800BR9801EP_R2.pdf

http://www.squared.com/us/products/...56a4d0067edd9/$FILE/AccuSine PCS Brochure.pdf

http://www.schneider-electric.com.s...08048256C60000B637D/$file/release-3211002.pdf

http://www.schneider.co.uk/products/dist/line_con.htm

etc. for more info
Passive harmonic filters are dependent on the system impedance, which is big drawback

 

[jOmega]

Most applications need to take down the 5th and possibly the 7th .... 5th harmonic traps are generally not tuned exactly to the 5th harmonic... there's some spill over towards the 7th and consequent benefit.

Passive filters, properly applied do a very good job. Most applications do not need the unwarranted expense of an active filter. They should be saved for the most severe situations. Proper harmonic analysis ahead of time can save a ton of money ... (hdwre + Installation costs).

 

[hydrae]

Fellow practicing Engs.
Thanks for your inputs

The client purchased and installed a 5% passive line reactor, works great now, it also improved minor vibrations when on normal power also.
I will know better next time...

Hydrae

 

[jOmega]

Hydrae:

Thank you very much for the feedback.

Glad the problem is resolved.

Kind regards,
jOmega

 

[jbartos]

Suggestion to hydrae (Mechanical) Oct 10, 2003 marked ///\\Fellow practicing Engs.
Thanks for your inputs

The client purchased and installed a 5% passive line reactor, works great now, it also improved minor vibrations when on normal power also.
///Was there a power supply quality analysis performed before the motor application and after the motor application?\\
I will know better next time...
///Perhaps, when the Utility comes to the Client requesting some further harmonic suppression equipment. Then, the reactor solution will have to be amended.\\Hydrae

 

[rconnett]

One other observation:

Bear in mind that, from a system standpoint, there are a number of control / system elements at work here - the AVR (as you all noted), the generator (speed) governor, mechanical spring time constant of the piping system, the drive characteristics, the ratio of kW load to motor load, presence of harmonics, etc. Adding the reactor may have dampened one critical portion of the system.

It seems intuitive to me that as the load increased on the system, the overall damping increased as well. When I hear about 'pipes shaking' it makes me think more of an overall systems response rather than zeroing in on harmonics & drive problems. (Remember the old 'poles' in the mechanical / electrical system transfer function - a few too many years since college I'm afraid here). An interesting approach might be to try slow down the time constant or response time of the AVR/drive/governor & see what happens.

That being said, when the problem gets solved, it's good for everyone.

 

[jOmega]

rconnett:

Have you ever seen a distribution sytem responding to an 'electrical' harmonic resonance ?

If severe enough, can be "Shake, Rattle, and Blow baby"

Not pretty !

 

[rconnett]

Actually, I saw 2 x 50MW arc furnaces at a steel mill shut down for a week because of a change in the overall system resonance point of a series of harmonic filter banks we had. Among other things, it also fried $300,000 worth of thyristors in an 100MVAR Static Var Control.....so yes, I have seen it & had to personally deal with the consequences.

My point was that mechanical vibrations (the pipes rattling)may arise from sources other than just harmionic resonance, and as EE's it's easy to focus in on just the electrical causes.