(AC) regen via BMS, good or bad idea

3594 Views 11 Replies 4 Participants Last post by Huub3, Jan 12, 2011

Huub3

Discussion Starter · Jan 9, 2011

Hello,

I am working on my AC setup, based on a big (too big

) ABB motor and a modified Hitachi SJ300. Similar setup to some of the AEVA members (kiwi4x4, coulomb, weber, woody, johnny etc) and comparable to what Eric Tisscher did on his Passat.

The SJ300 I have (the 75 kW version) does not have a brake chopper nor braking resistor build in.

These accessories are useful for normal AC operation, however in our EV we can do motor braking (=regen) in charging back to the batteries

However, certainly I want to avoid blowing up my batteries via the regen current, and here is my dilemma.

Limiting regen current is difficult, as it might interfere with a consistent driving/feeling, so I need to modulate not at the controller, but somewhere at the batteries, to avoid damaging them.

I was thinking whether this can be done via the balancing function of a BMS, in that the "over-" current is bled away via the balancing resistors, giving a distributed version of a normal braking resistor.

Related to this Idea I have two questions:
- Regen typically has high currents, well over 1C. This is higher than the adviced charging currents (mostly 1C, sometimes even as low as 0.3C). Do we have data/experience what these high regen-currents mean for battery life.
- who has experience with an AC+BMS set-up as described above. Is this a wise move? If yes, what power would typically be dissipated at the resistors. Do I need to beef-up the BMS to cope with the higher regen-currents?

Any other suggestions would be also very much welcomed.

Regards,

Huub

Save

1 - 12 of 12 Posts

· #2 · Jan 9, 2011

Regen has high currents but the real amount you will be pumping back into the batteries is negligible at best. I think when you need to worry the most is heavy regen down very long stretches of hwy with a FULL pack. A very unlikely scenario. Unless you start at the very top with a very full pack then I'd say you will be fine with regen.

I don't know too many folks who actually ride the brake down any type of grade for any more than a sec or two anyway. An occasional press on the brake to help control speed even in a coasting situation is more likely the truth. If you ride your brakes you are not in control.

Don't worry about regen. Use it when you need.

Pete

Save

tomofreno

· #3 · Jan 9, 2011 (Edited)

Limiting regen current is difficult, as it might interfere with a consistent driving/feeling, so I need to modulate not at the controller, but somewhere at the batteries, to avoid damaging them.

Not following you here. Regen is controlled by most controllers. It can be tailored as I want in mine, with different permitted max regen (expressed as a fraction of max controller current) permitted at different motor rpm ranges. The driving is very smooth. I don't see how you could control it with bms resistors since you must dissipate hundreds of Watt. You only want to limit current to the entire pack, so could have high power resistors mounted in some cooled enclosure away from anything flammable. It seems you have to use high current contactors and associated logic to switch these resistors in and out of the circuit (controller to batteries), and control how much resistance to switch in for different regen limiting at different deceleration rates. Not something I would want to do.

In normal driving with gradual slowing, regen currents are typically C/2 or less in my vehicle. They are only greater at higher deceleration rates. I have set the above controller settings so that it cannot exceed about 1C. That gives me enough to slow and stop the vehicle under almost all but more sudden, or emergency decelerations, with regen control through the accelerator pedal, touching the mechanical brakes lightly when the vehicle is almost stopped.

Not sure that helped you at all. Thought I should post something because I don't think you will get many answers from this group of mostly series DC motor users. I would think you would get more help from some of the people you mentioned at www.aeva.asn.au

Good luck. Don't set your vehicle on fire!

EDIT: If your controller always just puts out full regen upon deceleration, it will decelerate the vehicle extremely rapidly whether you have resistors dissipating some of the power or not. Not very drivable this way I would think.

See less See more

Save

etischer

· #4 · Jan 9, 2011

You want to reduce your max regen torque once you reach a certain voltage. In my conversion, I allow 100% regen up to 330 volts, and by 340 volts it has tapered down to 0%.

BMS resistors are designed to shunt around 0.75 amps.
Regen current would be upwards of 100 amps. I wouldn't rely on BMS to prevent an overcharge situation, but the problem can easily be solved programatically.

Save

Huub3

Discussion Starter · #5 · Jan 9, 2011 (Edited)

All,

thanks a lot for these answers.

Pete: good arguments to reduce my fears. No hills over here, and indeed you will not put more into the pack by regen that you took out of it in the first place to accelerate the car (over-unity discussion anyone

). My concern was therefore not so much with overcharging the batteries, but more with the much higher currents in regen compared to charging. But, indeed, I think my fears are not needed.

Tomofreno: what I meant was indeed that the regen is controlled by the motor controller, but that the regen current is depending on the state of the battery pack, meaning that you have different regen (=braking) behaviour with full and more empty packs. I was concerned this might be a safety risk. But your approach, making regen responsible for a normal ramp-down, but not too excessive, will mean that for emergency braking the natural reflex will remain to be hitting the brake pedal.

Eric: your response indicates that indeed regen should be made dependend on the battery state. I take it from your answer that this was something additionally that you programmed, and not something that the controller takes care of from itself (e.g. regen being dependent on bus voltage). Also, looking at the quality of your work and the effort you put into your car, I assume this solution is perfectly drivable, and not a compromise in the area of safety (braking behaviour) or comfort (driving feeling). Can you elaborate a bit on the way you did this programming. Knowing the SJ300, I know a lot is possible, but you have to think around corners a lot to achieve what is needed for our EV application (currently thinking/working on the "VFD OK" signal to close main contactor after pre-charge, not simple with the Hitachi I fear).

These have been absolutely answers that help me further a bit!

Regards,

Huub

See less See more

Save

etischer

· #6 · Jan 9, 2011 (Edited)

Huub3 said:
All,

Eric: I take it from your answer that this was something additionally that you programmed, and not something that the controller takes care of from itself (e.g. regen being dependent on bus voltage).

Huub

My inverter is preprogrammed from the factory to fault out and shut down if the voltage exceeds 400 volts. I added the regen taper function for a more graceful result. Anyone driving the car would be unaware that regen is changing.

My equation looks something like this:

Max regen torque = Max(0,0,(345 - DC bus voltage)) * 4

So with a battery voltage of 330v, max regen is 60%
If battery voltage >= 345v, max regen is 0%

I limit regen torque to 50%, this is about 100A

Save

Huub3

Discussion Starter · #7 · Jan 9, 2011

Eric,

thank you for this explaination. One question, you are running this calculation on the VFD itself? Or on some external processor?

I have not yet found a way to do custom programming on the SJ300. Closest I have found is the build-in PID controller, that I might misuse for this (but I was rather wanting to use the PID for something else).

So, trying to learn how to explore the boundaries of VFD programming.

Again, thanks a lot!

Huub

etischer said:
My inverter is preprogrammed from the factory to fault out and shut down if the voltage exceeds 400 volts. I added the regen taper function for a more graceful result. Anyone driving the car would be unaware that regen is changing.

My equation looks something like this:

Max regen torque = Max(0,0,(345 - DC bus voltage)) * 4

So with a battery voltage of 330v, max regen is 60%
If battery voltage >= 345v, max regen is 0%

I limit regen torque to 50%, this is about 100A

Save

etischer

· #8 · Jan 9, 2011 (Edited)

The VFD I'm using, has quite a bit of programming space available so all the math I'm doing is done in the drive.

Here is what my program looks like
http://etischer.com/awdev/program2.gif

Huub3 said:
Eric,

thank you for this explaination. One question, you are running this calculation on the VFD itself? Or on some external processor?

I have not yet found a way to do custom programming on the SJ300. Closest I have found is the build-in PID controller, that I might misuse for this (but I was rather wanting to use the PID for something else).

So, trying to learn how to explore the boundaries of VFD programming.

Again, thanks a lot!

Huub

Save

SimonRafferty

· #9 · Jan 9, 2011

I assume the regen current is proportional to the difference between the demand speed on the controller and the actual motor speed - so as you lift off the throttle, it applies regen braking to slow the motor down?

If this is the case, you could use a micro controller to modulate the throttle based on the pack voltage. If the pack is fully charged, when you lift off, the motor speed will stay the same. When it is discharged, the difference is limited to keep the pack below it's ideal charge voltage.

A few people on here have said that the energy recovered by regen braking is not worth the trouble. I disagree. I have an AC setup and where I live, it's relatively hilly. I am recovering about 8% of the battery capacity through regen. Making the vehicle nearly 10% more efficient is pretty good going!

I have mine set to charge at 4C (200A) but not to exceed 14.7v per battery.

The controller has a signal to switch off regen. This is enabled if any of the volt-blotchers detect over voltage on an individual battery.

Si

See less See more

Save

etischer

· #10 · Jan 9, 2011 (Edited)

That is my regen logic. Note, the numbers might not make sense because my laptop was probably offline when it was printed.

While I depress the brake pedal I get 50% regen
When I let off the brake, I get 15% regen
Once I press the throttle, I reset back to 0% regen until I press the brake again.

By tapping the brakes, I latch in 15% regen which maintains my speed going down hill. Tapping the throttle resets regen back to 0%. It behaves like cruise control in an ICE car (tapping the brake cancels cruise control).

I'd estimate I recoup about 5% from regen. It's nice to see 100A going back into the pack =)

SimonRafferty said:
I assume the regen current is proportional to the difference between the demand speed on the controller and the actual motor speed - so as you lift off the throttle, it applies regen braking to slow the motor down?

If this is the case, you could use a micro controller to modulate the throttle based on the pack voltage. If the pack is fully charged, when you lift off, the motor speed will stay the same. When it is discharged, the difference is limited to keep the pack below it's ideal charge voltage.

A few people on here have said that the energy recovered by regen braking is not worth the trouble. I disagree. I have an AC setup and where I live, it's relatively hilly. I am recovering about 8% of the battery capacity through regen. Making the vehicle nearly 10% more efficient is pretty good going!

I have mine set to charge at 4C (200A) but not to exceed 14.7v per battery.

The controller has a signal to switch off regen. This is enabled if any of the volt-blotchers detect over voltage on an individual battery.

Si

See less See more

Save

tomofreno

· #11 · Jan 10, 2011

Oh, sorry. My controller has overvoltage and undervoltage parameters. When the pack voltage reaches the overvoltage setpoint the controller stops regen. Otherwise max regen can be set by other parameters as a function of motor rpm as I said, or just given one max value for all rpm, such as 50% of max controller current.

Save

Huub3

Discussion Starter · #12 · Jan 12, 2011

All,

again, thanks a lot for all your replies, and details on implementation.

Getting a general answer can be helpful in setting a direction for own choices.

Getting so many relevant details as available in this forum, and as proven by the answers in this thread, is even better and helps avoiding mistakes and brings our EV's to the roads quicker, proving that electric driving is not longer a dream.

One time I hope to be able to give back to the forum what I took out of it. Many regards,

Huub

Save

1 - 12 of 12 Posts

This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.

Top Contributors this Month