Flow Battery Research Collective

I have still not been able to test my cell because I wanted to make sure I had my mystat calibrated correctly. For this I wanted to implement a calibration wizard. Unfortunately I was not prepared to deal with the very large single file, single layer codebase of the mystat control software. I therefore had to start refactoring this code. My progress is visible here: https://codeberg.org/sepi/mystat/src/branch/feat-refactor-modular

The advantages of my refactor are amongst others the following:

• less global state, easier to understand code
• multiple smaller modules making it easier to find code
• layered architecture, separating concerns more cleanly
• "plugin" system for easier extension
• easier testing of both hardware code and gui code due to decoupling into layers. A mock hardware class could be implemented to test the gui or other Tabd could be implemented as new classes
• availability of decoupled hardware layer that can be used independently of the gui
• easier implementation of new "Tab plugins" by subclassing and using the hardware and gui layer.

This is still work in progress and probably pretty buggy. I tried not to touch any of the implemented logic but there afe surely some problems left. Once I'm done with the refactor, I would like to write a wizard for calibration hoping to make it obsolete to consult the original paper for reference.

I'm open for comments and suggestions now and help with testing once my branch has stabilised a bit.

@kirk @danielfp248 I just did some measurements of flow rate with my setup. I measured the time to pass 40ml of deionised water through one side of the cell while the other side cycles deionised water. I tested with 11V-19V and ended up with values between 40 and 120mL/min. There is quite a bit of difference between the two pumps, the left passing up to 40% more water at the same voltage. I will repeat this once I have the PWM regulator set up correctly. I'm afraid though that the pumps have issues running at lower than the 11V, especially for more visquous fluids.

My questions are:

1. Is this 20-40ml/min measured with the cell in line or just the pump.
2. Is the figure obtained with water or electrolyte?
3. What would you recommend doing against the excess flow?

I can't speak for the FBRC but I think that your project sounds very interesting, especially combined with FBRC. I'm not sure if it will be possible to produce tangible results soon since the devkit is probably not very interesting for analysy due to its comparatively low environmental impact (very low number of kits and small amount of substances used). The second iteration of the project might be of more interest, since it is way closer to a real world product. Still, one very important aspect, chemistry, is still a moving target. In any case, sustainability should be a prime driver in the development of this project and recycling of the most important components should be considered from the start. It would probably also be very interesting to do a comparative analysyis between common lithium batteries and flow cells. The hope being that flow cells would be way more sustainable. Anyways, those are just my two cents as an interested follower of the FBRC project.

@kirk I finally found some time to advance the project. I managed to run the cell on deionised water for an hour without leakage nor bubbling or weird pressure buildup in the tanks. The issue was that one of the flow frames had a very large resistance to water flow. I guess this was some kond of print failure. Once I knew what to look for, it was very easy to diagnose by blowing through the two barbs.

I'll do some flow measurements next and then I can finally do some actual cycling. Yay!!

The tubing is gone now, thanks for buying at my store

@doho I can send you 2m if that's enough. If you manage to send me a private message, we can exchange address and account number. Else please post your email address here and we do the rest via mail!

On my phone it looks like this:

@doho that's a great idea, I'll give it a try!

@doho I sent you private messages. Im afraid I don't have 5m for sale anymore. Let me check what I can still sell when I'm home later tonight.

@kirk said in My build (very slowly progressing):

This sounds like internal leakage through/around the membrane or similar, or some type of flow obstruction in one flow path that is causing higher pressure on one side of the membrane and bulk fluid transfer from one side to the other. The pressures on each side of the cell should be roughly matched, total imbalance in an hour is too fast. All your printed parts seem to have clear internal flow channels?

That could be the culprit. When I finally find some time, I'll check if I can blow through the flow frame. An obstructed manifold would indeed explain why or gets sucked into the tanks. Thanks for the diagnostics!

Oh, about the gaskets, I have never heard of those materials but they sound promising to my untrained ears. Especially for a production version of the cell, the existing material will probably not cut it, at least not with the Zn/I chemistry.

@gus afaik, the firmware from kirks repo with the python mystat application is not actually the latest one (it's missing the extended range). Check the updated page on the mystat docs: https://fbrc.codeberg.page/rfb-dev-kit/pstat.html which links to thw correct version. I'm not 100% sure if that is the issue you are encountering but it might be.

Concerning the detached tubing, I could imagine the felt getting clogged somehow leading to an ovepressure making the tubes pop off. I'm having some issues with gas being pulled into one container, leading to pressure buildup in the container. I have no clue yet where this comes from.

Good luck with your further investigations! You might want to rzun your future experiments in a transparent PP box in the future. I will for sure once I'm ready for actual electrolyte.

After more than an hour, the negolyte tank almost overflowed and the posolyte one was almost empty. This seems weird to me because both pumps seem to run at the same RPM.

After installing tubes without my homeprinted barbs for easy removal, I managed to get the system running without a liquid leak. I've had it run for 5min with no visible leak but will let it run for an hour or so to make sure.

On the negolyte side, there is however some fizzing to be seen in the tank, also when I place the septum on the right tank, at some point it will pop off. There is thus a gas leak somewhere. Do you have a good strategy how to find this @danielfp248 @kirk ? I guess it's in the part where the liquid is sucked out of the cell and into the pump.

I managed to fix an issue with one pump running quite a bit slower than the other with equal voltage applied. I managed to trace the issue back to one for the gears in the planetary gearbox being slighly misplaced. Has anyone had those issues yet? After making sure all gears are inserted perfectly symmetrically, both pumps run pretty smoothly and less hot.

So I have gotten my pumps, off course they are the wrong ones, the brushed version and not the brushless. This creates problems since I can't easily adjust the RPMs. I instead got some c+eap PWM H-Bridge drivers to at least have some open loop control. Using this setup, I managed to do my first leak testin. As expected my homemade barbed tube connectors were not all tight and the cell leaked too. I'll try again without the connectors I added for easier cell removal. After having tightened the cell quite a bit, I think it os tight. This would need to be tested again with the aforementioned setup. So my buimd has slowed down considerably due to lack of time, but it hasn't halted :).

Thanks for the comprehensive reply, that sounds indeed very nice! I have more questions still:

1. Do you do anything differently than what is already published or is the challenge mostly to get something kind of proven working in the FBRC devkit?
2. Would it in general be an issue to just never charge the battery to high SoCs? Sure your volumetric energy density will go down but you might avoid all kinds of trouble this way. I imagine habing a very simple or lowtech battery of mediocre efficiency and density could still be of a lot of value to many people.
3. Are the low voltages due to ohmic losses in the separator and maybe electrolyte?
4. How is the patent situation around this technology?

This sounds very exciting but to be honest, as an outsider, I can't really judge. Could you explain how this chemistry compares to others and how the results are promising, etc? Also, is the low voltaic efficiency due to overpotentials because of relatively bad kinetics?

My pumps didn't have them, only the ones built into the pump. I saw that on some of your pictures, you have additional fittings between the different elements, I guess to make it easier to connect and disconnect the cell. In the end, it might make more sense to just shrink the barbs on the flowframe and/or tank a bit so the tubing comes off mor easily.