NHW11 battery pack voltate and SoC info

I have taken the graphs for the Gen II and just relabeled for the NHW11 from here. (which is basically taking the generic NiMH cell voltage chart and multiplying by number of cells)




Now I realize the SoC mapping to cell voltage or traction battery voltage depends on the battery temperature and the discharge / charge rate. In addition it appears depending on what MG1 is doing, you could have the same SoC reported for a wide range of voltages.

I haven't found an firm numbers for measuring in neutral or tracking the battery voltage, and I have looked through tons of Bob's graphs, and he estimates in at least one case that 294V might correspond to the 80% SOC. But many of people's graphs are around the 300V area mostly because of forced charging at the beginning of their data logging.

I was wondering if anyone could point me in the right direction for what module voltages would correspond to what SoC in a good battery module. I'm assuming older ones jump in the voltage and the reported SoC doesn't make sense. Does anyone know if the ECU just tracks voltage, or maybe dome dV/dt or dV/dI to determine an estimated SoC.

Some related threads:
Super Brain 989 | PriusChat
ReInVolt PHEV Plan | PriusChat
NHW11 Traction battery autopsy | PriusChat
Yahoo! Groups
Yahoo! Groups

 

Ok I guess what I'm asking is, are there more plots of the ECU reported SOC% versus module pairs voltage.

From Bob's work from here: ReInVolt PHEV Plan | PriusChat

We can see that force charging and hill descent with an extra 20A IB, the pack gets to a max of 330V (1.41V/cell) to 350V (1.5V/cell) (35C battery temps). In general the module pairs stay between 15.5V (1.3V/cell) and 18V (1.5V/cell). In the cooling section of his experiments with 1.2A load in N at 35C we see the pack go from 325V (1.39V/cell) to 295V (1.26V/cell) over 35 mins. And the reinvolt pack goes from 315V (1.35V/cell) to 290V (1.24V/cell) over 40mins at slightly higher temp. I would love to know what the reported SoC% might have been.

I'm just curious at trying to examine reduce battery capacity from cell voltage discharges, and I have a hunch the gas mileage might be related to the voltages and the SoC, and if battery packs happen to have a lower cell voltage (not necessarily capacity) maybe due to hot or cold climates or driving habits.

For example 315V is being reported at 60% SoC and 280V is around 50% (with 22mOhm cell resistances). I understand it is complicated with battery temp and MG1 interactions and surface charge and voltage depends on current charge/discharge, and measured cell resistances. I guess I need to research more the SoC spoofing, but I do suspect on these older battery packs might be a reason for difference in mpg (ie. incorrect behavior based on assumed SoC).


I am wondering if there are more data points on Bob's procedure outlined here:
Prius Gen2 Traction Battery Experience - Page 2 - Hybrid Car Forums

 

I meant in addition to the SOC, I meant data sets from anyone else who did this. Did anyone else post similar in N discharge voltage curves? My hunch is the batteries aren't at similar voltages and/or SOC.

 

Heh, all my googling for this on forums always return back to your reports and graphs.

I supposed it is worth looking more at the Gen II stuff and just scaling it, but they do have quite a difference performance and I assume the SOC takes that into account (from different resistance).

Also this chart Stealth_Envelope.gif by johnson487682 shows why I think many prius get worse mileage.


As the battery ages or maybe gets a history from certain driving conditions or habits, you operate around the 50% SOC and prevents S4 from using the battery enough, ie. IB limit is low so it requests the ICE more often. It really should try to operate around 70% SOC on older batteries maybe. (Though you are giving up battery life I suppose, so maybe low current limits is a good exchange for mpg instead of earlier battery replacement)

 

I took some data from djlarue Mar 17, 2010 2001 Traction Battery Failure - thrown codes P3000/P3006 and data from MiniScanner

And I think the reported SoC isn't very useful, but I think if you sort by IB currents you might get a reasonable SoC curve, maybe someone has better ideas.

The traction battery voltage here is taken by scaling the module pair voltage by 19, which seems to be good estimate based on the reported Vbat and module pair averages in real time.

 

I did some discharge in N measurements trying to determine estimated Ah left on pack.
Initial discharge at 0.6A and I realized it would take too long, so I increased to 2.2A 30 mins into the experiment. (MG1 inverter temp got up to 58C for the record without any cooling air in N)

Assuming the ECU reported SoC estimates are accurate, 3Ah at 0.6A discharge, or in 40min 70% to 40% at 2.2A which is about 4.8Ah.

The traction battery voltage drop over the first minute matches the data from Bob's measurements pretty closely over the first 30 mins, except for my IB was 0.6A and his was 1.2A, giving half the capacity of those packs. About 0.36V/min @ 0.6A and 0.9V/min @ 2.2A.





Anyways, I don't quite understand the calculations in this post but anyways I suppose you need to hook individual modules to a battery charger to get a real capacity estimate.

BTW, thanks to Bob Wilson for all those posting he did.

 

Sorry to anyone having read all this so far.. (I'm late to the party)

Start here with:

FreedomCAR Battery Test Manual For Power-Assist Hybrid Electric Vehicles, October 2003
http://avt.inl.gov/battery/pdf/freedomcar_manual_04_15_03.pdf

INL Battery Test Manual for Plug-In Hybrid Electric Vehicles, March 2008
http://www.fuelcon.com/cms/index.php?id=28&file=1A609B0&no_cache=1&uid=32

Anyways, some useful information here:
2010 Toyota Prius-6063 Hybrid BOT Battery Test Results

 

Ok, well no one has any input, but my estimate is 50% capacity battery which is just prior to the larger self-discharge region of battery life.

 

My thoughts were is more people with Scanguages could let things warm up, then with SoC at 60%, put it in N and everyone adjust the fan to have the same IB around 2A or something, then write down the Voltage, reported SoC, IB, (maybe module resistance) and Battery Temp (hi/lo) and wait 10 mins, then record again.

We could plot multiple batteries and get a pretty good estimate of battery capacity on these older batteries. Assuming the same battery temp and discharge current, everyone would be on the same curve maybe shifted left or right a little. But at least you could compare estimated Ah capacity.

 

Well, with parking brake to disable DRL and MFD off and fan off, I get pretty solid IB readings. Yes, as the pack voltage drops below 50% SoC the IB increases because of a fixed power drain (less voltage is more current). From 300V the IB was 2.2A (660W), at 276V it was 2.4A (~660W). But for a 10 min run at 1A or 2A, IB should be pretty steady.


0.6A - baseline IB current
+0.1A - MFD
+0.8A - headlights and running lights
+0.5A - DRL headlights
+0.8A - DRL headlights and running lights
+0.2A - cabin blower on lowest
+0.8A - cabin blower on high
+0.6A - defrost

2.8A - MFD, headlights, running lights, defrost, fan high
(or close to 0.6+0.1+0.8+0.8+0.6)



My proposal is just to record a few numbers (start and end points) 10 mins apart. So note the bat Temp and starting SoC and final SoC, but what matters is IB, Vpack after 10 mins when starting close to 60% SoC, then graph a few people's batteries.