1DA Torque Reading

[lucamifbuh]

Hello all,
This is my first post on this forum, but I have been actively following the community for a while.
I wanted to see whether anyone has some kind of experience with the 1DA torque value on the EV-CAN. I am a student working on a dissertation on regenerative braking, and I am using data from the CAN bus to facilitate data collection. However, I cannot understand something with the data and was hoping that while contibuting some information (I intend to make the data available to the community following completion of the project), someone can give a hypothesis on this anomaly.

So the below graph shows the test. I accelerated to a constant speed, and then used regenerative braking to slow the vehicle down (it is in standard Drive mode - the car is a AZE-0 30 kWh model). I plotted two power values, one from the battery (voltage * current*-1) and one from the torque value on the 1DA message (torque * RPM *(2pi/60)). I expect there to be a gap between the two owing to inefficiencies. However, it is clear that the gap is far larger when accelerating than decelerating. I do not have any explanation for this. Any ideas are greatly appreciated.

Thanks in advance!

 

[coulomb]

Not an actual answer, but I wonder how they are measuring mechanical torque. I doubt that there is an actual torque sensor, so they are presumably inferring the torque by some sort of maths based on motor currents, motor voltage, and motor speed. This inference may be flawed.

 

[lucamifbuh]

Not an actual answer, but I wonder how they are measuring mechanical torque. I doubt that there is an actual torque sensor, so they are presumably inferring the torque by some sort of maths based on motor currents, motor voltage, and motor speed. This inference may be flawed.

Yes, that did cross my mind. It is definitely a possibility. Although for the accuracy of the results, I am hoping that there is a more physical reason for it. I am also trying to understand what exactly uses the output. The dashboard?

 

[lorenfb]

Not an actual answer, but I wonder how they are measuring mechanical torque. I doubt that there is an actual torque sensor, so they are presumably inferring the torque by some sort of maths based on motor currents, motor voltage, and motor speed. This inference may be flawed.

It's battery resistance (effective) losses at different power levels (accel versus decel), i.e. I^2 * R losses,
that result in the power differential.

 

[lucamifbuh]

Not an actual answer, but I wonder how they are measuring mechanical torque. I doubt that there is an actual torque sensor, so they are presumably inferring the torque by some sort of maths based on motor currents, motor voltage, and motor speed. This inference may be flawed.

It's battery resistance (effective) losses at different power levels (accel versus decel), i.e. I^2 * R losses,
that result in the power differential.

The strange part to me isn't the differential, it's the substantial difference between the acceleration and deceleration differentials (nearly nonexistent in deceleration). Why would I^2*R losses only occur with current flowing in one direction, not the other? The power level isn't substantially different between acceleration and deceleration. If it helps, this test is in D mode. I tried it in B-mode and got the same result. Image below.

 

[coulomb]

Ah, brilliant! Thanks. There are of course roughly symmetrical losses charging or discharging, but the measurement system can't see the "actual" battery voltage "on the other side of the internal resistance".

So if the voltage at the battery terminals rises from say 360V (heavy discharge) to 400V briefly at zero current (so, 400V is the "actual" battery voltage), and rises to say 420V under regen, it looks like the battery is getting 420V times the charge current, when in reality 20/420ths of that is a hidden battery loss. But the system sees the full loss on the discharge side (380V times discharge current).

 

[lorenfb]

Not an actual answer, but I wonder how they are measuring mechanical torque. I doubt that there is an actual torque sensor, so they are presumably inferring the torque by some sort of maths based on motor currents, motor voltage, and motor speed. This inference may be flawed.

It's battery resistance (effective) losses at different power levels (accel versus decel), i.e. I^2 * R losses,
that result in the power differential.

The strange part to me isn't the differential, it's the substantial difference between the acceleration and deceleration differentials (nearly nonexistent in deceleration). Why would I^2*R losses only occur with current flowing in one direction, not the other? The power level isn't substantially different between acceleration and deceleration. If it helps, this test is in D mode. I tried it in B-mode and got the same result. Image below.

There should be no battery charging in D mode by the motor controller, although the voltage of the windings is there.

 

[lucamifbuh]

It's battery resistance (effective) losses at different power levels (accel versus decel), i.e. I^2 * R losses,
that result in the power differential.

The strange part to me isn't the differential, it's the substantial difference between the acceleration and deceleration differentials (nearly nonexistent in deceleration). Why would I^2*R losses only occur with current flowing in one direction, not the other? The power level isn't substantially different between acceleration and deceleration. If it helps, this test is in D mode. I tried it in B-mode and got the same result. Image below.

There should be no battery charging in D mode by the motor controller, although the voltage of the windings is there.

Am I misunderstanding this, or are you saying that the vehicle does not perform any regenerative braking without pressing your foot on the accelerator when on Drive mode?
I can confirm that there definitely is regen on drive mode, per my first graph, at least on my AZE-0.

 

[FinnishCoffee]

Am I misunderstanding this, or are you saying that the vehicle does not perform any regenerative braking without pressing your foot on the accelerator when on Drive mode?
I can confirm that there definitely is regen on drive mode, per my first graph, at least on my AZE-0.

Can also confirm that both ZE0(2012) and ZE1(2023) also regen in D.

 

[lucamifbuh]

Ah, brilliant! Thanks. There are of course roughly symmetrical losses charging or discharging, but the measurement system can't see the "actual" battery voltage "on the other side of the internal resistance".

So if the voltage at the battery terminals rises from say 360V (heavy discharge) to 400V briefly at zero current (so, 400V is the "actual" battery voltage), and rises to say 420V under regen, it looks like the battery is getting 420V times the charge current, when in reality 20/420ths of that is a hidden battery loss. But the system sees the full loss on the discharge side (380V times discharge current).

Thank you so much for the input! I purposely waited a little to reply to have a thought about this.
I will definitely look further into it, but it seems to explain what I am observing here quite well.
The LEAF community is truly amazing.

 

[lorenfb]

The strange part to me isn't the differential, it's the substantial difference between the acceleration and deceleration differentials (nearly nonexistent in deceleration). Why would I^2*R losses only occur with current flowing in one direction, not the other? The power level isn't substantially different between acceleration and deceleration. If it helps, this test is in D mode. I tried it in B-mode and got the same result. Image below.

There should be no battery charging in D mode by the motor controller, although the voltage of the windings is there.

Am I misunderstanding this, or are you saying that the vehicle does not perform any regenerative braking without pressing your foot on the accelerator when on Drive mode?
I can confirm that there definitely is regen on drive mode, per my first graph, at least on my AZE-0.

Yes, my error, but the regen is significantly less (< 50%) than in B mode. In both modes, regen is a function of SOC and battery temp.

 

[lorenfb]

The strange part to me isn't the differential, it's the substantial difference between the acceleration and deceleration differentials (nearly nonexistent in deceleration). Why would I^2*R losses only occur with current flowing in one direction, not the other? The power level isn't substantially different between acceleration and deceleration. If it helps, this test is in D mode. I tried it in B-mode and got the same result. Image below.

There should be no battery charging in D mode by the motor controller, although the voltage of the windings is there.

Am I misunderstanding this, or are you saying that the vehicle does not perform any regenerative braking without pressing your foot on the accelerator when on Drive mode?
I can confirm that there definitely is regen on drive mode, per my first graph, at least on my AZE-0.

My error, regen occurs in both modes, but in drive mode the regen is less than 50% of B mode. In both modes, regen is a function of SOC and battery temp.