VionX Energy, formerly Premium Power, and now Largo Clean Energy, contracted me to help with its thermal system design & modeling for its vanadium redox flow battery.
https://www.wbur.org/news/2017/01/09/vanadium-liquid-energy-storage
Process Engineering and Modeling
Heat transfer via a liquid coolant loop
Batteries of any kind generate heat during operation and it’s the role of the thermal management system to keep the battery components at safe & optimal temperatures. Flow batteries particularly lend themselves to liquid heat transfer since one of their main components, the electrolyte, is also a liquid, and liquid-to-liquid heat exchange is quite effective.
A heat transfer device might be a chiller, an air cooler, a heater, and/or thermal storage.
Leading with questions
A major purpose of modeling is to provide a strong basis for reasonable equipment sizing & selection. I find thermal model development is best guided with a battery of questions:
- Overheating prevented?
- Too cold prevented?
- What’s the utilization of the existing equipment?
- Existing equipment optimized? (optimized for: kWhr output? $/kWhr?)
- Possible to quantify the design improvements of the next product iteration?
Thermal modeling, Excel based
I start with modeling in Excel because I like to start small and simple, and because every business has Excel so I know I can ultimately hand the model back to the client to use/modify after my contract ends.
Below, I scraped weather data for a particularly hot day and set up the model to track the temperatures of key components while we trialed different scenarios of battery & heat transfer device operation.
Fairly quickly, we bumped into the short-comings of modeling in Excel: the analysis does not scale well in data length (i.e. beyond a couple dozen hours becomes cumbersome) nor is it particularly extensible in variables / dimensions (i.e. it’s a lot of work to add another temperature to calculate or another heat transfer route). Also, in trying keeping the model simple (no macros or VBA scripts), Excel is not a great for the simultaneous solving the systems of equations that arise from the energy balance of the heat transfer routes.
But the Excel model step is important for a few reasons:
1) it can reveal big issues without undertaking a much more complicated model,
2) it provide a ‘gut check’ basis for comparison when evaluating the legitimacy of the output of a more complicated model,
3) it is a proof-of-concept to help inform you if it’s worth spending the time & effort developing the more complicated more – it’s a Go/No-Go stage gate, which is typically good practice in project management
Thermal model, Matlab based
Pleased with the Excel model, we wanted a more comprehensive picture (include more heat transfer routes, sources and sinks), and to model seasonal effects (so at least one year of data & simulated operation). With these needs, Matlab was a good choice for the model development.
You could guess that Matlab (Matrix Laboratory) is particularly good at matrix math. Thus it’s a great fit for using matrix math to quickly solve the energy balance systems of equations at every hourly data point over the 8760 hours in a year.
One simple conclusion
I cannot name any specific details about any proprietary technology…
But I can offer one simple but profound piece of advice: Paint it white!
The heat any object picks up sitting in the sun can be tremendous, and white paint reflects a significant amount of heat away.
