Regenerative braking using IPMs

I have attached a paper that discusses regenerative braking. What I am curious about is with the use of an IPM, my assumption is that the IGBTs (or Mosfets) are simply shut down and the free-wheeling diodes route the energy back to the battery. What I do not understand is why one would want to modulate (PWM) the low or high side drivers as part of the regen. There are two senarios; (a) to put back energy into the battery, and, (b) simply provide braking assist. If anyone has some specific paper they can attach that sheds more light on this subject, I would appreciate it.

Is there some specific protective circuitry that is necessary or desirable relative to the battery?
kenkad

I believe it is the same for IPM as for any PM AC motor or for BLDC. Regenerative braking always puts energy back into the source (battery) and produces torque opposing the direction of motor rotation (braking). Using the motor for braking without putting energy into the source is called dynamic braking and uses a braking resistor in most cases to dissipate the energy. There is no additional circuitry required in the inverter and control of regeneration is accomplished with the modulation. It was discussed recently here: http://endless-sphere.com/forums/viewtopic.php?f=30&t=56508

major,
I did look over the other thread before. I have attached the 180 sine modulation for the bridge (scope shot). Is the implication that only the low side drivers are driven? You do not have to worry about over voltage to the battery? Is the timing on the low side drive the same as normally driving the motor? Do you have any document that describes this in detail?

In the scope shot, #0 and 1 are hi and lo side of phase A, #2 and 3 are phase B and #4 and 5 are phase C. Even numbers are high side and odd numbers are low side.
kenkad

I don't have any references handy for you. And I doubt I would do it justice trying to explain it better than I attempted in the other thread. As far as overvoltage to the battery, the control algorithm takes care of it similar to how the current is limited. Bus voltage is compared to a set value and as it approaches that, the torque is reduced or the current limit is lowered. In the extreme case, torque goes to zero and no charge current flows to the battery. This can happen on the first stop or two hot off the charger or on a long downhill with a high SOC.

The paper I attached does not go into enough detail for me. I attached the 180 degree modulation scope shot to show that each of the six sequence states has a fixed number of binary intervals. The amount of on/off time determines the torque expected from the motor when the motor is being driven. When in regen, I assumed that there is some parameter that has to be monitored (Voltage?) that will determine the on/off times for the lower bridge so that the voltage is not over the battery limit (controls voltage rise). The binary intervals are fixed by the motor RPM. So if the vehicle has gearing, then it would be also important to keep the motor RPM at the opportune setting that the low bridge modulation would work the best for the voltage generated. I am sure that someone has already done this but I cannot find a actual paper that provides the detail I expect. In case anyone is curious or interested, I did ask the IPM manufacturer and have not received any response.
.
I guess we will see if anyone responds.
kenkad

OK, so we will try to add a bit more to the discussion and see if there is any response.

Attachment 1, discusses how the upper free-wheeling diode and the lower bridge are used for regenerative braking. Obviously, the Bemf voltage has to be higher than the battery supply voltage, else the battery gets nothing. So lets say that the nominal max RPM for normal driving conditions (battery max V and motor Kv) is 3000 RPM. In order to get the Bemf higher to charge the battery, we need to spin the motor, in regenerative mode, at an RPM higher than 3000, say 3600 RPM.

Now, Attachment 2 shows a graph of what the efficiency is of the regeneration based on the on/off time of the lower bridge element (50-60% modulation). I suspect that this is different for different motor designs. By how much, I have no idea. Anyway, it is a starting point. Looking back at Attachment 1, the regeneration is probably as low as 7% and may be as high as 12% (again, motor design must have an effect so 12% is not necessarily a max possible)? This again is just a guess from what they are suggesting.

So the question is how beneficial is regeneration? Yes, yes, we know that it is supposed to preserve the brakes! But at what cost? To keep the Bemf high enough, we need to keep the RPM high, so this would suggest some type of transmission. It seems something like with an ICE, you shift down and the engine high RPM helps slow the vehicle. If you do not go this route, then you need some type of variable boost converter that takes the Bemf, at whatever value it is, and boosts it to be greater than the battery voltage. It does not appear that the amount of current flowing is an issue.

This means, that some decisions need to be made, one way or another, in the design of the electric motor drive mechanics or an additional boost converter in the controller electronics. Clearly, some type of Bemf sensor (and possible current sensor) will be needed that has the ability to do an integration of the 3-phase Bemf waveform (a totally separate small microcontroller based PCB module with SPI, or CAN, or RS485?). This will be the data element that decides the lower bridge modulation rate and the motor drive mechanics or boost converter operation. Maybe a controlled variation of the modulation and a variable transmission gearing? Wow, this does not appear too pleasant.

Maybe regeneration is simply not worth the effort, just install a better braking system.

Any response to this additional information?
kenkad

kenkad said:

If anyone has some specific paper they can attach that sheds more light on this subject, I would appreciate it.

Click to expand...

I find this site really useful, http://www.modularcircuits.com/blog/articles/h-bridge-secrets/ ; the paper seems to talk about an asynchronous mode but sign-magnitude is more efficient.

kenkad said:

....the Bemf voltage has to be higher than the battery supply voltage, else the battery gets nothing.

Click to expand...

That is not necessary. Both of your attachments explain how the bridge is used as a boost converter during regeneration to boost the voltage above the battery potential. Motor RPM does not need to be increased.

In attachment 1, figure 4 shows this and is described as follows:

from Figure 4a, we can see that when T4 is ON, bypass diode D6, T4 and the motor form a closed circuit, the braking energy is temporarily stored in the inductance. When T4 is OFF, bypass diode D1, D6 and the battery form another closed circuit. The battery is charged by the boosted voltage. Charging current can be controlled by the regulation of duty cycle in pulse-width modulation (PWM) control.

Click to expand...

Note that the current direction in figure 4a is shown incorrectly.

In attachment 2, page 17 shows a graph of boosted voltage. At 0% duty, all 6 transistors are off and you simply have the diodes rectifying the motor's generated voltage. As they start modulation as a boost converter, the DC bus voltage increases even though the speed is constant and the generated voltage remains the same.

HOPE this YT vid is Germain to this thread? Check out these regenerative braking ebikes you guys sent in!
[youtube]RFmnPXsHQ_k[/youtube]

In essence, the motor (generator) is shorted which stores energy in the magnetic fields of the stator. Then the switches are opened, and the energy in the collapsing magnetic field causes the usual inductive effect of raising the voltage (boost conversion) until it finds a current path, which is via the diodes to the battery.