High efficiency hub motor

[Green team twente]

Hi, Im Joost Potma from the green team twente and we're competing in Shell eco marathon Europe 2012. We're designing a car that will use a fuel cell to convert hydrogen into electricity. For the propulsion we will use 2 hub motors (one in every rear wheel). Because the goal is to design the most energy efficient vehicle we are looking for the most efficient hub motors. The nominal power we need will be around 200-400W and the peak power will be around 1500W (only during acceleration for about 7 seconds). This is the total power needed (divide by 2 for the power per motor). The peak efficiency needs to be around 250 rpm. So which hub motor can deliver this power and is very efficient. Also a low weight is desired.

 

[Hugues]

surely someone will reply to your question but just out of curiosity:

i read quickly the rules for the competition:
http://www.shell.com/home/content/ecomarathon/europe/for_participants/europe_rules/

I believe you can use batteries as source of energy. If so, why do you bother using a fuel cell to produce electricity ? surely adding weight and complexity.

 

[Green team twente]

There are several different classes you can compete in. Like petrol, diesel, hydrogen, solar, battery etc. We choose to compete in the hydrogen class. So we will compete against other vehicles driven by fuel cells.

For the weight. The fuel cell we will be using weighs only 3.8 kg. I don't know how this compares to battery's but its not much.

 

[Hugues]

I see.
Someone will surely propose something.

In the meantime, checkout ebikes.ca, they have a nice simulator, surely helpful for your needs.

 

[dogman dan]

Set speed and race for economy or a race for time with limited fuel?

The sim will help a lot, if you are planning on using bike hubmotors. Assuming power on all the time, dd motors, if pulse and glide coasting, then planetary gearmotors for the freewheeling.

Pick a motor with a winding that matches your desired rpm at your voltage at full throttle I'd think. Too fast a motor would put you in a low rpm, poor efficiency zone too much, but too slow would be slower than you want after adding drag and lifting weight up hills if any.

 

[Alan B]

What voltage and current does your fuel cell produce?

 

[Green team twente]

We have a set time in which we need to complete 10 laps of the track. And the one which uses the least fuel to complete these 10 laps wins.

The fuel cell delivers 13-20 V depending on the load, but we will use a dc/dc converter to boost the voltage to set value like, 48V is the maximum voltage we can use.
And it can produce a maximum of 500W. For the acceleration we will use supercapacitors to support the fuel cell and we will use them to store the regenerative braking energy.

 

[John in CR]

For max efficiency the motor you want is the 98% efficient CSIRO wheel motor http://www.csiro.au/en/Outcomes/Climate/Reducing-GHG/Solar-cars-use-CSIRO-motor.aspx . With ebike hubmotors you're looking at 85% peak efficiency on their best day, and doubtful at a voltage low enough for peak efficiency at 250rpm. I'll have a hubmotor in hand for testing soon that claims a 93% peak efficiency, but that's at almost triple your desired rpm.

 

[dogman dan]

But if the budget is low, I don't really know that one type is much more efficient than the other. All about the same not so great 70% efficient I think. Perhaps the larger rotors of a magic pie could be an advantage at the start up, but once on the track cruising at a set speed, nearly all dd hubmotors will get the same wh/mi pretty much.

More than anything else, it sounds like an aerodynamics contest. Everybody will be stuck with similar efficiency in the motor if using common scooter and bike stuff. You'll want the slower windings I'd think, for 250 rpm. Which one depends on the actual voltage the controller is fed.

 

[Drunkskunk]

A CSIRO would be the way to go, if you have the budget for it.
But you can boost your efficancy by eliminating the DC/DC converter and running 2 fuel cells in series

 

[Lebowski]

I would lose the DCDC converter. I tried to go this way but it's not efficient. Better is to rewind your motor
to the low fuel cell voltage, then you'll eliminate the inefficient DCDC converter.

Think about it... A DCDC boost converter works by turning a voltage into a current (DCDC's inductor), then
the current into a voltage (the 48 V), then back into a motor-current. At every stage you lose energy.
If you think about it some more you'll realise the motor with its controller-switches actually works like a
DCDC down converter, this down-converter uses the motor windings to deliver power to the back-emf
voltage source. It is even possible to use the no-DCDC converter topology but to use the motor inductor
to boost the voltage, the motor then runs faster than you would expect based on battery voltage
(in the motor world this is called field weakening but electronically speaking you're building a boost converter)

One more remark about DCDC, the real killer there is the ESR of the capacitor combined with the
very spiky nature of the input current of the motor controller (due to the PWM)

As far as the motor goes, there's some simple equations you can do which lead to P_loss/P_delivered
being low for:
- big motor volume (so big diameter and thickness), get 4 hub motors instead of 2. Motor current in each will
drop by half -> you only have half the power loss.
- lot's of copper content (many windings with thin wire (high volt motor) or little windings thick wire (low volt motor)
has no influence, it'e the amount of copper that counts)
- strong magnetic field
- high motor rpm, hub motors are at a real disadvantage here. Go to small rims if you really want a hub motor

And last, get a good controller. Don't get a simple hall sensor controlled one, the least you should get is
something that implements Field Oriented Control (FOC) but I myself think you can do even better than that

 

[Green team twente]

Thanks for all this information. The problem is that we are all mechanical engineering students. So taken in to account the time and our electrical knowledge I think its best for us to use standard part, otherwise I think it will become to difficult for us we won't have the car ready on time.

 

[Hugues]

Lebowski ! You scared the shxx out of the guy ! :wink:

Joost, don't worry, Leboswki is a good guy. But just between you and me, I'm also a mechanical engineer and i don't understand what he's saying most of the time. This guy is a genius.
If he did not put Zurich in his signature, i would think he's from another planet.

Now, coming back to your race, and i'm not going to help you much here,
if I were you, i would change class and go to the battery group, then connect as many LiPo's as you can put your hands on,
get that Cro hub they're talking about in the other thread, and those fancy 22 dollar a piece Mosfet, some 4 gauge wires,
light up the whole thing and leave those fuel cell guys behind looking at you as if they just saw Pamela Anderson running half naked

That may not be the most efficient setup, but it will surely be a truck load of fun !

 

[Lebowski]

.The problem is that we are all mechanical engineering students

a problem indeed

Under these difficult conditions I would get these motors:

http://www.goldenmotor.com/ go to hub motor (magic pie)
http://www.goldenmotor.com/magicpie/features_magicpie.html

I have no experience with them but they full fill some of the requirements I mentioned earlier. They're physically big,
lots of strong magnets and available with rims down to 16" so the motor speed will be high. Make sure the small
wheel size doesn't limit your top speed !

Try to get 2 fuel cells, connect them in series and get a 24V or 36V version of the Magic Pie

How do I get my hands on one of those fuel cells ?

 

[Green team twente]

I've looked at the magic pie but @ 36v 250 rpm the efficiency is only 75%, and I have my doubts by the quality. So is there a hub motor that has a efficiency > 80% @ 250 rpm (24-48V)?
The tire we will be using has a outer diameter of 22 inch. This is a tire that Michelin specially makes for the shell eco marathon, which has a extreme low rolling resistance of 0.0015.

Using 2 fuel cells is not an option because this will be to expensive and the power needed from both fuel cells will be divided by 2. If you lower the power the efficiency goes down also.
Fuel cells can be ordered from MES DEA located in Switzerland.

 

[Drunkskunk]

9 Continents makes some motors that run up to 85%

But at the low voltage you're going to be running, a hub motor isn't the best option. DC/DC converters will be 10-30% inefficent, so avoiding using them and running a slightly lower efficent system may be a far better option.

Have you considered RC drives? there is drivetrain loss, but the motors are more efficent and they will opperate at the voltage levels of a Fuel Cell

 

[Lebowski]

I've looked at the magic pie but @ 36v 250 rpm the efficiency is only 75%, and I have my doubts by the quality. So is there a hub motor that has a efficiency > 80% @ 250 rpm (24-48V)?

Just make sure you compare apples with apples.... the magic pie, according to their website, has the
controller build-in, so I can imagine their efficiency is the combined controller-motor efficiency
while pure motor manufacturers only post motor efficiency.

 

[Green team twente]

I think that if we look at standard hub motors the Nine Continent 2807 @ 36V is the best option.

We also need 2 high efficiency motor controllers (one for each motor). What motor controllers can you advise for the 9C 2807 @ 36V?

 

[hjns]

I think that if we look at standard hub motors the Nine Continent 2807 @ 36V is the best option.

We also need 2 high efficiency motor controllers (one for each motor). What motor controllers can you advise for the 9C 2807 @ 36V?

I think you are missing the point. The DC-DC converter that you need to get to 36V will create an efficiency loss of 30% and will produce a lot of heat. That means that whatever you are gaining with a superefficient motor at 36V is completely lost due to the DC-DC converter as compared to directly feeding the low voltage into a controller driving a hubmotor.

If you aim for 250 rpm with a low voltage, and you do not want to build a mid motor but are stuck with a DD or geared hubmotor, then try to feed the voltage from the fuel cell directly in to the controller. Then you need a fast motor to get to 250 rpm at 13-20V (that's what you told us the fuel cell delivers).

So, why don't you try the simulator here: http://ebike.ca/simulator/
I used the following assumptions:

• 16V under load, directly from the fuel cell
• 22 inch custom wheel
• 150 kg gross weight

With that, for example a Crystalyte 5302 will get you 205 rpm at 75% efficiency 200W max, which is pretty close to what you were aiming for.
Two of these motors will get you 400W. It will pull a maximum of 55 Amps, but once you are over 100rpm, you only need 35 Amps (per motor). So get a sensorless controller and modify it so that it can feed 40 Amps to the motor. Might as well get 2 of these Lyen controllers add some solder to the shunt, and be done with it. However, I do not know where to purchase a Clyte 5302...

Don't forget to use two independent throttles. Search on this forum for 2WD (2 wheel drive) setups to understand why.

Missing still some info: how much current can that fuel cell provide? At 20V, will it deliver 25 amp and provide indeed 500W???

 

[hjns]

Missing still some info: how much current can that fuel cell provide? At 20V, will it deliver 25 amp and provide indeed 500W???

In addition: did you realize that your DC-DC converter may be able to up the voltage, however at the cost of current? Apart from the efficiency loss, even at 60V max (which is allowed according to the rules), you still need 8.3 Amp in order to get to 500W. At 60V 8amp, a Crystalyte H3525 will run very efficient (+80%) over a large rpm range (from 200rpm to over 300rpm).

 

[Green team twente]

Here is a graph of the fuel cell performance. The values are per cell and our fuel cell will have 22 cell, so multiply all values (except current) by 22.

 

[markobetti]

http://www.cebi.com/cebi/content/index_html?a=8&b=151&c=222

some interesting stuff for MES-DEA cells

 

[Green team twente]

How about a dc/dc converter like this one: http://zahninc.com/1236300.html

It has an input of 10-20V and a output of 30-40V which seems to be sufficient for our needs and the efficiency is >90%

 

[hjns]

Wow, that fuel cell is able to provide between 462 and 605 Watts. Is thisthe correct one? And you will be using 0.4L or 1.0L hydrogen as fuel, leading to what capacity in kWh? In other words, will you have near unlimited current given the high energy density of the hydrogen and given the maximum voltage of 60V (48V nominal)? What is the target speed, and what is the total distance?

I read the rules again. Three or four wheels. Max weight 140kg / 205kg depending on class. Rims 13-17 inch in diameter. And the car with the least kJ per km wins, with the achievements of the car expressed as km/kWh.

That means that with a DC-DC converter you indeed would have more than enough Amps available. That changes things for me. In this case I would indeed consider to use a nice efficient DC-DC converter, go to the maximum of 60Volts if possible, and limit the current to such an extent that you remain within the 500W limit. Upping the voltage will increase efficiency, because for low voltage systems at higher speed, the largest losses are in the need for running high currents through the cables.

At 60V 8.3A (500W), you can pick any motor that is in your budget, because all of these hubmotors will easily eat the 60V with the small currents you are running. And due to the small currents, all will run very efficient, you do not need heavy cables, and weight will then become more important. And your fuel cell also weighs 4-5kg. Therefore, you may also want to reconsider 2WD, because the added weight of the extra hubmotor and the controls is significant and may reduce your efficiency as compared to a single wheel drive / trike. And we are talking about a 5kg motor + whatever controller etc, so that is 2-5% weight reduction as compared to the total allowed weight of the vehicle.

Work the sim, and simulate the range of efficiency for 48V and 60V. At 60V 8A you have a much larger range of efficiency as compared to 48V 10A. With the 9C 2807 you are above the 80% efficiency from 220rpm up to 400+rpm.

Regarding the DC-DC converter, I am absolutely no expert. However, as mentioned above, I would go for a 60V output, with max current of at least 10A (600W max). If that is not available, get a 12-48V converter with more than 10A (500W) The linked converter only has 8.3A at 40V (300W) which is much lower than the 500W that you should aim for. Or use 2 of thesein parallel....

Better hold any action until you get a reaction from a real guru on the ES forum....

 

[hjns]

BTW, you will want to get a Cycle Analyst. This will allow you to easily fine tune the total amount of current going to the controller, as well limit the maximum speed. Well worth the price. Oh, and this is another reason to keep to a single wheel drive. Keep it simple.