Modelling the BYD Blade Battery with WattCell: part I

Case study of modelling the BYD Blade battery with WattCell (and a bit of 'how to use').

Today, I want to show you how to use WattCell to rapidly model battery cell and play with some parameters to explore the potential performance of technology under study. We’re going to use WattCell to try some unconventional optimisations of the battery design. I’m going to first replicate the BYD Blade cell, which has been turning heads with its innovative design and impressive energy density. In second step we will check if we could push it further. Fair warning: we’re about to venture into the realm of “just because we can, doesn’t mean we should” - this is exercise for fun and discovering possibilities of the WattCell, not suggestion for design of real-world batteries.

Modelling the BYD Blade cell  

Let’s start by walking through the process of modelling the BYD Blade in WattCell. We know it’s a prismatic cell using LFP (Lithium Iron Phosphate) as the cathode material, with an aluminium case. We can find most specs on the internet ( [1]  [2]  [3]  ) and make educated guesses for the rest.

1. Choose the materials and cell format: We know it is LFP chemistry, with graphite anode and of course prismatic format. From teardown we also know that it is z-stacked.

1. Set electrodes properties: since we don’t know all the details we can actually leave all inputs as default for these materials. They are taken from built-in database of properties for most common materials. However, the teardown reports indicate areal capacity of cathode at 3.39 mAh/cm2. We can model it adjusting the cathode thickness. The other thing we need to change is the current collector for anode, which is copper for Li-ion batteries.

1. Specify Separator and Electrolyte: The same thing applies to separator. We don’t know exactly which one is used in the Blade cell, but Celgard 2325 is a good guess. However we need to change electrolyte to LiPF6. It is also worth increasing excess to 10% to get more realistic results. If you prefer, the Excess electrolyte slider can be also used to tune electrolyte volume per Ah ratio displayed below.

2. Rest of the cell configuration: We haven’t defined the n/p ratio and first cycle capacity loss (Initial Coulombic Efficiency) yet, but we can leave the default values here.
3. Define cell casing: We can take the dimensions of the cell from the sources linked at the top of this post. It is also worth changing tab material for anode to Ni.

4. Fine tuning: Now we have all known parameters specified. However, a quick glance at the calculated results (capacity, electrode dimensions, mass etc.). This discrepancy is caused by two things: a lot of parameters were just guesses and the blade cell has terminals on the sides, while model assumes the terminals on top of the cell. The second issue is easy to fix. By changing headspace (normally on top cell for tabs, safety devices etc.) to 1mm and shrinking width by 10mm we get the correct electrode sizes.

Now we are a bit closer to the expected capacity and mass.

To get even more precise results, we need to improve our guesses. Lets first check the output of modelled cell. One parameter worth checking is number of layers. We should get 26 layers according to teardown.

However, in our model we are short 2 layers - we’ve only got 24 instead of 26. Let’s assume that BYD applies a bit more advanced materials - we will change capacity of LFP to 170 mAh/g (so to get 3.4mAh/cm2 we will reduce thickness of cathode to 82um) and we will use thinner separator: Celgard 2320. After these updates the designed cell has 26 layers - same as the real one!

So we have all dimensions of the modelled cell the same as those of the physical counterpart. However, the capacity we get (~129Ah) is still quite a bit lower than nominal from datasheet (138Ah). At this point the only thing that affects calculated capacity and we haven’t tinkered with yet is the Initial Coulombic Efficiency . To get to 138Ah we have to increase it to 100%, which makes sense - the nominal capacity reported by BYD likely reflects the usable capacity after the formation cycle.

The last touch is adjusting the Extra mass to match the weight of the modelled cell with the specs. It is 150g in this case, to get total mass of 2630g. This covers all the cell parts outside of the model, such as safety vents, terminals, tapes etc.

Finally, we arrive at a model that closely matches the reported energy density of the BYD Blade cell.

And at this point I realised how long this post got. I’m very happy that you reached so far in reading this, but now it’s time for a little break. We will play with that cell model more in part II.