Electronics > Projects, Designs, and Technical Stuff

Lithium battery cell Simulator/Emulator for BMS testing

(1/2) > >>


I am designing a new project, it is a BMS for 12 LiFePO4 cells. It is the first time I am working with those type of batteries and as I understand many things can go wrong.

The BMS PCB is responsible of monitoring cell voltages, current, temperature, drive contactors, active balance on cells etc.

Is there any idea how I can simulate/emulate the cells on hardware level, to avoid any risks during development? I am thinking to simulate the cells using some supercaps and variable resistors.

Maybe there  is any certified (or not) cell emulator tool you can suggest?

A series of large value capacitors is exactly what you want. To create an imbalance that the BMS must then resolve, temporarily connect a resistor to discharge one of the caps to less than the others.

I suppose one could consider some ESP8266 or ESP32 modules as smart loads with voltage monitoring. The wireless communications provides the isolation for you.

I'm assuming the pack is very high current hence the decision to use contactors. Even then, you might want to take a look at using MOSFETs for switching since they're cheaper than you think.

The pack current will be below 20A. MOSFETs will be used, I just used the term "contactor" to simplify the description.

Is there any tool that can be used as HIL (Hardware in the loop) for testing? Something that could simulate (on hardware level) the cell and its characteristics to simulate/test the balancer? I know the cost might be high.

Given that you can't test reall imbalance or ageing with anything but a real cell (because you don't know how the capacity of your actual cells varries with batch, temperature and SoH etc) don't bother is my suggestion.

 Simply use an adjustable resistive divider chain to "prove" your BMS can read the correct cell voltages with a decent accuracy (+- 2mV would be typical) across the range you need to read (say 3v to 4.2v for example) and that can also prove that your BMS can in fact enable and disable your balancing circuit at the correct voltage etc.

The bigger the power rating of the resistors you use and the supply you test with the higher current you can test your balancing too. The voltage measurement part obviously needs to be decently high impedance to avoid actually discharging the cells, so a simply resistive divider with say 370 ohm resistors will work fine with that at around 10mA of series current (35mW) but get down to say 36 ohms (100mA and 360mW) or even 3.6 ohm (1000mA / 3.6W) and you can almost certainly prove out your balancing as well.

At some point, when you have confidence in your BMS system basic performance on your fake resistive "cells" you will have to move to real cells. Here, the characterisation and SW determination of SoC, SoH and IR becomes the difficult bit. DO not under estimate the SW complexities in a decent BMS, especially if you want decent estimation of SoH / IR etc

So, for 12 cells in series, get a 50v DC supply that can do say at least 25W (500mA) continuously, and use a string of resistors, to simulate each cell potential in your chain, and feed those mid node votlages into your BMS   :-+

I used to use tiny 2032 sized lithium-Ion cells in series stacks exactly for this purpose. Energy limited because cells are small, and you can charge/discharge/un-balance them easily since you don't need much current to do so. This was plenty to test a custom BMS for a 5kWh LiFePO4 pack.


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version