3 phase high current ripple suppression

1. Oct 23, 2012

chemnut220

I am building an on-board generator for a small aircraft (AUW = ~ 10kg). The fuel motor is rated at 1.8 kW and it directly drives a 3 phase AC motor. The rpm will be regulated at roughly 8k so I think the frequency of the AC output will be around 133 Hz. The AC motor produces 1 volt per 260 RPM (260 k/V). So I am looking at roughly 30 volts. I want to run this generator in parallel with batteries (8 cell lithium polymer) and let the generator maintain the batteries at 30 volts.

The problem that I am having (I am a chemist and have only taken a few college classes on electronics – coupled with a lot of tinkering and reading over the years) is managing the ripple voltage under load. The generator will experience constant loads of ~30 amps and spikes potentially approaching 60 amps (I understand that my fuel motor may not keep up with those spikes but it will be in parallel with batteries).
When I made this on a smaller scale I just used passive filtering (caps) and it worked great. But by my calcs I will need huge caps to manage the ripple on larger scale which equates to a lot of weight (.68f cap bank was around 1.5 kg – not to mention the space it would occupy and what it could do to my rectifier when charging them).
I started reading a lot about active filters and sophisticated rectifier circuits but I am not sure where to focus my reading.

What I am hoping to get from this forum are some tips/pointers on where to focus my reading, specifically regarding what circuits are out there that can manage the output of the AC motor and keep the ripple voltage to around 5-10% of my of the rectified DC signal – and not weight more than ~ 500 grams. I started reading about Vienna rectifiers and it looks like it can output a relatively clean DC signal? Is this the direction I should be looking to go – PWM rectifiers? What does that gain me?

I know I am not going to get a magic answer and I have come across several theses/dissertations on this topic (it seems quite advanced), but I am several months (and several thousand dollars) invested in this endeavor and do not expect over-night results – it’s a hobby/ obsession…

Any discussion/guidance will be much appreciated.

Thanks,
- Joe

Edit: i forgot to mention that the primary load on the generator is another 3-phase AC motor that is used for powering the aircraft. So maybe the smarter solution would be looking at circuits that take a 3 phase AC output and regulate it so it will go directly to the motor used to fly the plane... showing my ignorance...

Last edited: Oct 23, 2012
2. Oct 24, 2012

Enthalpy

If you connect a three-phase generator and motor directly:
- They will have a fixed ratio in their rotation speed
- You won't be able to store electricity in a battery between them.
Is this acceptable to you?

Some converters exist from and to three-phase, but presently they all go through DC.

Beware that batteries, especially lithium, need subtle and precise charging circuitry. These are fragile parts; for instance the end of charge is difficult to assess, and they may explode, catch fire, spat corrosive liquids and toxic gas. To the very least, separate people from batteries - as a chemist, you know that.

For a limited capacity, supercapacitors are easier to use, live longer, are more sturdy.

-----

Ripple filters tend to use inductors, sometimes in conjunction with capacitors. But in your case, and despite the ripple (from the three-phase rectifier) is only 4V at 800Hz, the inductor (which must not saturate at 60A) is too heavy.

A first improvement is to add a three-phase transformer which, with split secondary, transforms 900W to a separate three-phase with a 30° lag. Each three-phase gets a rectifier, their DC outputs are connected in series, and now the ripple shrinks to 1-cos(15°) or 4 times less and at 1600Hz, which is easier to filter.

[They make it with 13GW and 1MV at Itaipu ]

-----

At 1800W I'd say the lightweight solution is indeed a PFC rectifier, but it's seriously complicated! It's basically one resonant AC-to-DC for each pair of phases, and they must be driven properly.

http://scholar.lib.vt.edu/theses/available/etd-08142002-075617/unrestricted/Barbosa_ETD.pdf
on page 6

I doubt you'll find one to buy for 30V. Computer power supplies have one but for single-phase 110 or 230V.

3. Oct 24, 2012

Enthalpy

Some estimates for a 900W 115Hz three-phase transformer. Say, 15V peak and 40A peak per phase.

Let's take columns of 4cm2 and 0.1m height, of 400Hz laminations used at 1.3T peak. Each of the 6+ coils can have 40 turns of 0.09m*0.1mm aluminium foil.

(DC!) resistance is 14mohm, conduction losses are 70W which heats a lot, alu weighs 0.7kg and iron 1.6kg.

Not brilliant, is it?

-----

Estimates for an inductor that filters the ripple at 800Hz (not 1600Hz). Say, we need j2.5ohm at 800Hz or 500µH, and 60A DC.

The gap storing 0.9J DC needs 1.5T and 1cm3. Take 9cm2 iron and 1.1mm gap for 22 turns: the coil can use 0.05m*0.5mm aluminium foil. Resistance is 5.7mohm, loss is 21W to be cooled.

Aluminium mass is 0.36kg, iron mass 1.3kg, so:
- Not brilliant
- Copper foil would be better
- 9cm2 can be reduced a little bit
- Filtering at 800Hz looks less bad than the transformer!

A capacitor following the inductor can have j20S at 800Hz or 4000µF >30V which is small. You can increase the capacitor and reduce the inductor a bit.

4. Oct 24, 2012

chemnut220

Ugh – I forgot that tiny little fact the motor RPM (for my motor) is frequency dependent…

Batteries in parallel: I will work on managing the batteries later – I think/hope that will be much simpler.

Thanks a ton for providing those weight figures – the transformer seemed so promising and I had not even considered it.

I would love to use supercaps but I am unable to find any of them that meet my requirements. It looks like the majority of them are intended for lower voltage use, and those that are made for high voltage seem to be rated in the range of ~80 farads and look massive (not to mention expensive).

I guess it is time to start digging in to some literature on PFC rectifiers – that thesis will be a nice starting point. In the end, I hope that learning my way around this type of circuitry will provide for a general approach to light-weight, compact, higher power AC-DC rectification; and will be worth the time invested.

Like I said – over-night solutions here were not expected and what I was looking for here was for a nudge in the right direction. I didn’t want to spend months working on a complex circuit only to find that due to my lack of EE background, I overlooked a viable, simple solution.

5. Oct 24, 2012

chemnut220

If I could change my setup to substitute the direct coupled AC motor for an induction driven AC motor that operates at a much higher RPM, say 40K, which would bring my rectified frequency to 4kHz; would that gain me anything? It looks like I would need much less filtering capacitance at that frequency - but what would that do to a capacitor bank to operate at such high frequency?

6. Oct 25, 2012

Enthalpy

You mean, a higher frequency at the generator? For the electronics, it has only advantages. Even if electrolytic capacitors feel uneasy, there are other technologies (ceramic, plastic...). The coil alone, which gets smaller, could also filter without the help of a capacitor.

Then there's the mechanical limit. Faster generators and motors are lighter because in electric machines, only force has a cost, not peripheral speed. If you can rotate a generator by a turbine, it's much smaller than by a piston engine. In 1400MW power plants, turbo-generators have some 100m/s and are smaller than the turbines.

Though, you need to power a propeller which can't rotate at arbitrary frequency. A gear may well offset the weight advantage of smaller motor and electronics.

-----

About power electronics: a PFC circuit is difficult to develop for an electronics engineer skilled in analog cicuitry. Chemistry can be an excelent background for it, as one of the very best electronics engineers I met had a degree in chemistry. But it takes incompressible years and years to learn. Interesting, though.

Once you have DC in the battery, you'll still need to produce three-phase for the motor. Look at existing drives for squirrel-cage motors: at 1800W they aren't light. For sure, if developed for an aeroplane, weight would be optimized.

Supercapacitors can be connected in series; this may need a special charge monitor circuitry. Presently, parts with a higher voltage are used in autobus, hence the excessive size...