Sizing the Dump Load

[jimovonz]

I have been using this dump load now about five months or so with no apparent problems. I use a picaxe to monitor battery voltage and switch a bank of fets at around 1kHz. The  turbine that this is on is 12V, 3.6m in diameter, cuts in around 140 rpm and furls between 350 and 400rpm. Its quite a windy site and I see over 70A on occasion. The dump load has no trouble maintaining the set voltage but I noticed the other day that  it doesn't seem to get as hot as I expected. I had previously put this down to the good air flow it recieves whenever it dumping. After some thought I realised that the low resistance of the dump load compared to the alt meant that most of the power that was being dumped was being lossed in the stator rather than where I intended it in the external element. Due to the relatively low power requirements and the small battery I have at the site, the turbine is just about always dumping some portion of what its making. By my estimation, if the turbine is putting 1kW into the battery its  running at around 45% (electrical) efficiency and loosing some 1.2kW in the stator for a total of 2.2kW. There are 12 coils in the stator so hopefully this isn't pushing things too far (after reading DanB's post here I used 30% calcium carbonate in the resin). If that same power level is maintained but entirely switched to the dump load, the stator is all of a sudden dissapating 50% more at around 1.8kW. This is not something I had previously considered.

In the interest of prolonging the life of the turbine should the dump load be sized as close to the stator resistance as possible so that it shoulders its share of the load?

[willib]

i'm afraid i have more questions than answers ,is the picaxe the shunt regualtor you talk about in the other post?

are you using your hall sensor / ac current measuring device , with your picaxe?

how many Fets are you using? and is the picaxe putting out PWM to drive your dump load?

you could make the dump load resistance twice the stator resistance this would drop the power dissipated in the stator.

how many feet of what gauge wire are between the alt and the battery & dump load?

[SamoaPower]

I think there's a little confusion here as to the role of a dump load. The idea of a diversion load is two fold. One, and primary, is to provide a charge algorithm for the batteries. Two, is to supply a continous load for the machine for protection.

Since you're using PWM to control the current proportioning between the batteries and the dump load, based on the PWM voltage set point, the current from the machine doesn't change for a given set of conditions. The stator doesn't know how much is going to the batteries and/or to the dump load, nor does it care.

Alternator efficiency is primarily determined by the stator resistance and the output current for a given air rotor power input (wind). So, changing where that current is directed has no effect on the efficiency or stator dissipation.

The dump load resistance should be chosen so that it can dissipate the maximum power ouput from the machine at the PWM voltage set point, plus a bit for the odd gust. If you change the dump load resistance, the PWM will adjust its duty cycle to try to maintain the set point voltage, but if it is outside the range of the PWM, you will have a problem.

I have two concerns for your system as described. What you indicate is that you simply have an upper limit to the battery charge voltage. You don't mention a means to terminate the charge when the battery is full of to switch to a float mode. This is very minimal charge control and will result in decreased battery life unless you monitor closely.

Secondly, dissipating 1.2 kW in the stator would worry me a lot, particularly with a cast stator. I'm afraid it's a failure just waiting for the right wind to happen. An efficiency of 45-50% is not a requirement. Much better can be done.

[jimovonz]

Hi willib. In order....

I talk about picaxes all the time, which post?

This particular setup doesn't have a current measurement (though I have one that does so using the hall effect sensor). I just use the clamp meter.

I am using 4 x IRF3205 and a good sized heat sink, well exposed to the wind.

I am not using the pwmout from the picaxe but rather just 'manually' controlling the pin. I basically turn the fets on if the voltage is over the threshold and turn them off if under. Works out to around 800-1000 Hz. The pwm on the picaxe has a min frequency of around 4kHz. I wanted to keep the gate drive as simple as possible and because the battery/dump load doesn't care about the frequency, opted to keep it low.

Too high a dump load resistance would limit how useful it was as a dump load...

The battery is at the foot of the 30ft pole that the turbine is mounted on. I'm not sure of the size - it was some 3 phase power cable I salvaged from a disused plant at the local pulp mill. I think it worked out to between 8 & 9 AWG. Surprisingly flexible for its size.

[jimovonz]

Hi SamoaPower, I believe I have a reasonable grasp on role of the diversion/dump load. This specific application (wireless repeater) has relatively low power requirements (averages 2A draw @ 12V) and the turbine is well over sized given the level of wind at the site. I was simply putting something together that works using the parts on hand. Best result for least effort. The batteries are two well used (EV) T105's that I got for $NZ50. The batteries do surprisingly well considering. I can get up to 5 days on a full charge without the turbine (depending on how far I want to push it). Even from dead flat the diversion load will activate when the current is over 65A. I top up the batteries at least every two weeks and have never exposed the plates. Additional steps in the charging levels would be easy enough to implement with the picaxe.

That is what I always assumed. I may be over thinking this here, but if I am switching in a low resistance in parallel with the battery, even for a short time, is n't the additional power loss still in proportion to the relative resistances? To take it to the extreme: say the battery is full and requires very little current to maintain my preset voltage. My chosen dump load is very low resistance - a dead short. Isn't most all of the loss now in the stator?

I agree. But it works. This particular turbine has all coil leads accessable. They are currently in series/parallel per phase leaving me the option of placing all in parallel and experiement with a boost converter. Also the option of all in series and  using a buck converter... When/if I get time....

[Flux]

As long as there is a battery in the circuit, the dump load mean current will match the incoming current,

From your details on the dump load, it seems as though your 70A runs the dump at about 50% duty cycle. During each cycle of discharge pulse part of the current will come from the turbine direct and the remainder will come from the battrey, this will be returned during the fet off period.

The only thing the alternator will see is a negligible reduction in battery voltage for the on period of the pulse and obviously the lower the battery volts the worse the stator dissipation but we are talking millivolts and it's not relevant.

If you have a second rectifier feeding the dump, so it is dumping before the battery then the mean current will still be the same but there will be an increased rms current in the stator winding as the battery will not supply a component.

Possibly better on the battery ( but they don't seem to mind fast pulses) and worse for the alternator, but again a small effect.

1.5kW in a wind won't produce much temperature rise.

[Flux]

Perhaps I should make it clear that the last line was referring to the dump load( not the stator).

[jimovonz]

Thank you. Just one question, does that mean my system would work just as well by shorting across the battery (very low resistance dump load) in which case my fancy dump load element just serves to limit the current to something the fets can cope with?

[Flux]

Good question.

I suppose you can take anything to extremes and reach a point where you run into trouble.

Ideally the dump load should be sized to take care of the maximum possible charge current during a gust. Within reason using a lower resistor will be ok, but there comes a time when the peak current spikes drawn from the battery are not going to be good for it and as you say the peak current rating of the fets becomes silly.

The alternator is certainly not going to be bothered by this as long as you don't cause battery problems.

Pwm charging and pulse charging is considered to be better for batteries than smooth charging current, but somewhere there must be a limit to the current pulses.

Having a dump load that can dump twice the charging current is probably as low as you would want to go from a practical point of view.

Perhaps someone else has experience of large current pulses on battery life.

Flux

[ghurd]

"...as long as you don't cause battery problems."

Would it be a good idea to have a smaller dump load first, given the small battery bank?

Why risk damage dumping 131A, if half the time it is charging at 2A?

The line of thought... It isn't making 70A often.

A 10A (25A?) dump load first, then the large dump load at a slightly higher voltage.

The lower power dump load would handle the extra power most of the time.

The large dump load 'hours' will would be reduced, the C/1.7 131A battery drain will be less often, reducing chances of any battery problems.

G-

[Flux]

I originally assumed that these pwm dump controllers had inductors and freewheel diodes, in which case the battery would just see a variable resistor. The fact that the majority seem to manage without any inductance seems to indicate that within reason batteries don't object to this treatment.

Most of this operation seems to be in the region of surface charge so perhaps this limits the peak value of the pulse.

An interesting subject for experiment if ever I can find the time.

There are claims by some people that high discharge pulses are effective at desulphating, but that is another grey area that I don't have answers to.

Ghurd I am inclined to agree with your 2 stage pwm idea for high currents in relation to battery capacity.

Flux