Why does my petrol generator output voltage rise with increasing load?

[Odhiambo]

Dear All,
I originally posted this on the Knowledge Point website (knowledgepoint.org) -a great resource by the way for more development related off-grid questions. Someone on that forum suggested that this was the best place to ask such a question, so here goes. Your thoughts most appreciated.

I am trying to power up a micro-grid providing electricity to over 50 households and small businesses on a remote island in Lake Victoria (Kenya, East Africa). I am using a Victron 3kVA inverter/charger coupled with a solar/wind and storage battery set-up. I have been trying to connect a petrol generator to make up for a shortfall in power when the battery voltage gets too low. The 3.5 kVA rated petrol generator will charge the batteries through the inverter/charger if there is no load. As soon as there is a load (ie people using power) the voltage of the generator output rises. The more the load, the higher the voltage. Eventually the voltage rises above 270V and the inverter/charger cuts off the generator at which point it settles back to whatever I set the AVR (Automatic Voltage Regulator) to. The generator is a German-designed Golf.

Question: Why does the generator voltage rise? How can I get round this problem? Help! Any thoughts most appreciated!

[XeonPony]

Sounds like you may have a govenour issue, does the frequency rise with it? if so the govenor is set too tight needs more droop.

if the frequency stayes fixed there is another issue, possibly reactive current. flux will be more help there.

[joestue]

capacitive loads will increase the terminal volts, inductive reduces them.

I don't know off hand the orders of magnitude, but you might see 1 volt change in terminal volts per amp of reactive current.

the last generator i had, put out 260 volts no load, 240 under 6KW load, 60Hz for both.
it wouldn't take much to make it output 270 volts under no load, but under load it would quickly drop.

getting a small generator to 270 under full load certainly sounds like a regulator or governor problem.

you could have something interesting going on if the generator is a third harmonic resonant excited type.
you'll recognize these by a diode and capacitor located on the rotor, and there are no brushes.

[Flux]

It seems highly unlikely that this is an engine governor problem, rising speed characteristics with load are not usual with mechanical governors.

Battery chargers on small generators are something of a gamble, the old style rectified chargers usually went totally flat with small generators as the diode conduction caused a peak volt drop from the generator reactance.

Modern chargers usually have some form of power factor rectifier that draws current over most of the cycle and work much better, but certain waveforms may cause trouble, small generators often have up to 20% waveform distortion.

If the alternator has an avr it may have a simple voltage sensing circuit detecting near peak volts and flattening of the waveform may raise the rms ( or mean voltage).

If the alternator is of the capacitor excited variety then you are in the lap of the gods, they are very speed sensitive and regulate best with resistive loads.They drop volts badly with highly inductive loads and may react even more violently to highly capacitive loads as there is only core saturation to hold things down.

I can only suggest that you check the generator on its own with various loads and I expect this problem will vanish, in which case it will be an interaction with the inverter/charger. If the problem is still present with the generator on its own you will have to measure speed or frequency to see if it is a governor problem.

Flux

[Flux]

Having read your post again I see that you state that it is an avr generator.

I don't know how you are loading things but I was under the impression that an inverter/charger could only charge when it was not inverting. If you are loading via the inverter I think you will need a separate battery charger.

Most present day alternators in the 3.5kVA region are capacitor excited, if this is an older machine it will be built down to a price and the sensing of the avr could indeed be peak or very nearly so so a big rectifier biting off the peaks of the waveform could lead it to raise volts.

Flux