Water Cooled Batt Pack

[Mark911]

MSC Engineering is almost ready to test my water cooled Alta battery pack. Pack is ready to be swapped into my bike but need to finish all the plumbing first. System uses a small electric pump, remote radiator and a modified Alta pack. Currently, the pack has only two of the four modules actively cooled so I can get real comparison data under identical conditions.

The system will eventually feature a quick change dry break connector for attaching to an auxiliary chiller for cooling the modules below ambient for more thermal headroom. In addition, the chiller will allow cooling of the pack between motos while charging, a condition which currently takes an already hot pack and makes it even hotter.

The pack weight is unchanged, as there's no need for the 8 heavy heat spreaders. The only additional weight is the radiator, water lines, pump and of course, coolant. Probably 5-6 lbs. This rather extreme mod is only necessary if you have an issue with thermal limiting. Will it solve the problem? We'll find out over the next couple months of testing.

Details for the pack cold plate and retrofit is somewhat company private right now.

Note, I'm testing my A pack, which is more susceptible to thermal issues.

 

[rayivers]

Amazing! No doubt it will drop temps significantly.

 

[Mark911]

Here's some simulation I did using ANSYS.

The first is pack temperature during heavy discharge and subsequent charging. Essentially, it shows that after about 1.5 hrs of charging the pack would be around 53 C (without any mist cooling) assuming standard ambient conditions. Thermal limiting begins around 55-60C. A good rider would probably be thermal limiting in only a few laps if they lined up in that condition. On a warm day the pack would start out 10-15C warmer and thermal limiting would occur during the initial discharge cycle.

The second simulation is a pack that was pre-chilled prior to a heavy discharge and then charged while being cooled with chilled water. Temperatures never gets close to the limiting threshold during discharge (assuming standard ambient conditions) and almost reaches the initial pre-chilled temp during the charge cycle (which would take another 30 minutes to obtain). Most motos are at least 2 hours apart, so this condition is obtainable.

Of course, simulations are only as good as the models and constraints imposed, so it'll be interesting to see how it correlates in real life.

 

[TCMB371]

I've said it many times before, but we're super lucky to have you in this community. Can't wait to see how well this helps with battery thermals! IMO the thermal performance of the battery is the only downfall of our bikes, particularly for quicker riders.

 

[Mark911]

Temporary installation of water pump, good enough for testing. Fan pulls air through the radiator.

Thanks for the kind words BH. Back at you for sure.

 

[Mark911]

Bike is ready for initial testing

 

[Judaslefourbe]

We always thought the temp delta was too small to make radiator work. Curious to see if we were right.

 

[privateer703]

So it's just a small heat sink attached to the battery? It looks like the same battery housing casting, so it's not jacketed I'm guessing. Interested to hear how well it works.

 

[Mark911]

We always thought the temp delta was too small to make radiator work. Curious to see if we were right.

The delta T isn't much, particularly on a hot day, but you have the same situation with the passive (stock) thermal mgmt system but without the torturous and limited thermal path. However, I'm incorporating a thermostat which senses the cold plate temperature and won't turn on the pump until it reaches ambient condition for that moto. Otherwise, I could be actually heating up the pack! I've mentioned it before, but as my simulations show, the real benefit will probably be the ability to circulate chilled water to cool the cells below ambient prior to the race. If in doing so my tests show that the pump never turns on then I'll simply eliminate the radiator and all the associated plumbing/weight for conditions where the ambient temps make the radiator system too inefficient. However, I have other simulations that show on cooler days that the radiator works quite efficiently and there's no need to pre-chill.

 

[Mark911]

So it's just a small heat sink attached to the battery? It looks like the same battery housing casting, so it's not jacketed I'm guessing. Interested to hear how well it works.

Well, much more than a small heat sink. The housing is heavy modified internally and every cell is thermally connected to circulating water. This naturally resulted in a change in pack height to accommodate the required cold plate cross section. However, all the tolerance was applied to the outboard modules leaving the two inboard module Z height stations unchanged since there's no way to grow the pack without some kind of belly band and this would throw off the mounting locations. The fwd and aft modules can grow in Z approx 0.100 with the addition of some simple 3D printed spacers.

 

[C5tor]

The tubing next to the chain would worry me a bit. Maybe some extra armor around the tube?

 

[Philip]

So, Mark, could you then tweak the current sensor and safely extract 60-70hp out of your MX?

 

[Mark911]

So, Mark, could you then tweak the current sensor and safely extract 60-70hp out of your MX?

That much HP would require over 45kw, or (with losses) about 180 amps of battery power at 300v compared to the 100-125 amps stock. At 6P you would need to push 30 amps per cell. Even the VTC6 would have a hard time doing that without excess voltage sag (regardless of temperature). I'm not even sure the stock HV wiring and interconnects could handle that kind of current. So I'd say, no. Just getting the stock power to last an entire moto is the goal.

 

[Philip]

That much HP would require over 45kw, or (with losses) about 180 amps of battery power at 300v compared to the 100-125 amps stock. At 6P you would need to push 30 amps per cell. Even the VTC6 would have a hard time doing that without excess voltage sag (regardless of temperature). I'm not even sure the stock HV wiring and interconnects could handle that kind of current. So I'd say, no. Just getting the stock power to last an entire moto is the goal.

Yes, now I remember that, when I asked if I could upgrade my MX to MXR, @Judaslefourbe told me to sell my MX and buy an MXR because, besides the battery, I would also need a new harness, among other things.

 

[Mark911]

There's not a hell of alot of difference in the HV harness current carrying wise between models. The MXR does use module interconnect conductor pins (what I call them) that have a bit higher contact surface area and the top cover bus bars the same. Probably to reduce V drop more than anything else. The HV connections between Batt and Inverter are about the same gauge. There's a couple simple schematic changes and the actual connectors are way different, probably the main reason behind his harness comment. What I was referring to was the probable wire gauge change needed to make 60-70 hp possible. That's a different animal.

Assuming you removed the mostly empty space lower cover and replaced it with a shallower design, there's probably enough room under the stock battery to add an additional module. By increasing the nominal voltage to 378v it might be possible to make 60-70 hp using the stock wiring. I think the stock IGBT is rated for over 450v so it might work. You'd need to figure out a way to get the system to talk with the new module and you'd probably get a bunch of over-voltage faults from the self test system, however.

 

[wildman240]

Perhaps you've already modeled it for comparison, but why not tap into the coolant line already plumed and use the bulkhead radiator and pump built-in?

 

[Mark911]

Perhaps you've already modeled it for comparison, but why not tap into the coolant line already plumed and use the bulkhead radiator and pump built-in?

The motor runs much hotter than the batt pack. It would actually heat up the pack if I used the same water. Plus, the heat dissipation of the stock system wouldn't support another load. A split system is required.

 

[Mark911]

Progress report. Installed the dry break water fittings and connected the water chiller as shown. This setup is able to reduce the two lower module temps (the only ones with a cold plate for testing) from around 97F (a hot day) to under 68F in about 45 minutes. That's about 3.3F every 5 mins. However, it took another 15 mins to pull another 5F out, indicating I was almost reaching equilibrium and to pull much more I'd need to reduce my water temp lower than the 55 degrees I had set it for. I wasn't sure what the dew point was that day and didn't see any condensation on my water lines/fittings but before I went lower I'd want to check, so I left it at 55. An interesting observation was that the non-cooled upper modules came down in temps as well, about 6F, as the two housings are not thermally isolated and therefore experience thermal cross talk..

After that I rode the bike pretty hard for 15 minutes but wasn't able to get the upper two non-cooled modules to limit. Connected multitool and checked the module temps. The upper modules were around 48-49C (120F) and my water cooled modules were 10C (18F) lower at 38-39C (102F). I originally expected that the initial 30F delta T would have been maintained but obviously there's some heat transfer going on between lower and upper modules and the passive Alta thermal management is working against me by actually heating up the pack, at least until it reaches ambient temp. What the split is between the two I have no idea, I guess I could simulate it in ANSYS. How much did the thermal cross talk help the upper modules and hurt the lower ones? Not sure, but I would suspect the uppers would be 5-6F hotter otherwise, around 52C (125F), and possibly into the beginnings of thermal limiting. The lowers would obviously be colder than experienced. In hindsight, cooling half the pack to see the actual difference it made was a bit wishful thinking as they're much more thermally coupled than I expected. In fact, when the pack is considerably colder than the ambient conditions, like after chilling on hot days, it would make sense to totally thermally isolate the pack by wrapping it in some kind of insulation.

After collecting the above data, I connected the chiller and began charging using my Std charger at 220v. Since my lower two modules were only a few degrees warmer than before (97F vs 102F) I expected almost the same 3.3F per 5min reduction in temps less whatever heating experienced due to the charging. I set the chiller to 50F this time (instead of 55) and surprisingly it took about the same 45 minutes under charge to achieve sub 68F (20C) temperatures. Although I started to notice some condensation on the fittings immediately approximate to the chiller, none was noted on the coldplate itself as it was remote and slightly warmer. Again, it took considerably longer to pull another 2-3F out, but I was able to achieve 64F before the charger turned itself off, a bit over an hour.

After that I monitored how quickly the modules would warm up under static (not riding) conditions. It was close to a degree F every 10 minutes. So you could pull the chiller and only gain a couple degrees within the 20-30 minutes it took to get ready and line up for your next moto.

Due to all the thermal cross talk contaminating my results, I'm not sure if it's worth doing more testing with only one plate or simply install the second cold plate and continue testing. Frankly, its been so hot outside lately that the last thing I'm thinking about is going to the track, lol.



View showing quick connect dry breaks.

My chiller, good to 40F.

You can clearly see the delta T between upper and lower modules in the thermal shot. Anything red was around 95F, the ambient conditions at the time.

 

[snydes]

I do enjoy reading about these experiments. Did you conclude that the radiator wasn’t providing any worthwhile benefit?

 

[Mark911]

While riding I stopped a few times to check the fan/pump status (on of off). While I'm not sure exactly when the thermostat switched them on, it was for sure in the last 5 minutes (of approx 15). That means for at least the first 10 minutes the radiator wasn't doing a thing. I'll try to study my log for that ride and see if I can see when they switched on and from a trend line standpoint how much they helped/hurt the overall thermal solution.